feat: Introduce bot-digitador service for automated data entry in Integración Total. Add CLI and GUI interfaces for CSV processing, implement auto-start functionality on Windows, and enhance README with setup instructions and usage details. Include necessary configurations in docker-compose for service integration.

services/api_gateway/api/main.py

@@ -20,6 +20,7 @@ CLEANING_URL = os.getenv("CLEANING_URL", "http://cleaning:8002").rstrip("/")
 POBLACION_APTA_URL = os.getenv("POBLACION_APTA_URL", "http://poblacion_apta:8003").rstrip("/")
 COUNT_URL = os.getenv("COUNT_URL", "http://count-api:8080").rstrip("/")
 BOT_DESCARGA_URL = os.getenv("BOT_DESCARGA_URL", "http://bot-descarga:8200").rstrip("/")
+BOT_DIGITADOR_URL = os.getenv("BOT_DIGITADOR_URL", "http://bot-digitador:8210").rstrip("/")
 
 # Timeouts: JSON ligero vs uploads OCR/conteo
 _TIMEOUT_JSON = httpx.Timeout(connect=5.0, read=30.0, write=30.0, pool=5.0)
@@ -36,12 +37,13 @@ async def lifespan(app: FastAPI):
     limits=httpx.Limits(max_connections=100, max_keepalive_connections=20),
   )
   _log.info(
-    "Gateway iniciado — extraction=%s cleaning=%s poblacion_apta=%s count=%s bot_descarga=%s",
+    "Gateway iniciado — extraction=%s cleaning=%s poblacion_apta=%s count=%s bot_descarga=%s bot_digitador=%s",
     EXTRACTION_URL,
     CLEANING_URL,
     POBLACION_APTA_URL,
     COUNT_URL,
     BOT_DESCARGA_URL,
+    BOT_DIGITADOR_URL,
   )
   yield
   await _http_client.aclose()
@@ -84,6 +86,7 @@ def _is_upload_path(path: str) -> bool:
     or path.startswith("/poblacion-apta/jobs/from-csv")
     or path.startswith("/poblacion-apta/jobs/from-cleaning")
     or path.startswith("/bot-descarga/jobs")
+    or path.startswith("/bot-digitador/jobs")
   )
 
 
@@ -105,6 +108,7 @@ async def health() -> dict[str, Any]:
   poblacion_apta = await _safe_health(client, POBLACION_APTA_URL)
   counting = await _safe_health(client, COUNT_URL)
   bot_descarga = await _safe_health(client, BOT_DESCARGA_URL)
+  bot_digitador = await _safe_health(client, BOT_DIGITADOR_URL)
   return {
     "status": "ok" if extraction["ok"] and cleaning["ok"] and poblacion_apta["ok"] else "degraded",
     "services": {
@@ -113,6 +117,7 @@ async def health() -> dict[str, Any]:
       "poblacion_apta": poblacion_apta,
       "counting": counting,
       "bot_descarga": bot_descarga,
+      "bot_digitador": bot_digitador,
     },
   }
 
@@ -183,6 +188,11 @@ async def bot_descarga_proxy(path: str, request: Request) -> Response:
   return await _proxy_request(request, f"{BOT_DESCARGA_URL}/{path}")
 
 
+@app.api_route("/bot-digitador/{path:path}", methods=["GET", "POST", "PUT", "PATCH", "DELETE"])
+async def bot_digitador_proxy(path: str, request: Request) -> Response:
+  return await _proxy_request(request, f"{BOT_DIGITADOR_URL}/{path}")
+
+
 @app.api_route("/jobs", methods=["GET", "POST"])
 async def jobs_root_proxy(request: Request) -> Response:
   return await _proxy_request(request, f"{EXTRACTION_URL}/jobs")

services/bot_descarga/api/main.py

@@ -87,26 +87,30 @@ def claim_pending(
 
 @app.post("/jobs")
 async def create_job(
-    worker_id: str = Form(...),
+    worker_id: str = Form(""),
+    worker_ids: str = Form(""),
     job_id: str = Form(...),
     title: str = Form(""),
     file: UploadFile = File(...),
     _: str = Depends(require_bearer),
 ) -> dict:
-    if not worker_id.strip():
-        raise HTTPException(status_code=400, detail="worker_id obligatorio")
+    """Crea una corrida de descarga.
+
+    - 1 worker → flujo clásico (un solo job claimable).
+    - N workers (worker_ids JSON o todos online) → padre + N shards,
+      mismo downloads_prefix del padre.
+    """
+    import json as _json
+
+    from domain.excel_split import parse_radicado_rows, rows_to_xlsx_bytes, split_rows
+    from infrastructure.storage_minio import downloads_prefix as mk_prefix
+
     parent_job_id = job_id.strip()
     if not parent_job_id:
         raise HTTPException(status_code=400, detail="job_id (Job Desktop) obligatorio")
     if not store.parent_job_exists(parent_job_id):
         raise HTTPException(status_code=404, detail="Trabajo (job) no encontrado")
 
-    if not store.is_worker_online(worker_id):
-        raise HTTPException(
-            status_code=409,
-            detail=f"Worker '{worker_id}' no está online. Enciende el agent primero.",
-        )
-
     data = await file.read()
     if not data:
         raise HTTPException(status_code=400, detail="Archivo vacío")
@@ -115,45 +119,224 @@ async def create_job(
     content_type = file.content_type or "application/octet-stream"
     title_clean = title.strip()
 
-    job = store.create_job(
-        worker_id=worker_id.strip(),
+    # Resolver workers
+    selected: List[str] = []
+    raw_ids = (worker_ids or "").strip()
+    if raw_ids:
+        try:
+            parsed = _json.loads(raw_ids)
+            if isinstance(parsed, list):
+                selected = [str(x).strip() for x in parsed if str(x).strip()]
+            elif isinstance(parsed, str) and parsed.strip():
+                selected = [parsed.strip()]
+        except _json.JSONDecodeError:
+            selected = [x.strip() for x in raw_ids.split(",") if x.strip()]
+    if not selected and worker_id.strip():
+        selected = [worker_id.strip()]
+
+    online = store.list_online_worker_ids()
+    online_set = set(online)
+    if not selected:
+        selected = list(online)
+    else:
+        missing = [w for w in selected if w not in online_set]
+        if missing:
+            raise HTTPException(
+                status_code=409,
+                detail=f"Workers no online: {', '.join(missing)}. Enciende el agent primero.",
+            )
+
+    if not selected:
+        raise HTTPException(
+            status_code=409,
+            detail="No hay bots online. Enciende al menos un agent.",
+        )
+
+    # Dedup preservando orden
+    seen: set[str] = set()
+    workers_unique: List[str] = []
+    for w in selected:
+        if w not in seen:
+            seen.add(w)
+            workers_unique.append(w)
+
+    # Parse filas
+    try:
+        rows = parse_radicado_rows(data, original)
+    except Exception as e:
+        raise HTTPException(status_code=400, detail=f"No se pudo leer el Excel/CSV: {e}") from e
+    if not rows:
+        raise HTTPException(status_code=400, detail="El archivo no tiene filas de radicado/NIT")
+
+    # Un solo bot (o una sola fila) → flujo actual
+    n_workers = min(len(workers_unique), len(rows))
+    if n_workers <= 1:
+        only = workers_unique[0]
+        job = store.create_job(
+            worker_id=only,
+            excel_key="",
+            original_filename=original,
+            title=title_clean,
+            job_id=parent_job_id,
+        )
+        task_id = store.create_downloads_task(
+            parent_job_id=parent_job_id,
+            download_job_id=job.id,
+            worker_id=only,
+            title=title_clean or f"Bot descarga ({original})",
+            original_filename=original,
+            worker_ids=[only],
+            mode="single",
+        )
+        try:
+            key = minio.upload_excel(job.id, data, original, content_type=content_type)
+        except Exception as e:
+            store.patch_job(job.id, status=JobStatus.FAILED.value, error=f"MinIO upload failed: {e}")
+            raise HTTPException(status_code=502, detail=f"Error subiendo a MinIO: {e}") from e
+
+        job = store.update_excel_key(job.id, key)
+        assert job is not None
+        store.add_event(
+            job.id,
+            "run_started",
+            f"Job creado y asignado a {only}",
+            metadata={
+                "excel_key": key,
+                "worker_id": only,
+                "job_id": parent_job_id,
+                "task_id": task_id,
+                "mode": "single",
+                "total_rows": len(rows),
+            },
+        )
+        return job.to_dict()
+
+    # Distribuido: padre + shards
+    workers_unique = workers_unique[:n_workers]
+    chunks = split_rows(rows, n_workers)
+    n = len(chunks)
+    workers_unique = workers_unique[:n]
+
+    parent = store.create_job(
+        worker_id=None,
         excel_key="",
         original_filename=original,
-        title=title_clean,
+        title=title_clean or f"Descarga distribuida ({original})",
         job_id=parent_job_id,
-    )
-    task_id = store.create_downloads_task(
-        parent_job_id=parent_job_id,
-        download_job_id=job.id,
-        worker_id=worker_id.strip(),
-        title=title_clean or f"Bot descarga ({original})",
-        original_filename=original,
-    )
-    job = store.get_job(job.id)
-    assert job is not None
-
-    try:
-        key = minio.upload_excel(job.id, data, original, content_type=content_type)
-    except Exception as e:
-        store.patch_job(job.id, status=JobStatus.FAILED.value, error=f"MinIO upload failed: {e}")
-        raise HTTPException(status_code=502, detail=f"Error subiendo a MinIO: {e}") from e
-
-    job = store.update_excel_key(job.id, key)
-    assert job is not None
-    store.add_event(
-        job.id,
-        "run_started",
-        f"Job creado y asignado a {worker_id}",
-        metadata={
-            "excel_key": key,
-            "worker_id": worker_id,
-            "job_id": parent_job_id,
-            "task_id": task_id,
+        shard_count=n,
+        status=JobStatus.PENDING.value,
+        result={
+            "mode": "distributed",
+            "shard_ids": [],
+            "total_rows": len(rows),
+            "total": len(rows),
+            "processed": 0,
+            "exitosos": 0,
+            "fallidos": 0,
         },
     )
+    prefix = parent.downloads_prefix or mk_prefix(parent.id)
+
+    task_id = store.create_downloads_task(
+        parent_job_id=parent_job_id,
+        download_job_id=parent.id,
+        worker_id=None,
+        title=title_clean or f"Bot descarga ×{n} ({original})",
+        original_filename=original,
+        worker_ids=workers_unique,
+        mode="distributed",
+    )
+
+    try:
+        full_key = minio.upload_excel(parent.id, data, original, content_type=content_type)
+    except Exception as e:
+        store.patch_job(parent.id, status=JobStatus.FAILED.value, error=f"MinIO upload failed: {e}")
+        raise HTTPException(status_code=502, detail=f"Error subiendo a MinIO: {e}") from e
+
+    store.update_excel_key(parent.id, full_key)
+
+    shard_ids: List[str] = []
+    for i, (chunk, wid) in enumerate(zip(chunks, workers_unique)):
+        shard_bytes = rows_to_xlsx_bytes(chunk)
+        shard = store.create_job(
+            worker_id=wid,
+            excel_key="",
+            original_filename=f"shard_{i + 1}_of_{n}.xlsx",
+            title=f"{title_clean or 'Descarga'} · shard {i + 1}/{n}",
+            job_id=parent_job_id,
+            task_id=task_id,
+            downloads_prefix_override=prefix,
+            parent_download_job_id=parent.id,
+            shard_index=i,
+            shard_count=n,
+            result={"total": len(chunk), "processed": 0},
+        )
+        try:
+            shard_key = minio.upload_excel(
+                shard.id,
+                shard_bytes,
+                f"shard_{i + 1}_of_{n}.xlsx",
+                content_type=(
+                    "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+                ),
+            )
+        except Exception as e:
+            store.patch_job(
+                parent.id,
+                status=JobStatus.FAILED.value,
+                error=f"MinIO shard upload failed: {e}",
+            )
+            raise HTTPException(
+                status_code=502, detail=f"Error subiendo shard a MinIO: {e}"
+            ) from e
+        store.update_excel_key(shard.id, shard_key)
+        shard_ids.append(shard.id)
+        store.add_event(
+            parent.id,
+            "info",
+            f"Shard {i + 1}/{n} → {wid} ({len(chunk)} filas)",
+            metadata={"shard_id": shard.id, "worker_id": wid, "rows": len(chunk)},
+        )
+
+    store.patch_job(
+        parent.id,
+        result={
+            "mode": "distributed",
+            "shard_ids": shard_ids,
+            "total_rows": len(rows),
+            "total": len(rows),
+            "processed": 0,
+            "exitosos": 0,
+            "fallidos": 0,
+            "downloads_prefix": prefix,
+            "worker_ids": workers_unique,
+        },
+    )
+    store.add_event(
+        parent.id,
+        "run_started",
+        f"Corrida distribuida: {n} bots, {len(rows)} filas",
+        metadata={
+            "excel_key": full_key,
+            "worker_ids": workers_unique,
+            "job_id": parent_job_id,
+            "task_id": task_id,
+            "mode": "distributed",
+            "shard_ids": shard_ids,
+            "total_rows": len(rows),
+        },
+    )
+    job = store.get_job(parent.id)
+    assert job is not None
     return job.to_dict()
 
 
+@app.post("/internal/reclaim")
+def reclaim_jobs(_: str = Depends(require_bearer)) -> dict:
+    n = store.reclaim_stale_running_jobs()
+    return {"reclaimed": n}
+
+
 @app.get("/jobs")
 def list_jobs(
     limit: int = Query(100, ge=1, le=500),
@@ -168,7 +351,15 @@ def get_job(download_job_id: str, _: str = Depends(require_bearer)) -> dict:
     job = store.get_job(download_job_id)
     if not job:
         raise HTTPException(status_code=404, detail="Job no encontrado")
-    return job.to_dict()
+    # Refrescar consolidación si es padre distribuido
+    if job.is_distributed_parent:
+        refreshed = store.refresh_parent_from_shards(download_job_id)
+        if refreshed:
+            job = refreshed
+    payload = job.to_dict()
+    if job.is_distributed_parent:
+        payload["shards"] = [s.to_dict() for s in store.list_shards(job.id)]
+    return payload
 
 
 @app.get("/jobs/{download_job_id}/files")
@@ -279,7 +470,8 @@ def post_event(
     body: EventBody,
     _: str = Depends(require_bearer),
 ) -> dict:
-    if not store.get_job(download_job_id):
+    job = store.get_job(download_job_id)
+    if not job:
         raise HTTPException(status_code=404, detail="Job no encontrado")
     event = store.add_event(
         download_job_id,
@@ -292,8 +484,7 @@ def post_event(
         metadata=body.metadata,
         timestamp=body.timestamp,
     )
-    job = store.get_job(download_job_id)
-    if job and body.index is not None and body.total is not None:
+    if body.index is not None and body.total is not None:
         result = dict(job.result or {})
         result["processed"] = body.index
         result["total"] = body.total
@@ -327,4 +518,7 @@ def cancel_flag(download_job_id: str, _: str = Depends(require_bearer)) -> dict:
     job = store.get_job(download_job_id)
     if not job:
         raise HTTPException(status_code=404, detail="Job no encontrado")
-    return {"cancel_requested": job.cancel_requested, "status": job.status}
+    return {
+        "cancel_requested": store.is_cancel_requested(download_job_id),
+        "status": job.status,
+    }

services/bot_descarga/domain/excel_split.py

@@ -0,0 +1,99 @@
+"""Parseo y partición de Excel/CSV de radicados para descarga distribuida."""
+from __future__ import annotations
+
+import csv
+import io
+import re
+from typing import Any
+
+from openpyxl import Workbook, load_workbook
+
+_HEADER_NAMES = ("radicado", "proceso", "numero", "número")
+
+
+def _clean_nit(valor: Any) -> str:
+    valor = str(valor).strip()
+    valor = re.sub(r'[\\/:*?"<>|]', "_", valor)
+    return valor if valor else "SIN_NIT"
+
+
+def parse_radicado_rows(data: bytes, filename: str = "input.xlsx") -> list[tuple[str, str]]:
+    """Devuelve lista de (radicado, nit)."""
+    name = (filename or "input.xlsx").lower()
+    if name.endswith(".csv"):
+        return _parse_csv(data)
+    return _parse_xlsx(data)
+
+
+def _parse_csv(data: bytes) -> list[tuple[str, str]]:
+    text = None
+    for enc in ("utf-8-sig", "utf-8", "latin-1", "cp1252"):
+        try:
+            text = data.decode(enc)
+            break
+        except UnicodeDecodeError:
+            continue
+    if text is None:
+        text = data.decode("utf-8", errors="replace")
+    sample = text[:4096]
+    try:
+        dialect = csv.Sniffer().sniff(sample, delimiters="|,;\t")
+        delim = dialect.delimiter
+    except csv.Error:
+        delim = "|" if sample.count("|") >= sample.count(",") else ","
+    reader = csv.reader(io.StringIO(text), delimiter=delim)
+    out: list[tuple[str, str]] = []
+    for fila in reader:
+        if not fila:
+            continue
+        radicado = fila[0] if len(fila) > 0 else None
+        nit = fila[1] if len(fila) > 1 else None
+        if not radicado or str(radicado).strip() in ("", "None"):
+            continue
+        radicado_s = str(radicado).strip()
+        if radicado_s.lower() in _HEADER_NAMES:
+            continue
+        out.append((radicado_s, _clean_nit(nit)))
+    return out
+
+
+def _parse_xlsx(data: bytes) -> list[tuple[str, str]]:
+    wb = load_workbook(io.BytesIO(data), data_only=True)
+    ws = wb.active
+    out: list[tuple[str, str]] = []
+    for fila in ws.iter_rows(values_only=True):
+        radicado = fila[0] if len(fila) > 0 else None
+        nit = fila[1] if len(fila) > 1 else None
+        if not radicado or str(radicado).strip() in ("", "None"):
+            continue
+        radicado_s = str(radicado).strip()
+        if radicado_s.lower() in _HEADER_NAMES:
+            continue
+        out.append((radicado_s, _clean_nit(nit)))
+    return out
+
+
+def split_rows(rows: list[tuple[str, str]], n: int) -> list[list[tuple[str, str]]]:
+    """Parte filas en n trozos lo más equilibrados posible (chunks contiguos)."""
+    if n <= 1 or not rows:
+        return [rows]
+    n = min(n, len(rows))
+    size, rem = divmod(len(rows), n)
+    chunks: list[list[tuple[str, str]]] = []
+    idx = 0
+    for i in range(n):
+        take = size + (1 if i < rem else 0)
+        chunks.append(rows[idx : idx + take])
+        idx += take
+    return [c for c in chunks if c]
+
+
+def rows_to_xlsx_bytes(rows: list[tuple[str, str]]) -> bytes:
+    wb = Workbook()
+    ws = wb.active
+    ws.append(["radicado", "nit"])
+    for radicado, nit in rows:
+        ws.append([radicado, nit])
+    buf = io.BytesIO()
+    wb.save(buf)
+    return buf.getvalue()

services/bot_descarga/domain/models.py

@@ -40,7 +40,7 @@ class Worker:
 @dataclass
 class Job:
     id: str
-    worker_id: str
+    worker_id: Optional[str]
     status: str = JobStatus.PENDING.value
     title: str = ""
     excel_key: str = ""
@@ -54,10 +54,25 @@ class Job:
     finished_at: Optional[str] = None
     job_id: Optional[str] = None
     task_id: Optional[str] = None
+    parent_download_job_id: Optional[str] = None
+    shard_index: Optional[int] = None
+    shard_count: Optional[int] = None
 
     def to_dict(self) -> dict:
         return asdict(self)
 
+    @property
+    def is_distributed_parent(self) -> bool:
+        return (
+            self.parent_download_job_id is None
+            and self.shard_count is not None
+            and int(self.shard_count) > 1
+        )
+
+    @property
+    def is_shard(self) -> bool:
+        return bool(self.parent_download_job_id)
+
 
 @dataclass
 class JobEventRecord:

services/bot_descarga/infrastructure/database.py

@@ -103,7 +103,12 @@ class DownloadJobRow(Base):
     task_id: Mapped[str | None] = mapped_column(
         String(36), ForeignKey("tasks.id"), nullable=True, index=True
     )
-    worker_id: Mapped[str] = mapped_column(String(36), index=True)
+    worker_id: Mapped[str | None] = mapped_column(String(36), nullable=True, index=True)
+    parent_download_job_id: Mapped[str | None] = mapped_column(
+        String(36), ForeignKey("download_jobs.id"), nullable=True, index=True
+    )
+    shard_index: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    shard_count: Mapped[int | None] = mapped_column(Integer, nullable=True)
     status: Mapped[str] = mapped_column(String(32), default="pending", index=True)
     title: Mapped[str | None] = mapped_column(Text, nullable=True)
     excel_key: Mapped[str] = mapped_column(Text, default="")

services/bot_descarga/infrastructure/storage_minio.py

@@ -194,21 +194,25 @@ class MinioStorage:
 
         Si el manifest falta, falla o queda vacío, hace fallback a list_keys
         bajo el prefijo (cubre varias carpetas NIT planas).
+        En corridas distribuidas varios shards sobrescriben el manifest parcial:
+        siempre se fusiona con list_keys del prefijo.
         """
         prefix = prefix or downloads_prefix(job_id)
-        files: List[dict] = []
+        by_key: dict[str, dict] = {}
         try:
             raw = self.get_bytes(manifest_key(job_id))
             payload = json.loads(raw.decode("utf-8"))
             for key in payload.get("files") or []:
                 entry = self._file_entry(str(key))
                 if entry is not None:
-                    files.append(entry)
+                    by_key[entry["key"]] = entry
         except Exception:
-            files = []
+            pass
 
-        if not files:
-            files = self._list_pdf_entries(prefix)
+        for entry in self._list_pdf_entries(prefix):
+            by_key[entry["key"]] = entry
+
+        files = list(by_key.values())
         return {"job_id": job_id, "prefix": prefix, "files": files, "count": len(files)}
 
     def _list_pdf_entries(self, prefix: str) -> List[dict]:

services/bot_descarga/infrastructure/store.py

@@ -73,6 +73,9 @@ class JobStore:
             finished_at=_iso(row.finished_at),
             job_id=row.job_id,
             task_id=row.task_id,
+            parent_download_job_id=row.parent_download_job_id,
+            shard_index=row.shard_index,
+            shard_count=row.shard_count,
         )
 
     def _refresh_worker_status(self, worker: Worker) -> Worker:
@@ -129,7 +132,13 @@ class JobStore:
                 WorkerStatus.BUSY.value if row.current_job_id else WorkerStatus.ONLINE.value
             )
             session.flush()
-            return self._row_worker(row)
+            worker = self._row_worker(row)
+        # Reclaim fuera de la sesión del heartbeat para no bloquear
+        try:
+            self.reclaim_stale_running_jobs()
+        except Exception:
+            pass
+        return worker
 
     def get_worker(self, worker_id: str) -> Optional[Worker]:
         with session_scope() as session:
@@ -139,12 +148,23 @@ class JobStore:
             return self._row_worker(row)
 
     def list_workers(self) -> List[Worker]:
+        try:
+            self.reclaim_stale_running_jobs()
+        except Exception:
+            pass
         with session_scope() as session:
             rows = session.scalars(
                 select(DownloadWorkerRow).order_by(DownloadWorkerRow.id)
             ).all()
             return [self._row_worker(r) for r in rows]
 
+    def list_online_worker_ids(self) -> List[str]:
+        return [
+            w.id
+            for w in self.list_workers()
+            if w.status in (WorkerStatus.ONLINE.value, WorkerStatus.BUSY.value)
+        ]
+
     def is_worker_online(self, worker_id: str) -> bool:
         worker = self.get_worker(worker_id)
         if not worker:
@@ -162,31 +182,41 @@ class JobStore:
 
     def create_job(
         self,
-        worker_id: str,
+        worker_id: Optional[str],
         excel_key: str,
         original_filename: str = "",
         title: str = "",
         *,
         job_id: Optional[str] = None,
         task_id: Optional[str] = None,
+        downloads_prefix_override: Optional[str] = None,
+        parent_download_job_id: Optional[str] = None,
+        shard_index: Optional[int] = None,
+        shard_count: Optional[int] = None,
+        result: Optional[Dict[str, Any]] = None,
+        status: Optional[str] = None,
+        download_id: Optional[str] = None,
     ) -> Job:
         import uuid as uuid_mod
 
-        download_id = str(uuid_mod.uuid4())
-        prefix = downloads_prefix(download_id)
+        download_id = download_id or str(uuid_mod.uuid4())
+        prefix = downloads_prefix_override or downloads_prefix(download_id)
         with session_scope() as session:
             row = DownloadJobRow(
                 id=download_id,
                 job_id=job_id,
                 task_id=task_id,
                 worker_id=worker_id,
-                status=JobStatus.PENDING.value,
+                parent_download_job_id=parent_download_job_id,
+                shard_index=shard_index,
+                shard_count=shard_count,
+                status=status or JobStatus.PENDING.value,
                 title=title or f"Descarga {original_filename}",
                 excel_key=excel_key,
                 downloads_prefix=prefix,
                 original_filename=original_filename,
                 cancel_requested=False,
-                result={},
+                result=result or {},
                 created_at=utcnow(),
             )
             session.add(row)
@@ -205,34 +235,57 @@ class JobStore:
         limit: int = 100,
         *,
         job_id: Optional[str] = None,
+        parents_only: bool = True,
     ) -> List[Job]:
         with session_scope() as session:
-            stmt = select(DownloadJobRow).order_by(DownloadJobRow.created_at.desc()).limit(limit)
+            stmt = select(DownloadJobRow).order_by(DownloadJobRow.created_at.desc())
             if job_id:
-                stmt = (
-                    select(DownloadJobRow)
-                    .where(DownloadJobRow.job_id == job_id)
-                    .order_by(DownloadJobRow.created_at.desc())
-                    .limit(limit)
-                )
+                stmt = stmt.where(DownloadJobRow.job_id == job_id)
+            if parents_only:
+                stmt = stmt.where(DownloadJobRow.parent_download_job_id.is_(None))
+            stmt = stmt.limit(limit)
             rows = session.scalars(stmt).all()
             return [self._row_job(r) for r in rows]
 
+    def list_shards(self, parent_id: str) -> List[Job]:
+        with session_scope() as session:
+            rows = session.scalars(
+                select(DownloadJobRow)
+                .where(DownloadJobRow.parent_download_job_id == parent_id)
+                .order_by(DownloadJobRow.shard_index.asc())
+            ).all()
+            return [self._row_job(r) for r in rows]
+
     def claim_pending(self, worker_id: str) -> Optional[Job]:
+        try:
+            self.reclaim_stale_running_jobs()
+        except Exception:
+            pass
         with session_scope() as session:
             worker = session.get(DownloadWorkerRow, worker_id)
             if not worker or worker.current_job_id:
                 return None
 
-            row = session.scalars(
+            # No claimar padres orquestadores (shard_count > 1 sin parent)
+            rows = session.scalars(
                 select(DownloadJobRow)
                 .where(
                     DownloadJobRow.worker_id == worker_id,
                     DownloadJobRow.status == JobStatus.PENDING.value,
                 )
                 .order_by(DownloadJobRow.created_at.asc())
-                .limit(1)
-            ).first()
+            ).all()
+            row = None
+            for candidate in rows:
+                is_parent = (
+                    candidate.parent_download_job_id is None
+                    and candidate.shard_count is not None
+                    and int(candidate.shard_count) > 1
+                )
+                if is_parent:
+                    continue
+                row = candidate
+                break
             if not row:
                 return None
 
@@ -243,9 +296,19 @@ class JobStore:
             worker.status = WorkerStatus.BUSY.value
             worker.last_heartbeat = now
 
-            if row.task_id:
-                task = session.get(Task, row.task_id)
-                if task:
+            # Task del padre (shards heredan task_id del padre)
+            task_id = row.task_id
+            if row.parent_download_job_id:
+                parent = session.get(DownloadJobRow, row.parent_download_job_id)
+                if parent:
+                    task_id = parent.task_id
+                    if parent.status == JobStatus.PENDING.value:
+                        parent.status = JobStatus.RUNNING.value
+                        parent.started_at = parent.started_at or now
+
+            if task_id:
+                task = session.get(Task, task_id)
+                if task and task.status == JobStatus.PENDING.value:
                     task.status = JobStatus.RUNNING.value
                     task.started_at = now
                     task.updated_at = now
@@ -271,12 +334,44 @@ class JobStore:
                 row.status = JobStatus.CANCELLED.value
                 row.finished_at = now
                 self._sync_task(session, row, status=JobStatus.CANCELLED.value)
+
+            # Cancelar shards hijos si es padre distribuido
+            if row.parent_download_job_id is None and row.shard_count and int(row.shard_count) > 1:
+                children = session.scalars(
+                    select(DownloadJobRow).where(
+                        DownloadJobRow.parent_download_job_id == job_id
+                    )
+                ).all()
+                for child in children:
+                    if child.status in (
+                        JobStatus.COMPLETED.value,
+                        JobStatus.FAILED.value,
+                        JobStatus.CANCELLED.value,
+                    ):
+                        continue
+                    child.cancel_requested = True
+                    if child.status == JobStatus.PENDING.value:
+                        child.status = JobStatus.CANCELLED.value
+                        child.finished_at = now
+                        worker = session.get(DownloadWorkerRow, child.worker_id) if child.worker_id else None
+                        if worker and worker.current_job_id == child.id:
+                            worker.current_job_id = None
+                            worker.status = WorkerStatus.ONLINE.value
+
             session.flush()
             return self._row_job(row)
 
     def is_cancel_requested(self, job_id: str) -> bool:
         job = self.get_job(job_id)
-        return bool(job and job.cancel_requested)
+        if not job:
+            return False
+        if job.cancel_requested:
+            return True
+        if job.parent_download_job_id:
+            parent = self.get_job(job.parent_download_job_id)
+            if parent and parent.cancel_requested:
+                return True
+        return False
 
     def _sync_task(
         self,
@@ -287,9 +382,15 @@ class JobStore:
         result: Optional[Dict[str, Any]] = None,
         error: Optional[str] = None,
     ) -> None:
-        if not row.task_id:
+        # Sync siempre contra el padre si es shard
+        target = row
+        if row.parent_download_job_id:
+            parent = session.get(DownloadJobRow, row.parent_download_job_id)
+            if parent:
+                target = parent
+        if not target.task_id:
             return
-        task = session.get(Task, row.task_id)
+        task = session.get(Task, target.task_id)
         if not task:
             return
         now = utcnow()
@@ -298,7 +399,7 @@ class JobStore:
         if result is not None:
             task.result = {
                 **dict(task.result or {}),
-                "download_job_id": row.id,
+                "download_job_id": target.id,
                 **result,
             }
         if error is not None:
@@ -319,6 +420,7 @@ class JobStore:
         result: Optional[Dict[str, Any]] = None,
         error: Optional[str] = None,
     ) -> Optional[Job]:
+        parent_id: Optional[str] = None
         with session_scope() as session:
             row = session.get(DownloadJobRow, job_id)
             if not row:
@@ -340,22 +442,263 @@ class JobStore:
                 row.finished_at = row.finished_at or utcnow()
                 finished = True
 
-            self._sync_task(
-                session,
-                row,
-                status=new_status if status else None,
-                result=result,
-                error=error,
-            )
+            # No sync task directo en shards; se consolida en el padre
+            if not row.parent_download_job_id:
+                self._sync_task(
+                    session,
+                    row,
+                    status=new_status if status else None,
+                    result=result,
+                    error=error,
+                )
 
             if finished:
-                worker = session.get(DownloadWorkerRow, row.worker_id)
+                worker = session.get(DownloadWorkerRow, row.worker_id) if row.worker_id else None
                 if worker and worker.current_job_id == job_id:
                     worker.current_job_id = None
                     worker.status = WorkerStatus.ONLINE.value
 
+            parent_id = row.parent_download_job_id
             session.flush()
-            return self._row_job(row)
+            job = self._row_job(row)
+
+        if parent_id:
+            self.refresh_parent_from_shards(parent_id)
+            refreshed = self.get_job(job_id)
+            return refreshed or job
+        return job
+
+    def refresh_parent_from_shards(self, parent_id: str) -> Optional[Job]:
+        with session_scope() as session:
+            parent = session.get(DownloadJobRow, parent_id)
+            if not parent:
+                return None
+            children = list(
+                session.scalars(
+                    select(DownloadJobRow).where(
+                        DownloadJobRow.parent_download_job_id == parent_id
+                    )
+                ).all()
+            )
+            if not children:
+                return self._row_job(parent)
+
+            statuses = [c.status for c in children]
+            terminal = {
+                JobStatus.COMPLETED.value,
+                JobStatus.FAILED.value,
+                JobStatus.CANCELLED.value,
+            }
+            any_active = any(
+                s in (JobStatus.PENDING.value, JobStatus.RUNNING.value) for s in statuses
+            )
+            all_terminal = all(s in terminal for s in statuses)
+
+            total = 0
+            exitosos = 0
+            fallidos = 0
+            uploaded = 0
+            for c in children:
+                r = dict(c.result or {})
+                total += int(r.get("total") or 0)
+                exitosos += int(r.get("exitosos") or 0)
+                fallidos += int(r.get("fallidos") or 0)
+                uploaded += int(r.get("uploaded_count") or 0)
+                # Progreso parcial mientras corre
+                if c.status == JobStatus.RUNNING.value:
+                    total = max(total, int(r.get("total") or 0))
+                    exitosos = max(exitosos, int(r.get("processed") or r.get("exitosos") or 0))
+
+            # Agregar progreso en vuelo (processed) de shards running
+            processed = 0
+            for c in children:
+                r = dict(c.result or {})
+                if c.status in terminal:
+                    processed += int(r.get("exitosos") or 0) + int(r.get("fallidos") or 0)
+                else:
+                    processed += int(r.get("processed") or 0)
+
+            shard_ids = [c.id for c in children]
+            parent_result = dict(parent.result or {})
+            parent_result.update(
+                {
+                    "mode": "distributed",
+                    "shard_ids": shard_ids,
+                    "total": total or int(parent_result.get("total_rows") or 0),
+                    "total_rows": int(parent_result.get("total_rows") or total or 0),
+                    "exitosos": exitosos,
+                    "fallidos": fallidos,
+                    "processed": processed,
+                    "uploaded_count": uploaded,
+                    "downloads_prefix": parent.downloads_prefix,
+                    "shards": [
+                        {
+                            "id": c.id,
+                            "worker_id": c.worker_id,
+                            "status": c.status,
+                            "shard_index": c.shard_index,
+                            "result": dict(c.result or {}),
+                            "error": c.error,
+                        }
+                        for c in children
+                    ],
+                }
+            )
+            parent.result = parent_result
+
+            now = utcnow()
+            if any_active:
+                parent.status = JobStatus.RUNNING.value
+                parent.started_at = parent.started_at or now
+                parent.finished_at = None
+                parent.error = None
+                self._sync_task(
+                    session,
+                    parent,
+                    status=JobStatus.RUNNING.value,
+                    result=parent_result,
+                )
+            elif all_terminal:
+                any_failed = any(s == JobStatus.FAILED.value for s in statuses)
+                any_cancelled = any(s == JobStatus.CANCELLED.value for s in statuses)
+                all_cancelled = all(s == JobStatus.CANCELLED.value for s in statuses)
+                if all_cancelled:
+                    parent.status = JobStatus.CANCELLED.value
+                elif any_failed:
+                    parent.status = JobStatus.FAILED.value
+                    parent.error = next(
+                        (c.error for c in children if c.error),
+                        "Uno o más shards fallaron",
+                    )
+                elif any_cancelled and not any_failed:
+                    # mezclado completed+cancelled → completed parcial
+                    parent.status = JobStatus.COMPLETED.value
+                else:
+                    parent.status = JobStatus.COMPLETED.value
+                parent.finished_at = parent.finished_at or now
+                self._sync_task(
+                    session,
+                    parent,
+                    status=parent.status,
+                    result=parent_result,
+                    error=parent.error,
+                )
+
+            session.flush()
+            return self._row_job(parent)
+
+    def reclaim_stale_running_jobs(self) -> int:
+        """Jobs running cuyo worker está offline → pending + reasignar a otro online."""
+        reclaimed = 0
+        with session_scope() as session:
+            now = utcnow()
+            cutoff = now - timedelta(seconds=HEARTBEAT_OFFLINE_SECONDS)
+            workers = list(session.scalars(select(DownloadWorkerRow)).all())
+            offline_ids: set[str] = set()
+            online_free: List[DownloadWorkerRow] = []
+            for w in workers:
+                hb = w.last_heartbeat
+                if hb is None:
+                    offline_ids.add(w.id)
+                    continue
+                if hb.tzinfo is None:
+                    hb = hb.replace(tzinfo=timezone.utc)
+                if hb < cutoff:
+                    offline_ids.add(w.id)
+                    if w.current_job_id:
+                        w.current_job_id = None
+                    w.status = WorkerStatus.OFFLINE.value
+                elif not w.current_job_id:
+                    online_free.append(w)
+
+            if not offline_ids:
+                session.flush()
+                return 0
+
+            running = session.scalars(
+                select(DownloadJobRow).where(
+                    DownloadJobRow.status == JobStatus.RUNNING.value,
+                    DownloadJobRow.worker_id.in_(list(offline_ids)),
+                )
+            ).all()
+
+            rr = 0
+            for row in running:
+                # Solo shards o jobs simples (no padres orquestadores)
+                is_parent = (
+                    row.parent_download_job_id is None
+                    and row.shard_count is not None
+                    and int(row.shard_count) > 1
+                )
+                if is_parent:
+                    continue
+
+                row.status = JobStatus.PENDING.value
+                row.started_at = None
+                row.cancel_requested = False
+
+                if online_free:
+                    target = online_free[rr % len(online_free)]
+                    rr += 1
+                    row.worker_id = target.id
+                # Si no hay online free, queda pending con el mismo worker_id
+                # (se reasignará cuando haya free en el próximo reclaim)
+
+                reclaimed += 1
+                session.add(
+                    DownloadJobEventRow(
+                        download_job_id=row.id,
+                        event_type="warning",
+                        message=(
+                            f"Worker offline; job reclaimed"
+                            + (f" → {row.worker_id}" if row.worker_id else "")
+                        ),
+                        timestamp=now,
+                        metadata_json={"reclaimed": True, "worker_id": row.worker_id},
+                    )
+                )
+                if row.parent_download_job_id:
+                    session.add(
+                        DownloadJobEventRow(
+                            download_job_id=row.parent_download_job_id,
+                            event_type="warning",
+                            message=f"Shard {row.shard_index} reclamado (worker offline)",
+                            timestamp=now,
+                            metadata_json={
+                                "shard_id": row.id,
+                                "worker_id": row.worker_id,
+                            },
+                        )
+                    )
+
+            # Reasignar pending de workers offline a online free
+            if online_free:
+                pending_orphan = session.scalars(
+                    select(DownloadJobRow).where(
+                        DownloadJobRow.status == JobStatus.PENDING.value,
+                        DownloadJobRow.worker_id.in_(list(offline_ids)),
+                    )
+                ).all()
+                for row in pending_orphan:
+                    is_parent = (
+                        row.parent_download_job_id is None
+                        and row.shard_count is not None
+                        and int(row.shard_count) > 1
+                    )
+                    if is_parent:
+                        continue
+                    target = online_free[rr % len(online_free)]
+                    rr += 1
+                    if row.worker_id != target.id:
+                        row.worker_id = target.id
+                        reclaimed += 1
+
+            session.flush()
+
+        # Refrescar padres afectados
+        # (se hace en siguiente patch; aquí no hace falta)
+
+        return reclaimed
 
     def add_event(
         self,
@@ -385,6 +728,28 @@ class JobStore:
             )
             session.add(ev)
             session.flush()
+            # Mirror a padre si es shard
+            row = session.get(DownloadJobRow, job_id)
+            if row and row.parent_download_job_id:
+                session.add(
+                    DownloadJobEventRow(
+                        download_job_id=row.parent_download_job_id,
+                        event_type=event_type,
+                        message=message,
+                        timestamp=ts,
+                        radicado=radicado,
+                        nit=nit,
+                        idx=index,
+                        total=total,
+                        metadata_json={
+                            **(metadata or {}),
+                            "shard_id": job_id,
+                            "shard_index": row.shard_index,
+                            "worker_id": row.worker_id,
+                        },
+                    )
+                )
+                session.flush()
             return JobEventRecord(
                 id=int(ev.id),
                 job_id=job_id,
@@ -429,9 +794,11 @@ class JobStore:
         *,
         parent_job_id: str,
         download_job_id: str,
-        worker_id: str,
+        worker_id: Optional[str],
         title: str,
         original_filename: str,
+        worker_ids: Optional[List[str]] = None,
+        mode: str = "single",
     ) -> str:
         """Crea Task(type=downloads) y la asocia al download_job. Devuelve task_id."""
         with session_scope() as session:
@@ -443,10 +810,12 @@ class JobStore:
                 title=title or f"Bot descarga ({original_filename})",
                 request={
                     "worker_id": worker_id,
+                    "worker_ids": worker_ids or ([worker_id] if worker_id else []),
                     "original_filename": original_filename,
                     "download_job_id": download_job_id,
+                    "mode": mode,
                 },
-                result={"download_job_id": download_job_id},
+                result={"download_job_id": download_job_id, "mode": mode},
                 created_at=now,
                 updated_at=now,
             )

services/bot_descarga/requirements.txt

@@ -3,6 +3,7 @@ uvicorn[standard]>=0.27.0
 python-multipart>=0.0.9
 boto3>=1.34.0
 httpx>=0.27.0
+openpyxl>=3.1.0
 pydantic>=2.0.0
 sqlalchemy>=2.0.0
 psycopg[binary]>=3.1.0

services/bot_descarga/tests/test_excel_split.py

@@ -0,0 +1,33 @@
+"""Tests de partición Excel para descarga distribuida."""
+from __future__ import annotations
+
+from domain.excel_split import parse_radicado_rows, rows_to_xlsx_bytes, split_rows
+
+
+def test_split_rows_balanced():
+    rows = [(str(i), "NIT") for i in range(100)]
+    chunks = split_rows(rows, 2)
+    assert len(chunks) == 2
+    assert len(chunks[0]) == 50
+    assert len(chunks[1]) == 50
+
+
+def test_split_rows_uneven():
+    rows = [(str(i), "N") for i in range(10)]
+    chunks = split_rows(rows, 3)
+    assert len(chunks) == 3
+    assert sum(len(c) for c in chunks) == 10
+    assert [len(c) for c in chunks] == [4, 3, 3]
+
+
+def test_roundtrip_xlsx():
+    rows = [("RAD-1", "900"), ("RAD-2", "901")]
+    data = rows_to_xlsx_bytes(rows)
+    parsed = parse_radicado_rows(data, "shard.xlsx")
+    assert parsed == rows
+
+
+def test_split_more_workers_than_rows():
+    rows = [("A", "1"), ("B", "2")]
+    chunks = split_rows(rows, 5)
+    assert len(chunks) == 2

services/bot_digitador/Dockerfile

@@ -0,0 +1,28 @@
+FROM python:3.12-slim
+
+ENV PYTHONDONTWRITEBYTECODE=1 \
+    PYTHONUNBUFFERED=1 \
+    PYTHONPATH=/app
+
+WORKDIR /app
+
+RUN apt-get update \
+    && apt-get install -y --no-install-recommends curl \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY requirements.txt /app/requirements.txt
+RUN pip install --no-cache-dir -r /app/requirements.txt
+
+COPY domain /app/domain
+COPY infrastructure /app/infrastructure
+COPY api /app/api
+
+ENV BOT_API_HOST=0.0.0.0
+ENV BOT_API_PORT=8210
+
+EXPOSE 8210
+
+HEALTHCHECK --interval=10s --timeout=5s --retries=5 --start-period=15s \
+  CMD curl -f http://127.0.0.1:8210/health || exit 1
+
+CMD ["uvicorn", "api.main:app", "--host", "0.0.0.0", "--port", "8210"]

services/bot_digitador/api/main.py

@@ -0,0 +1,600 @@
+from __future__ import annotations
+
+import os
+from pathlib import Path
+from typing import Any, Dict, List, Optional
+
+from fastapi import Depends, FastAPI, File, Form, HTTPException, Query, UploadFile
+from fastapi.responses import Response
+from pydantic import BaseModel, Field
+
+from domain.models import JobStatus
+from infrastructure.auth import require_bearer
+from infrastructure.storage_minio import MinioSettings, MinioStorage
+from infrastructure.store import JobStore
+
+store = JobStore()
+minio = MinioStorage(MinioSettings.from_env())
+
+POBLACION_APTA_URL = os.getenv("POBLACION_APTA_URL", "http://poblacion_apta:8003").rstrip("/")
+
+app = FastAPI(title="BotDigitador Remote API", version="1.0.0")
+
+
+class RegisterWorkerBody(BaseModel):
+    worker_id: str
+    hostname: str = ""
+    max_concurrent: int = 1
+
+
+class HeartbeatBody(BaseModel):
+    worker_id: str
+
+
+class EventBody(BaseModel):
+    event_type: str
+    message: str
+    radicado: Optional[str] = None
+    nit: Optional[str] = None
+    index: Optional[int] = None
+    total: Optional[int] = None
+    metadata: Dict[str, Any] = Field(default_factory=dict)
+    timestamp: Optional[str] = None
+
+
+class PatchJobBody(BaseModel):
+    status: Optional[str] = None
+    result: Optional[Dict[str, Any]] = None
+    error: Optional[str] = None
+
+
+def _media_type_for_key(key: str) -> str:
+    lower = key.lower()
+    if lower.endswith(".csv"):
+        return "text/csv; charset=utf-8"
+    if lower.endswith(".log") or lower.endswith(".txt"):
+        return "text/plain; charset=utf-8"
+    if lower.endswith(".json"):
+        return "application/json"
+    return "application/octet-stream"
+
+
+@app.get("/health")
+def health() -> dict:
+    return {"status": "ok", "service": "bot-digitador"}
+
+
+@app.post("/workers/register")
+def register_worker(body: RegisterWorkerBody, _: str = Depends(require_bearer)) -> dict:
+    worker = store.register_worker(
+        body.worker_id,
+        hostname=body.hostname,
+        max_concurrent=body.max_concurrent,
+    )
+    return worker.to_dict()
+
+
+@app.post("/workers/heartbeat")
+def heartbeat(body: HeartbeatBody, _: str = Depends(require_bearer)) -> dict:
+    worker = store.heartbeat(body.worker_id)
+    if not worker:
+        raise HTTPException(status_code=404, detail="Worker no registrado")
+    return worker.to_dict()
+
+
+@app.get("/workers")
+def list_workers(_: str = Depends(require_bearer)) -> List[dict]:
+    return [w.to_dict() for w in store.list_workers()]
+
+
+@app.get("/workers/me/pending")
+def claim_pending(
+    worker_id: str = Query(...),
+    _: str = Depends(require_bearer),
+) -> dict:
+    job = store.claim_pending(worker_id)
+    if not job:
+        return {"job": None}
+    return {"job": job.to_dict()}
+
+
+@app.post("/jobs")
+async def create_job(
+    job_id: str = Form(...),
+    worker_id: str = Form(""),
+    worker_ids: str = Form(""),
+    title: str = Form(""),
+    file: Optional[UploadFile] = File(None),
+    poblacion_apta_job_id: str = Form(""),
+    subset: str = Form("aptos"),
+    _: str = Depends(require_bearer),
+) -> dict:
+    """Crea una corrida de digitación.
+
+    Exactamente uno de: file (CSV) o poblacion_apta_job_id.
+    - 1 worker → flujo clásico.
+    - N workers → padre + N shards CSV, mismo reports_prefix del padre.
+    """
+    import json as _json
+
+    import httpx
+
+    from domain.csv_split import parse_digitador_rows, rows_to_csv_bytes, split_rows
+    from infrastructure.storage_minio import reports_prefix as mk_prefix
+
+    parent_job_id = job_id.strip()
+    if not parent_job_id:
+        raise HTTPException(status_code=400, detail="job_id (Job Desktop) obligatorio")
+    if not store.parent_job_exists(parent_job_id):
+        raise HTTPException(status_code=404, detail="Trabajo (job) no encontrado")
+
+    pa_id = (poblacion_apta_job_id or "").strip()
+    data: bytes = b""
+    original = "input.csv"
+    content_type = "text/csv"
+    source_type = "upload"
+
+    file_bytes: bytes = b""
+    if file is not None:
+        file_bytes = await file.read()
+
+    if pa_id and file_bytes:
+        raise HTTPException(
+            status_code=400,
+            detail="Proporcione exactamente uno de: file o poblacion_apta_job_id",
+        )
+
+    if pa_id:
+        source_type = "poblacion_apta"
+        subset_norm = (subset or "aptos").strip().lower()
+        if subset_norm != "aptos":
+            raise HTTPException(
+                status_code=400,
+                detail="Solo se digitan aptos (PASO_BOT). Use subset=aptos.",
+            )
+        url = f"{POBLACION_APTA_URL}/jobs/{pa_id}/export.csv"
+        try:
+            async with httpx.AsyncClient(timeout=120.0) as client:
+                resp = await client.get(url, params={"subset": subset_norm})
+        except Exception as e:
+            raise HTTPException(
+                status_code=502,
+                detail=f"No se pudo contactar poblacion_apta: {e}",
+            ) from e
+        if resp.status_code == 404:
+            raise HTTPException(status_code=404, detail="Job de población apta no encontrado")
+        if not resp.is_success:
+            raise HTTPException(
+                status_code=502,
+                detail=f"Error exportando CSV de población apta: HTTP {resp.status_code}",
+            )
+        data = resp.content
+        original = f"poblacion_apta_{subset_norm}_{pa_id[:8]}.csv"
+        content_type = "text/csv; charset=utf-8"
+    elif file_bytes:
+        data = file_bytes
+        original = (file.filename if file else None) or "input.csv"
+        content_type = (file.content_type if file else None) or "text/csv"
+        source_type = "upload"
+    else:
+        raise HTTPException(
+            status_code=400,
+            detail="Proporcione exactamente uno de: file o poblacion_apta_job_id",
+        )
+
+    if not data:
+        raise HTTPException(status_code=400, detail="CSV vacío")
+
+    title_clean = title.strip()
+
+    # Resolver workers
+    selected: List[str] = []
+    raw_ids = (worker_ids or "").strip()
+    if raw_ids:
+        try:
+            parsed = _json.loads(raw_ids)
+            if isinstance(parsed, list):
+                selected = [str(x).strip() for x in parsed if str(x).strip()]
+            elif isinstance(parsed, str) and parsed.strip():
+                selected = [parsed.strip()]
+        except _json.JSONDecodeError:
+            selected = [x.strip() for x in raw_ids.split(",") if x.strip()]
+    if not selected and worker_id.strip():
+        selected = [worker_id.strip()]
+
+    online = store.list_online_worker_ids()
+    online_set = set(online)
+    if not selected:
+        selected = list(online)
+    else:
+        missing = [w for w in selected if w not in online_set]
+        if missing:
+            raise HTTPException(
+                status_code=409,
+                detail=f"Workers no online: {', '.join(missing)}. Enciende el agent primero.",
+            )
+
+    if not selected:
+        raise HTTPException(
+            status_code=409,
+            detail="No hay bots online. Enciende al menos un agent.",
+        )
+
+    seen: set[str] = set()
+    workers_unique: List[str] = []
+    for w in selected:
+        if w not in seen:
+            seen.add(w)
+            workers_unique.append(w)
+
+    try:
+        rows = parse_digitador_rows(data, original)
+    except ValueError as e:
+        raise HTTPException(status_code=400, detail=str(e)) from e
+    except Exception as e:
+        raise HTTPException(status_code=400, detail=f"No se pudo leer el CSV: {e}") from e
+    if not rows:
+        raise HTTPException(status_code=400, detail="El CSV no tiene filas para digitar")
+
+    # Normalizar a CSV canónico del bot (siempre pipe + headers snake_case)
+    canonical_bytes = rows_to_csv_bytes(rows)
+
+    n_workers = min(len(workers_unique), len(rows))
+    if n_workers <= 1:
+        only = workers_unique[0]
+        job = store.create_job(
+            worker_id=only,
+            csv_key="",
+            original_filename=original,
+            title=title_clean,
+            job_id=parent_job_id,
+            result={"source_type": source_type, "total_rows": len(rows), "total": len(rows)},
+        )
+        task_id = store.create_digitador_task(
+            parent_job_id=parent_job_id,
+            digitador_job_id=job.id,
+            worker_id=only,
+            title=title_clean or f"Bot digitador ({original})",
+            original_filename=original,
+            worker_ids=[only],
+            mode="single",
+            source_type=source_type,
+        )
+        try:
+            key = minio.upload_csv(job.id, canonical_bytes, original, content_type=content_type)
+        except Exception as e:
+            store.patch_job(job.id, status=JobStatus.FAILED.value, error=f"MinIO upload failed: {e}")
+            raise HTTPException(status_code=502, detail=f"Error subiendo a MinIO: {e}") from e
+
+        job = store.update_csv_key(job.id, key)
+        assert job is not None
+        store.add_event(
+            job.id,
+            "run_started",
+            f"Job creado y asignado a {only}",
+            metadata={
+                "csv_key": key,
+                "worker_id": only,
+                "job_id": parent_job_id,
+                "task_id": task_id,
+                "mode": "single",
+                "total_rows": len(rows),
+                "source_type": source_type,
+            },
+        )
+        return job.to_dict()
+
+    workers_unique = workers_unique[:n_workers]
+    chunks = split_rows(rows, n_workers)
+    n = len(chunks)
+    workers_unique = workers_unique[:n]
+
+    parent = store.create_job(
+        worker_id=None,
+        csv_key="",
+        original_filename=original,
+        title=title_clean or f"Digitación distribuida ({original})",
+        job_id=parent_job_id,
+        shard_count=n,
+        status=JobStatus.PENDING.value,
+        result={
+            "mode": "distributed",
+            "shard_ids": [],
+            "total_rows": len(rows),
+            "total": len(rows),
+            "processed": 0,
+            "ok": 0,
+            "error": 0,
+            "timeout": 0,
+            "existentes": 0,
+            "source_type": source_type,
+        },
+    )
+    prefix = parent.reports_prefix or mk_prefix(parent.id)
+
+    task_id = store.create_digitador_task(
+        parent_job_id=parent_job_id,
+        digitador_job_id=parent.id,
+        worker_id=None,
+        title=title_clean or f"Bot digitador ×{n} ({original})",
+        original_filename=original,
+        worker_ids=workers_unique,
+        mode="distributed",
+        source_type=source_type,
+    )
+
+    try:
+        full_key = minio.upload_csv(parent.id, canonical_bytes, original, content_type=content_type)
+    except Exception as e:
+        store.patch_job(parent.id, status=JobStatus.FAILED.value, error=f"MinIO upload failed: {e}")
+        raise HTTPException(status_code=502, detail=f"Error subiendo a MinIO: {e}") from e
+
+    store.update_csv_key(parent.id, full_key)
+
+    shard_ids: List[str] = []
+    for i, (chunk, wid) in enumerate(zip(chunks, workers_unique)):
+        shard_bytes = rows_to_csv_bytes(chunk)
+        shard = store.create_job(
+            worker_id=wid,
+            csv_key="",
+            original_filename=f"shard_{i + 1}_of_{n}.csv",
+            title=f"{title_clean or 'Digitador'} · shard {i + 1}/{n}",
+            job_id=parent_job_id,
+            task_id=task_id,
+            reports_prefix_override=prefix,
+            parent_digitador_job_id=parent.id,
+            shard_index=i,
+            shard_count=n,
+            result={"total": len(chunk), "processed": 0, "source_type": source_type},
+        )
+        try:
+            shard_key = minio.upload_csv(
+                shard.id,
+                shard_bytes,
+                f"shard_{i + 1}_of_{n}.csv",
+                content_type="text/csv; charset=utf-8",
+            )
+        except Exception as e:
+            store.patch_job(
+                parent.id,
+                status=JobStatus.FAILED.value,
+                error=f"MinIO shard upload failed: {e}",
+            )
+            raise HTTPException(
+                status_code=502, detail=f"Error subiendo shard a MinIO: {e}"
+            ) from e
+        store.update_csv_key(shard.id, shard_key)
+        shard_ids.append(shard.id)
+        store.add_event(
+            parent.id,
+            "info",
+            f"Shard {i + 1}/{n} → {wid} ({len(chunk)} filas)",
+            metadata={"shard_id": shard.id, "worker_id": wid, "rows": len(chunk)},
+        )
+
+    store.patch_job(
+        parent.id,
+        result={
+            "mode": "distributed",
+            "shard_ids": shard_ids,
+            "total_rows": len(rows),
+            "total": len(rows),
+            "processed": 0,
+            "ok": 0,
+            "error": 0,
+            "timeout": 0,
+            "existentes": 0,
+            "reports_prefix": prefix,
+            "worker_ids": workers_unique,
+            "source_type": source_type,
+        },
+    )
+    store.add_event(
+        parent.id,
+        "run_started",
+        f"Corrida distribuida: {n} bots, {len(rows)} filas",
+        metadata={
+            "csv_key": full_key,
+            "worker_ids": workers_unique,
+            "job_id": parent_job_id,
+            "task_id": task_id,
+            "mode": "distributed",
+            "shard_ids": shard_ids,
+            "total_rows": len(rows),
+            "source_type": source_type,
+        },
+    )
+    job = store.get_job(parent.id)
+    assert job is not None
+    return job.to_dict()
+
+
+@app.post("/internal/reclaim")
+def reclaim_jobs(_: str = Depends(require_bearer)) -> dict:
+    n = store.reclaim_stale_running_jobs()
+    return {"reclaimed": n}
+
+
+@app.get("/jobs")
+def list_jobs(
+    limit: int = Query(100, ge=1, le=500),
+    job_id: Optional[str] = Query(None, description="Filtrar por Job Desktop"),
+    _: str = Depends(require_bearer),
+) -> List[dict]:
+    return [j.to_dict() for j in store.list_jobs(limit=limit, job_id=job_id)]
+
+
+@app.get("/jobs/{digitador_job_id}")
+def get_job(digitador_job_id: str, _: str = Depends(require_bearer)) -> dict:
+    job = store.get_job(digitador_job_id)
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    if job.is_distributed_parent:
+        refreshed = store.refresh_parent_from_shards(digitador_job_id)
+        if refreshed:
+            job = refreshed
+    payload = job.to_dict()
+    if job.is_distributed_parent:
+        payload["shards"] = [s.to_dict() for s in store.list_shards(job.id)]
+    return payload
+
+
+@app.get("/jobs/{digitador_job_id}/files")
+def list_job_files(digitador_job_id: str, _: str = Depends(require_bearer)) -> dict:
+    job = store.get_job(digitador_job_id)
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    try:
+        return minio.list_report_files(digitador_job_id, prefix=job.reports_prefix)
+    except Exception as e:
+        raise HTTPException(status_code=502, detail=f"Error listando MinIO: {e}") from e
+
+
+@app.get("/jobs/{digitador_job_id}/object")
+def get_job_object(
+    digitador_job_id: str,
+    key: str = Query(..., min_length=1, description="Object key bajo reports_prefix"),
+    _: str = Depends(require_bearer),
+):
+    """Stream de un reporte MinIO (csv/log/json bajo el prefijo del job)."""
+    job = store.get_job(digitador_job_id)
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+
+    prefix = (job.reports_prefix or f"digitados/{digitador_job_id}/").lstrip("/")
+    if not prefix.endswith("/"):
+        prefix += "/"
+    normalized = key.lstrip("/")
+    if not normalized.startswith(prefix):
+        raise HTTPException(status_code=403, detail="Key fuera del prefijo del job")
+    lower = normalized.lower()
+    if not (lower.endswith(".csv") or lower.endswith(".log") or lower.endswith(".json")):
+        raise HTTPException(status_code=400, detail="Solo se permiten objetos csv/log/json")
+    if ".." in normalized.split("/"):
+        raise HTTPException(status_code=400, detail="Key no permitida")
+
+    try:
+        data = minio.get_bytes(normalized)
+    except Exception as e:
+        raise HTTPException(status_code=502, detail=f"Error leyendo MinIO: {e}") from e
+
+    filename = Path(normalized).name
+    return Response(
+        content=data,
+        media_type=_media_type_for_key(normalized),
+        headers={"Content-Disposition": f'inline; filename="{filename}"'},
+    )
+
+
+@app.get("/jobs/{digitador_job_id}/events")
+def get_events(
+    digitador_job_id: str,
+    since: int = Query(0, ge=0),
+    _: str = Depends(require_bearer),
+) -> dict:
+    if not store.get_job(digitador_job_id):
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    events = store.list_events(digitador_job_id, since_id=since)
+    return {"events": [e.to_dict() for e in events]}
+
+
+@app.post("/jobs/{digitador_job_id}/cancel")
+def cancel_job(digitador_job_id: str, _: str = Depends(require_bearer)) -> dict:
+    job = store.request_cancel(digitador_job_id)
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    store.add_event(digitador_job_id, "warning", "Cancelación solicitada por el usuario")
+    return job.to_dict()
+
+
+@app.get("/jobs/{digitador_job_id}/file")
+def get_job_file(digitador_job_id: str, _: str = Depends(require_bearer)):
+    job = store.get_job(digitador_job_id)
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    if not job.csv_key:
+        raise HTTPException(status_code=404, detail="CSV no disponible")
+
+    try:
+        data = minio.get_bytes(job.csv_key)
+    except Exception as e:
+        raise HTTPException(status_code=502, detail=f"Error leyendo MinIO: {e}") from e
+
+    filename = Path(job.csv_key).name
+    return Response(
+        content=data,
+        media_type="text/csv; charset=utf-8",
+        headers={"Content-Disposition": f'attachment; filename="{filename}"'},
+    )
+
+
+@app.get("/jobs/{digitador_job_id}/file-url")
+def get_job_file_url(digitador_job_id: str, _: str = Depends(require_bearer)) -> dict:
+    job = store.get_job(digitador_job_id)
+    if not job or not job.csv_key:
+        raise HTTPException(status_code=404, detail="Job/archivo no encontrado")
+    try:
+        url = minio.presigned_get(job.csv_key)
+    except Exception as e:
+        raise HTTPException(status_code=502, detail=str(e)) from e
+    return {"url": url, "csv_key": job.csv_key}
+
+
+@app.post("/jobs/{digitador_job_id}/events")
+def post_event(
+    digitador_job_id: str,
+    body: EventBody,
+    _: str = Depends(require_bearer),
+) -> dict:
+    job = store.get_job(digitador_job_id)
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    event = store.add_event(
+        digitador_job_id,
+        body.event_type,
+        body.message,
+        radicado=body.radicado,
+        nit=body.nit,
+        index=body.index,
+        total=body.total,
+        metadata=body.metadata,
+        timestamp=body.timestamp,
+    )
+    if body.index is not None and body.total is not None:
+        result = dict(job.result or {})
+        result["processed"] = body.index
+        result["total"] = body.total
+        if body.nit:
+            result["nit_actual"] = body.nit
+        store.patch_job(digitador_job_id, result=result)
+    return event.to_dict()
+
+
+@app.patch("/jobs/{digitador_job_id}")
+def patch_job(
+    digitador_job_id: str,
+    body: PatchJobBody,
+    _: str = Depends(require_bearer),
+) -> dict:
+    if body.status and body.status not in {s.value for s in JobStatus}:
+        raise HTTPException(status_code=400, detail="status inválido")
+    job = store.patch_job(
+        digitador_job_id,
+        status=body.status,
+        result=body.result,
+        error=body.error,
+    )
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    return job.to_dict()
+
+
+@app.get("/jobs/{digitador_job_id}/cancel-flag")
+def cancel_flag(digitador_job_id: str, _: str = Depends(require_bearer)) -> dict:
+    job = store.get_job(digitador_job_id)
+    if not job:
+        raise HTTPException(status_code=404, detail="Job no encontrado")
+    return {
+        "cancel_requested": store.is_cancel_requested(digitador_job_id),
+        "status": job.status,
+    }

services/bot_digitador/domain/csv_split.py

@@ -0,0 +1,157 @@
+"""Parseo y partición de CSV de filas a digitar (bot digitador)."""
+from __future__ import annotations
+
+import csv
+import io
+from typing import Any
+
+REQUIRED_BOT_COLUMNS = (
+    "radicado",
+    "documento",
+    "ingreso",
+    "total_cargos",
+    "autorizacion",
+    "tipo_documento",
+    "tipo_servicio",
+    "copago",
+    "cuota_moderadora",
+)
+
+IT_TO_BOT = {
+    "RADICADO": "radicado",
+    "NUMERO": "documento",
+    "FECHA": "ingreso",
+    "BRUTO": "total_cargos",
+    "AUTORIZACION": "autorizacion",
+    "TIPO_ID": "tipo_documento",
+    "SERVICIO": "tipo_servicio",
+    "COPAGO": "copago",
+    "MODERADORA": "cuota_moderadora",
+}
+
+
+def _normalize_header(name: str) -> str:
+    raw = str(name or "").strip()
+    if not raw:
+        return ""
+    upper = raw.upper()
+    if upper in IT_TO_BOT:
+        return IT_TO_BOT[upper]
+    lower = raw.lower().strip()
+    # Aceptar snake_case ya normalizado
+    if lower in REQUIRED_BOT_COLUMNS:
+        return lower
+    # Alias comunes
+    aliases = {
+        "número": "documento",
+        "numero": "documento",
+        "fecha": "ingreso",
+        "bruto": "total_cargos",
+        "tipo_id": "tipo_documento",
+        "servicio": "tipo_servicio",
+        "moderadora": "cuota_moderadora",
+    }
+    return aliases.get(lower, lower)
+
+
+def _decode_text(data: bytes) -> str:
+    for enc in ("utf-8-sig", "utf-8", "latin-1", "cp1252"):
+        try:
+            return data.decode(enc)
+        except UnicodeDecodeError:
+            continue
+    return data.decode("utf-8", errors="replace")
+
+
+def _detect_delimiter(sample: str) -> str:
+    try:
+        dialect = csv.Sniffer().sniff(sample, delimiters="|,;\t")
+        return dialect.delimiter
+    except csv.Error:
+        return "|" if sample.count("|") >= sample.count(",") else ","
+
+
+def _validate_columns(headers: list[str]) -> None:
+    present = set(headers)
+    missing = [c for c in REQUIRED_BOT_COLUMNS if c not in present]
+    if missing:
+        raise ValueError(f"Columnas faltantes: {', '.join(missing)}")
+
+
+def parse_digitador_rows(data: bytes, filename: str = "input.csv") -> list[dict[str, str]]:
+    """Devuelve filas normalizadas con headers del bot (snake_case)."""
+    _ = filename  # reservado por firma / diagnóstico futuro
+    text = _decode_text(data)
+    if not text.strip():
+        return []
+
+    delim = _detect_delimiter(text[:4096])
+    reader = csv.DictReader(io.StringIO(text), delimiter=delim)
+    if not reader.fieldnames:
+        raise ValueError("CSV sin encabezados")
+
+    rename_map: dict[str, str] = {}
+    normalized_headers: list[str] = []
+    for raw in reader.fieldnames:
+        norm = _normalize_header(raw)
+        rename_map[raw] = norm
+        if norm and norm not in normalized_headers:
+            normalized_headers.append(norm)
+
+    _validate_columns(normalized_headers)
+
+    out: list[dict[str, str]] = []
+    for row in reader:
+        if not row:
+            continue
+        mapped: dict[str, str] = {c: "" for c in REQUIRED_BOT_COLUMNS}
+        empty = True
+        for raw_key, value in row.items():
+            if raw_key is None:
+                continue
+            col = rename_map.get(raw_key) or _normalize_header(raw_key)
+            if col not in REQUIRED_BOT_COLUMNS:
+                continue
+            cell = "" if value is None else str(value).strip()
+            mapped[col] = cell
+            if cell:
+                empty = False
+        if empty:
+            continue
+        # Filtrar filas que parecen ser un segundo header
+        if mapped.get("radicado", "").lower() in ("radicado", "radicado"):
+            if mapped.get("documento", "").lower() in ("documento", "numero", "número", ""):
+                continue
+        out.append(mapped)
+    return out
+
+
+def split_rows(rows: list[dict[str, Any]], n: int) -> list[list[dict[str, Any]]]:
+    """Parte filas en n trozos lo más equilibrados posible (chunks contiguos)."""
+    if n <= 1 or not rows:
+        return [rows]
+    n = min(n, len(rows))
+    size, rem = divmod(len(rows), n)
+    chunks: list[list[dict[str, Any]]] = []
+    idx = 0
+    for i in range(n):
+        take = size + (1 if i < rem else 0)
+        chunks.append(rows[idx : idx + take])
+        idx += take
+    return [c for c in chunks if c]
+
+
+def rows_to_csv_bytes(rows: list[dict[str, Any]]) -> bytes:
+    """Serializa filas a CSV pipe-delimited con headers del bot (utf-8-sig)."""
+    buf = io.StringIO()
+    writer = csv.DictWriter(
+        buf,
+        fieldnames=list(REQUIRED_BOT_COLUMNS),
+        delimiter="|",
+        lineterminator="\n",
+        extrasaction="ignore",
+    )
+    writer.writeheader()
+    for row in rows:
+        writer.writerow({c: str(row.get(c, "") or "") for c in REQUIRED_BOT_COLUMNS})
+    return buf.getvalue().encode("utf-8-sig")

services/bot_digitador/domain/models.py

@@ -0,0 +1,91 @@
+from __future__ import annotations
+
+from dataclasses import asdict, dataclass, field
+from datetime import datetime, timezone
+from enum import Enum
+from typing import Any, Dict, Optional
+
+
+def utc_now() -> str:
+    return datetime.now(timezone.utc).astimezone().isoformat(timespec="seconds")
+
+
+class WorkerStatus(str, Enum):
+    ONLINE = "online"
+    OFFLINE = "offline"
+    BUSY = "busy"
+
+
+class JobStatus(str, Enum):
+    PENDING = "pending"
+    RUNNING = "running"
+    COMPLETED = "completed"
+    FAILED = "failed"
+    CANCELLED = "cancelled"
+
+
+@dataclass
+class Worker:
+    id: str
+    hostname: str = ""
+    max_concurrent: int = 1
+    status: str = WorkerStatus.ONLINE.value
+    last_heartbeat: str = field(default_factory=utc_now)
+    current_job_id: Optional[str] = None
+
+    def to_dict(self) -> dict:
+        return asdict(self)
+
+
+@dataclass
+class Job:
+    id: str
+    worker_id: Optional[str]
+    status: str = JobStatus.PENDING.value
+    title: str = ""
+    csv_key: str = ""
+    reports_prefix: str = ""
+    original_filename: str = ""
+    cancel_requested: bool = False
+    result: Dict[str, Any] = field(default_factory=dict)
+    error: Optional[str] = None
+    created_at: str = field(default_factory=utc_now)
+    started_at: Optional[str] = None
+    finished_at: Optional[str] = None
+    job_id: Optional[str] = None
+    task_id: Optional[str] = None
+    parent_digitador_job_id: Optional[str] = None
+    shard_index: Optional[int] = None
+    shard_count: Optional[int] = None
+
+    def to_dict(self) -> dict:
+        return asdict(self)
+
+    @property
+    def is_distributed_parent(self) -> bool:
+        return (
+            self.parent_digitador_job_id is None
+            and self.shard_count is not None
+            and int(self.shard_count) > 1
+        )
+
+    @property
+    def is_shard(self) -> bool:
+        return bool(self.parent_digitador_job_id)
+
+
+@dataclass
+class JobEventRecord:
+    id: int
+    job_id: str
+    event_type: str
+    message: str
+    timestamp: str
+    radicado: Optional[str] = None
+    nit: Optional[str] = None
+    index: Optional[int] = None
+    total: Optional[int] = None
+    metadata: Dict[str, Any] = field(default_factory=dict)
+
+    def to_dict(self) -> dict:
+        return asdict(self)

services/bot_digitador/infrastructure/auth.py

@@ -0,0 +1,20 @@
+from __future__ import annotations
+
+import os
+from typing import Optional
+
+from fastapi import Header, HTTPException
+
+
+def get_api_token() -> str:
+    return os.getenv("BOT_API_TOKEN", "dev-token-change-me")
+
+
+def require_bearer(authorization: Optional[str] = Header(default=None)) -> str:
+    expected = get_api_token()
+    if not authorization or not authorization.lower().startswith("bearer "):
+        raise HTTPException(status_code=401, detail="Authorization Bearer requerido")
+    token = authorization.split(" ", 1)[1].strip()
+    if token != expected:
+        raise HTTPException(status_code=401, detail="Token inválido")
+    return token

services/bot_digitador/infrastructure/database.py

@@ -0,0 +1,164 @@
+"""Persistencia Postgres del servicio bot-digitador (DB compartida IPSExtractor)."""
+from __future__ import annotations
+
+import os
+import uuid
+from contextlib import contextmanager
+from datetime import datetime, timezone
+from typing import Any, Iterator
+
+from sqlalchemy import (
+    JSON,
+    Boolean,
+    DateTime,
+    ForeignKey,
+    Integer,
+    String,
+    Text,
+    create_engine,
+    text,
+)
+from sqlalchemy.orm import DeclarativeBase, Mapped, Session, mapped_column, sessionmaker
+
+DATABASE_URL = os.getenv(
+    "DATABASE_URL",
+    "postgresql+psycopg://ipsextractor:ipsextractor@database:5432/ipsextractor",
+)
+
+engine = create_engine(DATABASE_URL, pool_pre_ping=True)
+SessionLocal = sessionmaker(bind=engine, autoflush=False, autocommit=False)
+
+
+class Base(DeclarativeBase):
+    pass
+
+
+def utcnow() -> datetime:
+    return datetime.now(timezone.utc)
+
+
+def new_id() -> str:
+    return str(uuid.uuid4())
+
+
+def _iso(dt: datetime | None) -> str | None:
+    if dt is None:
+        return None
+    if dt.tzinfo is None:
+        dt = dt.replace(tzinfo=timezone.utc)
+    return dt.astimezone().isoformat(timespec="seconds")
+
+
+class Job(Base):
+    """Contenedor Desktop (lectura / FK)."""
+
+    __tablename__ = "jobs"
+
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=new_id)
+    status: Mapped[str] = mapped_column(String(32), default="pending", index=True)
+    title: Mapped[str | None] = mapped_column(Text, nullable=True)
+    job_metadata: Mapped[dict[str, Any]] = mapped_column("metadata", JSON, default=dict)
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
+    updated_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
+
+
+class Task(Base):
+    __tablename__ = "tasks"
+
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=new_id)
+    job_id: Mapped[str | None] = mapped_column(
+        String(36), ForeignKey("jobs.id"), nullable=True, index=True
+    )
+    type: Mapped[str] = mapped_column(String(64), index=True)
+    status: Mapped[str] = mapped_column(String(32), default="pending", index=True)
+    title: Mapped[str | None] = mapped_column(Text, nullable=True)
+    request: Mapped[dict[str, Any]] = mapped_column(JSON, default=dict)
+    result: Mapped[dict[str, Any]] = mapped_column(JSON, default=dict)
+    error: Mapped[str | None] = mapped_column(Text, nullable=True)
+    owner_id: Mapped[str | None] = mapped_column(String(36), nullable=True, index=True)
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
+    updated_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
+    started_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+    finished_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+
+
+class DigitadorWorkerRow(Base):
+    __tablename__ = "digitador_workers"
+
+    id: Mapped[str] = mapped_column(String(36), primary_key=True)
+    hostname: Mapped[str] = mapped_column(Text, default="")
+    max_concurrent: Mapped[int] = mapped_column(Integer, default=1)
+    status: Mapped[str] = mapped_column(String(32), index=True)
+    last_heartbeat: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
+    current_job_id: Mapped[str | None] = mapped_column(String(36), nullable=True)
+
+
+class DigitadorJobRow(Base):
+    __tablename__ = "digitador_jobs"
+
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=new_id)
+    job_id: Mapped[str | None] = mapped_column(
+        String(36), ForeignKey("jobs.id"), nullable=True, index=True
+    )
+    task_id: Mapped[str | None] = mapped_column(
+        String(36), ForeignKey("tasks.id"), nullable=True, index=True
+    )
+    worker_id: Mapped[str | None] = mapped_column(String(36), nullable=True, index=True)
+    parent_digitador_job_id: Mapped[str | None] = mapped_column(
+        String(36), ForeignKey("digitador_jobs.id"), nullable=True, index=True
+    )
+    shard_index: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    shard_count: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    status: Mapped[str] = mapped_column(String(32), default="pending", index=True)
+    title: Mapped[str | None] = mapped_column(Text, nullable=True)
+    csv_key: Mapped[str] = mapped_column(Text, default="")
+    reports_prefix: Mapped[str] = mapped_column(Text, default="")
+    original_filename: Mapped[str] = mapped_column(Text, default="")
+    cancel_requested: Mapped[bool] = mapped_column(Boolean, default=False)
+    result: Mapped[dict[str, Any]] = mapped_column(JSON, default=dict)
+    error: Mapped[str | None] = mapped_column(Text, nullable=True)
+    created_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
+    started_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+    finished_at: Mapped[datetime | None] = mapped_column(DateTime(timezone=True), nullable=True)
+
+
+class DigitadorJobEventRow(Base):
+    __tablename__ = "digitador_job_events"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    digitador_job_id: Mapped[str] = mapped_column(
+        String(36), ForeignKey("digitador_jobs.id"), index=True
+    )
+    event_type: Mapped[str] = mapped_column(String(64))
+    message: Mapped[str] = mapped_column(Text)
+    timestamp: Mapped[datetime] = mapped_column(DateTime(timezone=True), default=utcnow)
+    radicado: Mapped[str | None] = mapped_column(String(255), nullable=True)
+    nit: Mapped[str | None] = mapped_column(String(64), nullable=True)
+    idx: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    total: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    metadata_json: Mapped[dict[str, Any]] = mapped_column("metadata", JSON, default=dict)
+
+
+def init_db() -> None:
+    if DATABASE_URL.startswith("postgresql"):
+        with engine.connect().execution_options(isolation_level="AUTOCOMMIT") as connection:
+            connection.execute(text("SELECT pg_advisory_lock(74512005)"))
+            try:
+                Base.metadata.create_all(bind=connection)
+            finally:
+                connection.execute(text("SELECT pg_advisory_unlock(74512005)"))
+        return
+    Base.metadata.create_all(bind=engine)
+
+
+@contextmanager
+def session_scope() -> Iterator[Session]:
+    session = SessionLocal()
+    try:
+        yield session
+        session.commit()
+    except Exception:
+        session.rollback()
+        raise
+    finally:
+        session.close()

services/bot_digitador/infrastructure/storage_minio.py

@@ -0,0 +1,230 @@
+from __future__ import annotations
+
+import json
+import os
+from dataclasses import dataclass
+from pathlib import Path
+from typing import BinaryIO, List, Optional
+
+
+@dataclass(frozen=True)
+class MinioSettings:
+    endpoint: str = "https://s3.builker.xyz"
+    bucket: str = "pys-cuentas-medicas"
+    access_key: str = ""
+    secret_key: str = ""
+    region: str = "us-east-1"
+    anonymous: bool = True
+
+    @classmethod
+    def from_env(cls) -> "MinioSettings":
+        access_key = (os.getenv("MINIO_ACCESS_KEY") or "").strip()
+        secret_key = (os.getenv("MINIO_SECRET_KEY") or "").strip()
+        anon_flag = (os.getenv("MINIO_ANONYMOUS") or "").strip().lower()
+        if anon_flag in ("0", "false", "no"):
+            anonymous = False
+        elif anon_flag in ("1", "true", "yes"):
+            anonymous = True
+        else:
+            anonymous = not (access_key and secret_key)
+        return cls(
+            endpoint=os.getenv("MINIO_ENDPOINT", "https://s3.builker.xyz").rstrip("/"),
+            bucket=os.getenv("MINIO_BUCKET", "pys-cuentas-medicas"),
+            access_key=access_key,
+            secret_key=secret_key,
+            region=os.getenv("MINIO_REGION", "us-east-1"),
+            anonymous=anonymous,
+        )
+
+
+def csv_key(job_id: str, filename: str = "input.csv") -> str:
+    return f"uploads/{job_id}/{filename}"
+
+
+def reports_prefix(job_id: str) -> str:
+    return f"digitados/{job_id}/"
+
+
+def meta_key(job_id: str) -> str:
+    return f"uploads/{job_id}/meta.json"
+
+
+def manifest_key(job_id: str) -> str:
+    return f"digitados/{job_id}/_manifest.json"
+
+
+_REPORT_NAMES = {
+    "resultados.csv",
+    "omitidos.csv",
+    "ejecucion_bot.log",
+}
+
+
+def _s3_client(settings: MinioSettings):
+    import boto3
+    from botocore import UNSIGNED
+    from botocore.client import Config
+
+    s3_cfg = {
+        "addressing_style": "path",
+        "request_checksum_calculation": "when_required",
+        "response_checksum_validation": "when_required",
+    }
+    if settings.anonymous:
+        return boto3.client(
+            "s3",
+            endpoint_url=settings.endpoint,
+            region_name=settings.region,
+            config=Config(signature_version=UNSIGNED, s3=s3_cfg),
+        )
+    return boto3.client(
+        "s3",
+        endpoint_url=settings.endpoint,
+        aws_access_key_id=settings.access_key,
+        aws_secret_access_key=settings.secret_key,
+        region_name=settings.region,
+        config=Config(signature_version="s3v4", s3=s3_cfg),
+    )
+
+
+class MinioStorage:
+    """Cliente S3-compatible para MinIO (uploads/digitados por job_id)."""
+
+    def __init__(self, settings: Optional[MinioSettings] = None):
+        self.settings = settings or MinioSettings.from_env()
+        self._client = None
+
+    @property
+    def client(self):
+        if self._client is None:
+            self._client = _s3_client(self.settings)
+        return self._client
+
+    @property
+    def bucket(self) -> str:
+        return self.settings.bucket
+
+    def put_bytes(self, key: str, data: bytes, content_type: str = "application/octet-stream") -> str:
+        self.client.put_object(
+            Bucket=self.bucket,
+            Key=key,
+            Body=data,
+            ContentType=content_type,
+        )
+        return key
+
+    def put_file(self, key: str, path: str | Path, content_type: str = "application/octet-stream") -> str:
+        path = Path(path)
+        return self.put_bytes(key, path.read_bytes(), content_type=content_type)
+
+    def put_stream(self, key: str, stream: BinaryIO, content_type: str = "application/octet-stream") -> str:
+        data = stream.read()
+        return self.put_bytes(key, data, content_type=content_type)
+
+    def get_bytes(self, key: str) -> bytes:
+        obj = self.client.get_object(Bucket=self.bucket, Key=key)
+        return obj["Body"].read()
+
+    def download_file(self, key: str, dest: str | Path) -> Path:
+        dest = Path(dest)
+        dest.parent.mkdir(parents=True, exist_ok=True)
+        self.client.download_file(self.bucket, key, str(dest))
+        return dest
+
+    def exists(self, key: str) -> bool:
+        try:
+            self.client.head_object(Bucket=self.bucket, Key=key)
+            return True
+        except Exception:
+            return False
+
+    def list_keys(self, prefix: str) -> List[str]:
+        keys: List[str] = []
+        token = None
+        while True:
+            kwargs = {"Bucket": self.bucket, "Prefix": prefix}
+            if token:
+                kwargs["ContinuationToken"] = token
+            resp = self.client.list_objects_v2(**kwargs)
+            for item in resp.get("Contents") or []:
+                keys.append(item["Key"])
+            if not resp.get("IsTruncated"):
+                break
+            token = resp.get("NextContinuationToken")
+        return keys
+
+    def presigned_get(self, key: str, expires_in: int = 3600) -> str:
+        if self.settings.anonymous:
+            return f"{self.settings.endpoint}/{self.bucket}/{key}"
+        return self.client.generate_presigned_url(
+            "get_object",
+            Params={"Bucket": self.bucket, "Key": key},
+            ExpiresIn=expires_in,
+        )
+
+    def upload_csv(
+        self,
+        job_id: str,
+        data: bytes,
+        original_filename: str,
+        content_type: str = "text/csv",
+    ) -> str:
+        ext = Path(original_filename).suffix.lower() or ".csv"
+        canonical = "input.csv" if ext in (".csv", ".txt") else "input.csv"
+        key = csv_key(job_id, canonical)
+        self.put_bytes(key, data, content_type=content_type)
+        meta = {
+            "original_filename": original_filename,
+            "size": len(data),
+            "canonical": canonical,
+        }
+        self.put_bytes(meta_key(job_id), json.dumps(meta).encode("utf-8"), "application/json")
+        return key
+
+    def write_manifest(self, job_id: str, keys: List[str]) -> str:
+        key = manifest_key(job_id)
+        payload = {"job_id": job_id, "files": keys, "count": len(keys)}
+        self.put_bytes(key, json.dumps(payload, ensure_ascii=False).encode("utf-8"), "application/json")
+        return key
+
+    def list_report_files(self, job_id: str, prefix: str | None = None) -> dict:
+        """Lista reportes de digitación (RESULTADOS/OMITIDOS/log + list_keys).
+
+        Fusiona manifest con listado del prefijo (cubriendo shards distribuidos).
+        """
+        prefix = prefix or reports_prefix(job_id)
+        by_key: dict[str, dict] = {}
+        try:
+            raw = self.get_bytes(manifest_key(job_id))
+            payload = json.loads(raw.decode("utf-8"))
+            for key in payload.get("files") or []:
+                entry = self._file_entry(str(key))
+                if entry is not None:
+                    by_key[entry["key"]] = entry
+        except Exception:
+            pass
+
+        for entry in self._list_report_entries(prefix):
+            by_key[entry["key"]] = entry
+
+        files = list(by_key.values())
+        return {"job_id": job_id, "prefix": prefix, "files": files, "count": len(files)}
+
+    def _list_report_entries(self, prefix: str) -> List[dict]:
+        files: List[dict] = []
+        for key in self.list_keys(prefix):
+            entry = self._file_entry(key)
+            if entry is not None:
+                files.append(entry)
+        return files
+
+    @staticmethod
+    def _file_entry(key: str) -> dict | None:
+        name = Path(key).name
+        lower = name.lower()
+        if lower.startswith("_manifest") and lower.endswith(".json"):
+            return None
+        # RESULTADOS.csv, OMITIDOS.csv, ejecucion_bot.log (+ variantes _shard_N)
+        if lower in _REPORT_NAMES or lower.endswith(".csv") or lower.endswith(".log"):
+            return {"key": key, "name": name}
+        return None

services/bot_digitador/infrastructure/store.py

@@ -0,0 +1,833 @@
+"""Repositorio sobre PostgreSQL (tablas digitador_* + sync de tasks)."""
+from __future__ import annotations
+
+from datetime import datetime, timedelta, timezone
+from typing import Any, Dict, List, Optional
+
+from sqlalchemy import select
+
+from domain.models import (
+    Job,
+    JobEventRecord,
+    JobStatus,
+    Worker,
+    WorkerStatus,
+    utc_now,
+)
+from infrastructure.database import (
+    DigitadorJobEventRow,
+    DigitadorJobRow,
+    DigitadorWorkerRow,
+    Task,
+    _iso,
+    init_db,
+    session_scope,
+    utcnow,
+)
+from infrastructure.storage_minio import reports_prefix
+
+HEARTBEAT_OFFLINE_SECONDS = 45
+
+
+def _parse_iso(value: str | None) -> datetime | None:
+    if not value:
+        return None
+    dt = datetime.fromisoformat(value)
+    if dt.tzinfo is None:
+        dt = dt.replace(tzinfo=timezone.utc)
+    return dt
+
+
+class JobStore:
+    """API de persistencia compatible con el patrón bot_descarga."""
+
+    def __init__(self, db_path: str | None = None):
+        # db_path ignorado (compat); usa DATABASE_URL.
+        init_db()
+
+    def _row_worker(self, row: DigitadorWorkerRow) -> Worker:
+        worker = Worker(
+            id=row.id,
+            hostname=row.hostname or "",
+            max_concurrent=int(row.max_concurrent or 1),
+            status=row.status,
+            last_heartbeat=_iso(row.last_heartbeat) or utc_now(),
+            current_job_id=row.current_job_id,
+        )
+        return self._refresh_worker_status(worker)
+
+    def _row_job(self, row: DigitadorJobRow) -> Job:
+        return Job(
+            id=row.id,
+            worker_id=row.worker_id,
+            status=row.status,
+            title=row.title or "",
+            csv_key=row.csv_key or "",
+            reports_prefix=row.reports_prefix or "",
+            original_filename=row.original_filename or "",
+            cancel_requested=bool(row.cancel_requested),
+            result=dict(row.result or {}),
+            error=row.error,
+            created_at=_iso(row.created_at) or utc_now(),
+            started_at=_iso(row.started_at),
+            finished_at=_iso(row.finished_at),
+            job_id=row.job_id,
+            task_id=row.task_id,
+            parent_digitador_job_id=row.parent_digitador_job_id,
+            shard_index=row.shard_index,
+            shard_count=row.shard_count,
+        )
+
+    def _refresh_worker_status(self, worker: Worker) -> Worker:
+        try:
+            hb = datetime.fromisoformat(worker.last_heartbeat)
+            if hb.tzinfo is None:
+                hb = hb.replace(tzinfo=timezone.utc)
+            age = datetime.now(timezone.utc) - hb.astimezone(timezone.utc)
+            if age > timedelta(seconds=HEARTBEAT_OFFLINE_SECONDS):
+                worker.status = WorkerStatus.OFFLINE.value
+            elif worker.current_job_id:
+                worker.status = WorkerStatus.BUSY.value
+            else:
+                worker.status = WorkerStatus.ONLINE.value
+        except Exception:
+            worker.status = WorkerStatus.OFFLINE.value
+        return worker
+
+    def register_worker(
+        self,
+        worker_id: str,
+        hostname: str = "",
+        max_concurrent: int = 1,
+    ) -> Worker:
+        with session_scope() as session:
+            row = session.get(DigitadorWorkerRow, worker_id)
+            now = utcnow()
+            if row:
+                row.hostname = hostname
+                row.max_concurrent = max_concurrent
+                row.status = WorkerStatus.ONLINE.value
+                row.last_heartbeat = now
+                row.current_job_id = None
+            else:
+                row = DigitadorWorkerRow(
+                    id=worker_id,
+                    hostname=hostname,
+                    max_concurrent=max_concurrent,
+                    status=WorkerStatus.ONLINE.value,
+                    last_heartbeat=now,
+                    current_job_id=None,
+                )
+                session.add(row)
+            session.flush()
+            return self._row_worker(row)
+
+    def heartbeat(self, worker_id: str) -> Optional[Worker]:
+        with session_scope() as session:
+            row = session.get(DigitadorWorkerRow, worker_id)
+            if not row:
+                return None
+            row.last_heartbeat = utcnow()
+            row.status = (
+                WorkerStatus.BUSY.value if row.current_job_id else WorkerStatus.ONLINE.value
+            )
+            session.flush()
+            worker = self._row_worker(row)
+        try:
+            self.reclaim_stale_running_jobs()
+        except Exception:
+            pass
+        return worker
+
+    def get_worker(self, worker_id: str) -> Optional[Worker]:
+        with session_scope() as session:
+            row = session.get(DigitadorWorkerRow, worker_id)
+            if not row:
+                return None
+            return self._row_worker(row)
+
+    def list_workers(self) -> List[Worker]:
+        try:
+            self.reclaim_stale_running_jobs()
+        except Exception:
+            pass
+        with session_scope() as session:
+            rows = session.scalars(
+                select(DigitadorWorkerRow).order_by(DigitadorWorkerRow.id)
+            ).all()
+            return [self._row_worker(r) for r in rows]
+
+    def list_online_worker_ids(self) -> List[str]:
+        return [
+            w.id
+            for w in self.list_workers()
+            if w.status in (WorkerStatus.ONLINE.value, WorkerStatus.BUSY.value)
+        ]
+
+    def is_worker_online(self, worker_id: str) -> bool:
+        worker = self.get_worker(worker_id)
+        if not worker:
+            return False
+        return worker.status in (WorkerStatus.ONLINE.value, WorkerStatus.BUSY.value)
+
+    def update_csv_key(self, job_id: str, csv_key: str) -> Optional[Job]:
+        with session_scope() as session:
+            row = session.get(DigitadorJobRow, job_id)
+            if not row:
+                return None
+            row.csv_key = csv_key
+            session.flush()
+            return self._row_job(row)
+
+    def create_job(
+        self,
+        worker_id: Optional[str],
+        csv_key: str,
+        original_filename: str = "",
+        title: str = "",
+        *,
+        job_id: Optional[str] = None,
+        task_id: Optional[str] = None,
+        reports_prefix_override: Optional[str] = None,
+        parent_digitador_job_id: Optional[str] = None,
+        shard_index: Optional[int] = None,
+        shard_count: Optional[int] = None,
+        result: Optional[Dict[str, Any]] = None,
+        status: Optional[str] = None,
+        digitador_id: Optional[str] = None,
+    ) -> Job:
+        import uuid as uuid_mod
+
+        digitador_id = digitador_id or str(uuid_mod.uuid4())
+        prefix = reports_prefix_override or reports_prefix(digitador_id)
+        with session_scope() as session:
+            row = DigitadorJobRow(
+                id=digitador_id,
+                job_id=job_id,
+                task_id=task_id,
+                worker_id=worker_id,
+                parent_digitador_job_id=parent_digitador_job_id,
+                shard_index=shard_index,
+                shard_count=shard_count,
+                status=status or JobStatus.PENDING.value,
+                title=title or f"Digitador {original_filename}",
+                csv_key=csv_key,
+                reports_prefix=prefix,
+                original_filename=original_filename,
+                cancel_requested=False,
+                result=result or {},
+                created_at=utcnow(),
+            )
+            session.add(row)
+            session.flush()
+            return self._row_job(row)
+
+    def get_job(self, job_id: str) -> Optional[Job]:
+        with session_scope() as session:
+            row = session.get(DigitadorJobRow, job_id)
+            if not row:
+                return None
+            return self._row_job(row)
+
+    def list_jobs(
+        self,
+        limit: int = 100,
+        *,
+        job_id: Optional[str] = None,
+        parents_only: bool = True,
+    ) -> List[Job]:
+        with session_scope() as session:
+            stmt = select(DigitadorJobRow).order_by(DigitadorJobRow.created_at.desc())
+            if job_id:
+                stmt = stmt.where(DigitadorJobRow.job_id == job_id)
+            if parents_only:
+                stmt = stmt.where(DigitadorJobRow.parent_digitador_job_id.is_(None))
+            stmt = stmt.limit(limit)
+            rows = session.scalars(stmt).all()
+            return [self._row_job(r) for r in rows]
+
+    def list_shards(self, parent_id: str) -> List[Job]:
+        with session_scope() as session:
+            rows = session.scalars(
+                select(DigitadorJobRow)
+                .where(DigitadorJobRow.parent_digitador_job_id == parent_id)
+                .order_by(DigitadorJobRow.shard_index.asc())
+            ).all()
+            return [self._row_job(r) for r in rows]
+
+    def claim_pending(self, worker_id: str) -> Optional[Job]:
+        try:
+            self.reclaim_stale_running_jobs()
+        except Exception:
+            pass
+        with session_scope() as session:
+            worker = session.get(DigitadorWorkerRow, worker_id)
+            if not worker or worker.current_job_id:
+                return None
+
+            rows = session.scalars(
+                select(DigitadorJobRow)
+                .where(
+                    DigitadorJobRow.worker_id == worker_id,
+                    DigitadorJobRow.status == JobStatus.PENDING.value,
+                )
+                .order_by(DigitadorJobRow.created_at.asc())
+            ).all()
+            row = None
+            for candidate in rows:
+                is_parent = (
+                    candidate.parent_digitador_job_id is None
+                    and candidate.shard_count is not None
+                    and int(candidate.shard_count) > 1
+                )
+                if is_parent:
+                    continue
+                row = candidate
+                break
+            if not row:
+                return None
+
+            now = utcnow()
+            row.status = JobStatus.RUNNING.value
+            row.started_at = now
+            worker.current_job_id = row.id
+            worker.status = WorkerStatus.BUSY.value
+            worker.last_heartbeat = now
+
+            task_id = row.task_id
+            if row.parent_digitador_job_id:
+                parent = session.get(DigitadorJobRow, row.parent_digitador_job_id)
+                if parent:
+                    task_id = parent.task_id
+                    if parent.status == JobStatus.PENDING.value:
+                        parent.status = JobStatus.RUNNING.value
+                        parent.started_at = parent.started_at or now
+
+            if task_id:
+                task = session.get(Task, task_id)
+                if task and task.status == JobStatus.PENDING.value:
+                    task.status = JobStatus.RUNNING.value
+                    task.started_at = now
+                    task.updated_at = now
+
+            session.flush()
+            return self._row_job(row)
+
+    def request_cancel(self, job_id: str) -> Optional[Job]:
+        with session_scope() as session:
+            row = session.get(DigitadorJobRow, job_id)
+            if not row:
+                return None
+            if row.status in (
+                JobStatus.COMPLETED.value,
+                JobStatus.FAILED.value,
+                JobStatus.CANCELLED.value,
+            ):
+                return self._row_job(row)
+
+            now = utcnow()
+            row.cancel_requested = True
+            if row.status == JobStatus.PENDING.value:
+                row.status = JobStatus.CANCELLED.value
+                row.finished_at = now
+                self._sync_task(session, row, status=JobStatus.CANCELLED.value)
+
+            if row.parent_digitador_job_id is None and row.shard_count and int(row.shard_count) > 1:
+                children = session.scalars(
+                    select(DigitadorJobRow).where(
+                        DigitadorJobRow.parent_digitador_job_id == job_id
+                    )
+                ).all()
+                for child in children:
+                    if child.status in (
+                        JobStatus.COMPLETED.value,
+                        JobStatus.FAILED.value,
+                        JobStatus.CANCELLED.value,
+                    ):
+                        continue
+                    child.cancel_requested = True
+                    if child.status == JobStatus.PENDING.value:
+                        child.status = JobStatus.CANCELLED.value
+                        child.finished_at = now
+                        worker = (
+                            session.get(DigitadorWorkerRow, child.worker_id)
+                            if child.worker_id
+                            else None
+                        )
+                        if worker and worker.current_job_id == child.id:
+                            worker.current_job_id = None
+                            worker.status = WorkerStatus.ONLINE.value
+
+            session.flush()
+            return self._row_job(row)
+
+    def is_cancel_requested(self, job_id: str) -> bool:
+        job = self.get_job(job_id)
+        if not job:
+            return False
+        if job.cancel_requested:
+            return True
+        if job.parent_digitador_job_id:
+            parent = self.get_job(job.parent_digitador_job_id)
+            if parent and parent.cancel_requested:
+                return True
+        return False
+
+    def _sync_task(
+        self,
+        session,
+        row: DigitadorJobRow,
+        *,
+        status: Optional[str] = None,
+        result: Optional[Dict[str, Any]] = None,
+        error: Optional[str] = None,
+    ) -> None:
+        target = row
+        if row.parent_digitador_job_id:
+            parent = session.get(DigitadorJobRow, row.parent_digitador_job_id)
+            if parent:
+                target = parent
+        if not target.task_id:
+            return
+        task = session.get(Task, target.task_id)
+        if not task:
+            return
+        now = utcnow()
+        if status:
+            task.status = status
+        if result is not None:
+            task.result = {
+                **dict(task.result or {}),
+                "digitador_job_id": target.id,
+                **result,
+            }
+        if error is not None:
+            task.error = error
+        task.updated_at = now
+        if status in (
+            JobStatus.COMPLETED.value,
+            JobStatus.FAILED.value,
+            JobStatus.CANCELLED.value,
+        ):
+            task.finished_at = now
+
+    def patch_job(
+        self,
+        job_id: str,
+        *,
+        status: Optional[str] = None,
+        result: Optional[Dict[str, Any]] = None,
+        error: Optional[str] = None,
+    ) -> Optional[Job]:
+        parent_id: Optional[str] = None
+        with session_scope() as session:
+            row = session.get(DigitadorJobRow, job_id)
+            if not row:
+                return None
+
+            new_status = status or row.status
+            if result is not None:
+                row.result = result
+            if error is not None:
+                row.error = error
+            row.status = new_status
+
+            finished = False
+            if new_status in (
+                JobStatus.COMPLETED.value,
+                JobStatus.FAILED.value,
+                JobStatus.CANCELLED.value,
+            ):
+                row.finished_at = row.finished_at or utcnow()
+                finished = True
+
+            if not row.parent_digitador_job_id:
+                self._sync_task(
+                    session,
+                    row,
+                    status=new_status if status else None,
+                    result=result,
+                    error=error,
+                )
+
+            if finished:
+                worker = session.get(DigitadorWorkerRow, row.worker_id) if row.worker_id else None
+                if worker and worker.current_job_id == job_id:
+                    worker.current_job_id = None
+                    worker.status = WorkerStatus.ONLINE.value
+
+            parent_id = row.parent_digitador_job_id
+            session.flush()
+            job = self._row_job(row)
+
+        if parent_id:
+            self.refresh_parent_from_shards(parent_id)
+            refreshed = self.get_job(job_id)
+            return refreshed or job
+        return job
+
+    def refresh_parent_from_shards(self, parent_id: str) -> Optional[Job]:
+        with session_scope() as session:
+            parent = session.get(DigitadorJobRow, parent_id)
+            if not parent:
+                return None
+            children = list(
+                session.scalars(
+                    select(DigitadorJobRow).where(
+                        DigitadorJobRow.parent_digitador_job_id == parent_id
+                    )
+                ).all()
+            )
+            if not children:
+                return self._row_job(parent)
+
+            statuses = [c.status for c in children]
+            terminal = {
+                JobStatus.COMPLETED.value,
+                JobStatus.FAILED.value,
+                JobStatus.CANCELLED.value,
+            }
+            any_active = any(
+                s in (JobStatus.PENDING.value, JobStatus.RUNNING.value) for s in statuses
+            )
+            all_terminal = all(s in terminal for s in statuses)
+
+            total = 0
+            ok = 0
+            error_count = 0
+            timeout = 0
+            existentes = 0
+            for c in children:
+                r = dict(c.result or {})
+                total += int(r.get("total") or 0)
+                ok += int(r.get("ok") or 0)
+                error_count += int(r.get("error") or 0)
+                timeout += int(r.get("timeout") or 0)
+                existentes += int(r.get("existentes") or 0)
+
+            processed = 0
+            for c in children:
+                r = dict(c.result or {})
+                if c.status in terminal:
+                    processed += (
+                        int(r.get("ok") or 0)
+                        + int(r.get("error") or 0)
+                        + int(r.get("timeout") or 0)
+                        + int(r.get("existentes") or 0)
+                    )
+                else:
+                    processed += int(r.get("processed") or 0)
+
+            shard_ids = [c.id for c in children]
+            parent_result = dict(parent.result or {})
+            parent_result.update(
+                {
+                    "mode": "distributed",
+                    "shard_ids": shard_ids,
+                    "total": total or int(parent_result.get("total_rows") or 0),
+                    "total_rows": int(parent_result.get("total_rows") or total or 0),
+                    "ok": ok,
+                    "error": error_count,
+                    "timeout": timeout,
+                    "existentes": existentes,
+                    "processed": processed,
+                    "reports_prefix": parent.reports_prefix,
+                    "shards": [
+                        {
+                            "id": c.id,
+                            "worker_id": c.worker_id,
+                            "status": c.status,
+                            "shard_index": c.shard_index,
+                            "result": dict(c.result or {}),
+                            "error": c.error,
+                        }
+                        for c in children
+                    ],
+                }
+            )
+            parent.result = parent_result
+
+            now = utcnow()
+            if any_active:
+                parent.status = JobStatus.RUNNING.value
+                parent.started_at = parent.started_at or now
+                parent.finished_at = None
+                parent.error = None
+                self._sync_task(
+                    session,
+                    parent,
+                    status=JobStatus.RUNNING.value,
+                    result=parent_result,
+                )
+            elif all_terminal:
+                any_failed = any(s == JobStatus.FAILED.value for s in statuses)
+                any_cancelled = any(s == JobStatus.CANCELLED.value for s in statuses)
+                all_cancelled = all(s == JobStatus.CANCELLED.value for s in statuses)
+                if all_cancelled:
+                    parent.status = JobStatus.CANCELLED.value
+                elif any_failed:
+                    parent.status = JobStatus.FAILED.value
+                    parent.error = next(
+                        (c.error for c in children if c.error),
+                        "Uno o más shards fallaron",
+                    )
+                elif any_cancelled and not any_failed:
+                    parent.status = JobStatus.COMPLETED.value
+                else:
+                    parent.status = JobStatus.COMPLETED.value
+                parent.finished_at = parent.finished_at or now
+                self._sync_task(
+                    session,
+                    parent,
+                    status=parent.status,
+                    result=parent_result,
+                    error=parent.error,
+                )
+
+            session.flush()
+            return self._row_job(parent)
+
+    def reclaim_stale_running_jobs(self) -> int:
+        """Jobs running cuyo worker está offline → pending + reasignar a otro online."""
+        reclaimed = 0
+        with session_scope() as session:
+            now = utcnow()
+            cutoff = now - timedelta(seconds=HEARTBEAT_OFFLINE_SECONDS)
+            workers = list(session.scalars(select(DigitadorWorkerRow)).all())
+            offline_ids: set[str] = set()
+            online_free: List[DigitadorWorkerRow] = []
+            for w in workers:
+                hb = w.last_heartbeat
+                if hb is None:
+                    offline_ids.add(w.id)
+                    continue
+                if hb.tzinfo is None:
+                    hb = hb.replace(tzinfo=timezone.utc)
+                if hb < cutoff:
+                    offline_ids.add(w.id)
+                    if w.current_job_id:
+                        w.current_job_id = None
+                    w.status = WorkerStatus.OFFLINE.value
+                elif not w.current_job_id:
+                    online_free.append(w)
+
+            if not offline_ids:
+                session.flush()
+                return 0
+
+            running = session.scalars(
+                select(DigitadorJobRow).where(
+                    DigitadorJobRow.status == JobStatus.RUNNING.value,
+                    DigitadorJobRow.worker_id.in_(list(offline_ids)),
+                )
+            ).all()
+
+            rr = 0
+            for row in running:
+                is_parent = (
+                    row.parent_digitador_job_id is None
+                    and row.shard_count is not None
+                    and int(row.shard_count) > 1
+                )
+                if is_parent:
+                    continue
+
+                row.status = JobStatus.PENDING.value
+                row.started_at = None
+                row.cancel_requested = False
+
+                if online_free:
+                    target = online_free[rr % len(online_free)]
+                    rr += 1
+                    row.worker_id = target.id
+
+                reclaimed += 1
+                session.add(
+                    DigitadorJobEventRow(
+                        digitador_job_id=row.id,
+                        event_type="warning",
+                        message=(
+                            "Worker offline; job reclaimed"
+                            + (f" → {row.worker_id}" if row.worker_id else "")
+                        ),
+                        timestamp=now,
+                        metadata_json={"reclaimed": True, "worker_id": row.worker_id},
+                    )
+                )
+                if row.parent_digitador_job_id:
+                    session.add(
+                        DigitadorJobEventRow(
+                            digitador_job_id=row.parent_digitador_job_id,
+                            event_type="warning",
+                            message=f"Shard {row.shard_index} reclamado (worker offline)",
+                            timestamp=now,
+                            metadata_json={
+                                "shard_id": row.id,
+                                "worker_id": row.worker_id,
+                            },
+                        )
+                    )
+
+            if online_free:
+                pending_orphan = session.scalars(
+                    select(DigitadorJobRow).where(
+                        DigitadorJobRow.status == JobStatus.PENDING.value,
+                        DigitadorJobRow.worker_id.in_(list(offline_ids)),
+                    )
+                ).all()
+                for row in pending_orphan:
+                    is_parent = (
+                        row.parent_digitador_job_id is None
+                        and row.shard_count is not None
+                        and int(row.shard_count) > 1
+                    )
+                    if is_parent:
+                        continue
+                    target = online_free[rr % len(online_free)]
+                    rr += 1
+                    if row.worker_id != target.id:
+                        row.worker_id = target.id
+                        reclaimed += 1
+
+            session.flush()
+
+        return reclaimed
+
+    def add_event(
+        self,
+        job_id: str,
+        event_type: str,
+        message: str,
+        *,
+        radicado: Optional[str] = None,
+        nit: Optional[str] = None,
+        index: Optional[int] = None,
+        total: Optional[int] = None,
+        metadata: Optional[Dict[str, Any]] = None,
+        timestamp: Optional[str] = None,
+    ) -> JobEventRecord:
+        with session_scope() as session:
+            ts = _parse_iso(timestamp) or utcnow()
+            ev = DigitadorJobEventRow(
+                digitador_job_id=job_id,
+                event_type=event_type,
+                message=message,
+                timestamp=ts,
+                radicado=radicado,
+                nit=nit,
+                idx=index,
+                total=total,
+                metadata_json=metadata or {},
+            )
+            session.add(ev)
+            session.flush()
+            row = session.get(DigitadorJobRow, job_id)
+            if row and row.parent_digitador_job_id:
+                session.add(
+                    DigitadorJobEventRow(
+                        digitador_job_id=row.parent_digitador_job_id,
+                        event_type=event_type,
+                        message=message,
+                        timestamp=ts,
+                        radicado=radicado,
+                        nit=nit,
+                        idx=index,
+                        total=total,
+                        metadata_json={
+                            **(metadata or {}),
+                            "shard_id": job_id,
+                            "shard_index": row.shard_index,
+                            "worker_id": row.worker_id,
+                        },
+                    )
+                )
+                session.flush()
+            return JobEventRecord(
+                id=int(ev.id),
+                job_id=job_id,
+                event_type=event_type,
+                message=message,
+                timestamp=_iso(ts) or utc_now(),
+                radicado=radicado,
+                nit=nit,
+                index=index,
+                total=total,
+                metadata=metadata or {},
+            )
+
+    def list_events(self, job_id: str, since_id: int = 0) -> List[JobEventRecord]:
+        with session_scope() as session:
+            rows = session.scalars(
+                select(DigitadorJobEventRow)
+                .where(
+                    DigitadorJobEventRow.digitador_job_id == job_id,
+                    DigitadorJobEventRow.id > since_id,
+                )
+                .order_by(DigitadorJobEventRow.id.asc())
+            ).all()
+            return [
+                JobEventRecord(
+                    id=int(r.id),
+                    job_id=r.digitador_job_id,
+                    event_type=r.event_type,
+                    message=r.message,
+                    timestamp=_iso(r.timestamp) or utc_now(),
+                    radicado=r.radicado,
+                    nit=r.nit,
+                    index=r.idx,
+                    total=r.total,
+                    metadata=dict(r.metadata_json or {}),
+                )
+                for r in rows
+            ]
+
+    def create_digitador_task(
+        self,
+        *,
+        parent_job_id: str,
+        digitador_job_id: str,
+        worker_id: Optional[str],
+        title: str,
+        original_filename: str,
+        worker_ids: Optional[List[str]] = None,
+        mode: str = "single",
+        source_type: Optional[str] = None,
+    ) -> str:
+        """Crea Task(type=digitador) y la asocia al digitador_job. Devuelve task_id."""
+        with session_scope() as session:
+            now = utcnow()
+            request: Dict[str, Any] = {
+                "worker_id": worker_id,
+                "worker_ids": worker_ids or ([worker_id] if worker_id else []),
+                "original_filename": original_filename,
+                "digitador_job_id": digitador_job_id,
+                "mode": mode,
+            }
+            result: Dict[str, Any] = {"digitador_job_id": digitador_job_id, "mode": mode}
+            if source_type:
+                request["source_type"] = source_type
+                result["source_type"] = source_type
+            task = Task(
+                job_id=parent_job_id,
+                type="digitador",
+                status=JobStatus.PENDING.value,
+                title=title or f"Bot digitador ({original_filename})",
+                request=request,
+                result=result,
+                created_at=now,
+                updated_at=now,
+            )
+            session.add(task)
+            session.flush()
+            row = session.get(DigitadorJobRow, digitador_job_id)
+            if row:
+                row.task_id = task.id
+                row.job_id = parent_job_id
+            session.flush()
+            return task.id
+
+    def parent_job_exists(self, parent_job_id: str) -> bool:
+        from infrastructure.database import Job as IpsJob
+
+        with session_scope() as session:
+            return session.get(IpsJob, parent_job_id) is not None

services/bot_digitador/requirements.txt

@@ -0,0 +1,10 @@
+fastapi>=0.110.0
+uvicorn[standard]>=0.27.0
+python-multipart>=0.0.9
+boto3>=1.34.0
+httpx>=0.27.0
+openpyxl>=3.1.0
+pydantic>=2.0.0
+sqlalchemy>=2.0.0
+psycopg[binary]>=3.1.0
+pytest>=7.0.0

services/bot_digitador/tests/test_csv_split.py

@@ -0,0 +1,96 @@
+"""Tests de parseo/partición CSV para digitador."""
+from __future__ import annotations
+
+import pytest
+
+from domain.csv_split import (
+    REQUIRED_BOT_COLUMNS,
+    parse_digitador_rows,
+    rows_to_csv_bytes,
+    split_rows,
+)
+
+
+def _it_csv_bytes() -> bytes:
+    header = "|".join(
+        [
+            "RADICADO",
+            "NUMERO",
+            "FECHA",
+            "BRUTO",
+            "AUTORIZACION",
+            "TIPO_ID",
+            "SERVICIO",
+            "COPAGO",
+            "MODERADORA",
+        ]
+    )
+    row = "|".join(
+        ["R-1", "123", "2024-01-01", "1000", "A-9", "CC", "AMB", "0", "4500"]
+    )
+    return f"{header}\n{row}\n".encode("utf-8-sig")
+
+
+def _snake_csv_bytes() -> bytes:
+    header = "|".join(REQUIRED_BOT_COLUMNS)
+    row = "|".join(
+        ["R-2", "999", "2024-02-02", "2000", "A-1", "TI", "HOS", "100", "0"]
+    )
+    return f"{header}\n{row}\n".encode("utf-8")
+
+
+def test_it_to_bot_mapping():
+    rows = parse_digitador_rows(_it_csv_bytes(), "aptos.csv")
+    assert len(rows) == 1
+    assert rows[0]["radicado"] == "R-1"
+    assert rows[0]["documento"] == "123"
+    assert rows[0]["ingreso"] == "2024-01-01"
+    assert rows[0]["total_cargos"] == "1000"
+    assert rows[0]["autorizacion"] == "A-9"
+    assert rows[0]["tipo_documento"] == "CC"
+    assert rows[0]["tipo_servicio"] == "AMB"
+    assert rows[0]["copago"] == "0"
+    assert rows[0]["cuota_moderadora"] == "4500"
+
+
+def test_snake_case_headers():
+    rows = parse_digitador_rows(_snake_csv_bytes(), "input.csv")
+    assert len(rows) == 1
+    assert rows[0]["radicado"] == "R-2"
+    assert rows[0]["documento"] == "999"
+
+
+def test_missing_columns_raises():
+    bad = b"RADICADO|NUMERO\nR1|1\n"
+    with pytest.raises(ValueError) as exc:
+        parse_digitador_rows(bad, "bad.csv")
+    assert "Columnas faltantes" in str(exc.value)
+
+
+def test_split_rows_balanced():
+    rows = [{"radicado": str(i)} for i in range(100)]
+    chunks = split_rows(rows, 2)
+    assert len(chunks) == 2
+    assert len(chunks[0]) == 50
+    assert len(chunks[1]) == 50
+
+
+def test_split_rows_uneven():
+    rows = [{"radicado": str(i)} for i in range(10)]
+    chunks = split_rows(rows, 3)
+    assert len(chunks) == 3
+    assert sum(len(c) for c in chunks) == 10
+    assert [len(c) for c in chunks] == [4, 3, 3]
+
+
+def test_roundtrip_csv():
+    rows = parse_digitador_rows(_it_csv_bytes(), "aptos.csv")
+    data = rows_to_csv_bytes(rows)
+    parsed = parse_digitador_rows(data, "shard.csv")
+    assert parsed == rows
+
+
+def test_split_more_workers_than_rows():
+    rows = [{"radicado": "A"}, {"radicado": "B"}]
+    chunks = split_rows(rows, 5)
+    assert len(chunks) == 2

services/cleaning/domain/csv_cleaning/apply.py

@@ -49,6 +49,11 @@ def apply_rules(
     if col not in result.columns:
       log.append(f"[WARN] Columna no encontrada: {col}")
       return
+    # Eliminar columna: no tiene sentido transformar y luego borrar.
+    if "drop_column" in rule_ids:
+      result.drop(columns=[col], inplace=True)
+      log.append(f"[OK] '{col}' <- {get_rule_name('drop_column')}")
+      return
     split_has_homologation = (
       "doc_id_split_tipo_numero" in rule_ids and "doc_type_homologate" in rule_ids
     )

services/cleaning/domain/csv_cleaning/rules/builtins.py

@@ -140,5 +140,14 @@ register_rule(
   ),
   fn=special._rule_passthrough_series,
 )
+register_rule(
+  "drop_column",
+  name="Eliminar columna",
+  description=(
+    "Quita por completo esta columna del resultado limpio. "
+    "Útil para campos que no se necesitan en el export o en pasos posteriores."
+  ),
+  fn=special._rule_passthrough_series,
+)
 
 rebuild_cleaning_rules()

services/cleaning/domain/csv_cleaning/type_inference.py

@@ -99,7 +99,12 @@ def _infer_column_type(
 
 def _applicable_rules_for_type(inferred_type: str) -> list[str]:
   """Identificadores de reglas a mostrar según el tipo inferido (sin redundancias)."""
-  base = ["strip_special_spaces", "strip_special_keep_date_seps", "strip_spaces"]
+  base = [
+    "strip_special_spaces",
+    "strip_special_keep_date_seps",
+    "strip_spaces",
+    "drop_column",
+  ]
   if inferred_type == TYPE_INTEGER:
     return base + ["cop_mod_split", "cop_mod_bruto_neto"]
   if inferred_type == TYPE_DECIMAL:

services/db/alembic/versions/0007_download_job_shards.py

@@ -0,0 +1,83 @@
+"""Download jobs: parent/shard fields for distributed bots.
+
+Revision ID: 0007
+Revises: 0006
+Create Date: 2026-08-09
+
+"""
+from __future__ import annotations
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0007"
+down_revision: Union[str, None] = "0006"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    if not inspector.has_table("download_jobs"):
+        return
+    cols = {c["name"] for c in inspector.get_columns("download_jobs")}
+
+    with op.batch_alter_table("download_jobs") as batch:
+        if "parent_download_job_id" not in cols:
+            batch.add_column(
+                sa.Column(
+                    "parent_download_job_id",
+                    sa.String(36),
+                    sa.ForeignKey("download_jobs.id"),
+                    nullable=True,
+                )
+            )
+        if "shard_index" not in cols:
+            batch.add_column(sa.Column("shard_index", sa.Integer(), nullable=True))
+        if "shard_count" not in cols:
+            batch.add_column(sa.Column("shard_count", sa.Integer(), nullable=True))
+
+    # Padre orquestador no tiene worker asignado
+    op.alter_column(
+        "download_jobs",
+        "worker_id",
+        existing_type=sa.String(36),
+        nullable=True,
+    )
+
+    existing_indexes = {ix["name"] for ix in inspector.get_indexes("download_jobs")}
+    if "ix_download_jobs_parent" not in existing_indexes:
+        op.create_index(
+            "ix_download_jobs_parent",
+            "download_jobs",
+            ["parent_download_job_id"],
+        )
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    if not inspector.has_table("download_jobs"):
+        return
+    existing_indexes = {ix["name"] for ix in inspector.get_indexes("download_jobs")}
+    if "ix_download_jobs_parent" in existing_indexes:
+        op.drop_index("ix_download_jobs_parent", table_name="download_jobs")
+
+    cols = {c["name"] for c in inspector.get_columns("download_jobs")}
+    with op.batch_alter_table("download_jobs") as batch:
+        if "shard_count" in cols:
+            batch.drop_column("shard_count")
+        if "shard_index" in cols:
+            batch.drop_column("shard_index")
+        if "parent_download_job_id" in cols:
+            batch.drop_column("parent_download_job_id")
+
+    op.alter_column(
+        "download_jobs",
+        "worker_id",
+        existing_type=sa.String(36),
+        nullable=False,
+    )

services/db/alembic/versions/0008_bot_digitador_tables.py

@@ -0,0 +1,152 @@
+"""Bot digitador: digitador_workers, digitador_jobs, digitador_job_events.
+
+Revision ID: 0008
+Revises: 0007
+Create Date: 2026-08-09
+
+"""
+from __future__ import annotations
+
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+from sqlalchemy.dialects import postgresql
+
+revision: str = "0008"
+down_revision: Union[str, None] = "0007"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def _table_exists(conn: sa.engine.Connection, table: str) -> bool:
+    return sa.inspect(conn).has_table(table)
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    json_type = (
+        postgresql.JSONB()
+        if conn.dialect.name == "postgresql"
+        else sa.JSON()
+    )
+
+    if not _table_exists(conn, "digitador_workers"):
+        op.create_table(
+            "digitador_workers",
+            sa.Column("id", sa.String(36), primary_key=True),
+            sa.Column("hostname", sa.Text(), nullable=False, server_default=""),
+            sa.Column("max_concurrent", sa.Integer(), nullable=False, server_default="1"),
+            sa.Column("status", sa.String(32), nullable=False),
+            sa.Column("last_heartbeat", sa.DateTime(timezone=True), nullable=False),
+            sa.Column("current_job_id", sa.String(36), nullable=True),
+        )
+        op.create_index("ix_digitador_workers_status", "digitador_workers", ["status"])
+
+    if not _table_exists(conn, "digitador_jobs"):
+        op.create_table(
+            "digitador_jobs",
+            sa.Column("id", sa.String(36), primary_key=True),
+            sa.Column("job_id", sa.String(36), sa.ForeignKey("jobs.id"), nullable=True),
+            sa.Column("task_id", sa.String(36), sa.ForeignKey("tasks.id"), nullable=True),
+            sa.Column("worker_id", sa.String(36), nullable=True),
+            sa.Column(
+                "parent_digitador_job_id",
+                sa.String(36),
+                sa.ForeignKey("digitador_jobs.id"),
+                nullable=True,
+            ),
+            sa.Column("shard_index", sa.Integer(), nullable=True),
+            sa.Column("shard_count", sa.Integer(), nullable=True),
+            sa.Column("status", sa.String(32), nullable=False, server_default="pending"),
+            sa.Column("title", sa.Text(), nullable=True),
+            sa.Column("csv_key", sa.Text(), nullable=False, server_default=""),
+            sa.Column("reports_prefix", sa.Text(), nullable=False, server_default=""),
+            sa.Column("original_filename", sa.Text(), nullable=False, server_default=""),
+            sa.Column(
+                "cancel_requested",
+                sa.Boolean(),
+                nullable=False,
+                server_default=sa.false(),
+            ),
+            sa.Column(
+                "result",
+                json_type,
+                nullable=False,
+                server_default=(
+                    sa.text("'{}'::jsonb")
+                    if conn.dialect.name == "postgresql"
+                    else sa.text("'{}'")
+                ),
+            ),
+            sa.Column("error", sa.Text(), nullable=True),
+            sa.Column("created_at", sa.DateTime(timezone=True), nullable=False),
+            sa.Column("started_at", sa.DateTime(timezone=True), nullable=True),
+            sa.Column("finished_at", sa.DateTime(timezone=True), nullable=True),
+        )
+        op.create_index("ix_digitador_jobs_job_id", "digitador_jobs", ["job_id"])
+        op.create_index("ix_digitador_jobs_task_id", "digitador_jobs", ["task_id"])
+        op.create_index("ix_digitador_jobs_worker_id", "digitador_jobs", ["worker_id"])
+        op.create_index("ix_digitador_jobs_status", "digitador_jobs", ["status"])
+        op.create_index(
+            "ix_digitador_jobs_parent",
+            "digitador_jobs",
+            ["parent_digitador_job_id"],
+        )
+        op.create_index(
+            "ix_digitador_jobs_worker_status",
+            "digitador_jobs",
+            ["worker_id", "status"],
+        )
+
+    if not _table_exists(conn, "digitador_job_events"):
+        op.create_table(
+            "digitador_job_events",
+            sa.Column("id", sa.Integer(), primary_key=True, autoincrement=True),
+            sa.Column(
+                "digitador_job_id",
+                sa.String(36),
+                sa.ForeignKey("digitador_jobs.id"),
+                nullable=False,
+            ),
+            sa.Column("event_type", sa.String(64), nullable=False),
+            sa.Column("message", sa.Text(), nullable=False),
+            sa.Column("timestamp", sa.DateTime(timezone=True), nullable=False),
+            sa.Column("radicado", sa.String(255), nullable=True),
+            sa.Column("nit", sa.String(64), nullable=True),
+            sa.Column("idx", sa.Integer(), nullable=True),
+            sa.Column("total", sa.Integer(), nullable=True),
+            sa.Column(
+                "metadata",
+                json_type,
+                nullable=False,
+                server_default=(
+                    sa.text("'{}'::jsonb")
+                    if conn.dialect.name == "postgresql"
+                    else sa.text("'{}'")
+                ),
+            ),
+        )
+        op.create_index(
+            "ix_digitador_job_events_job_id",
+            "digitador_job_events",
+            ["digitador_job_id", "id"],
+        )
+
+
+def downgrade() -> None:
+    conn = op.get_bind()
+    if _table_exists(conn, "digitador_job_events"):
+        op.drop_index("ix_digitador_job_events_job_id", table_name="digitador_job_events")
+        op.drop_table("digitador_job_events")
+    if _table_exists(conn, "digitador_jobs"):
+        op.drop_index("ix_digitador_jobs_worker_status", table_name="digitador_jobs")
+        op.drop_index("ix_digitador_jobs_parent", table_name="digitador_jobs")
+        op.drop_index("ix_digitador_jobs_status", table_name="digitador_jobs")
+        op.drop_index("ix_digitador_jobs_worker_id", table_name="digitador_jobs")
+        op.drop_index("ix_digitador_jobs_task_id", table_name="digitador_jobs")
+        op.drop_index("ix_digitador_jobs_job_id", table_name="digitador_jobs")
+        op.drop_table("digitador_jobs")
+    if _table_exists(conn, "digitador_workers"):
+        op.drop_index("ix_digitador_workers_status", table_name="digitador_workers")
+        op.drop_table("digitador_workers")

services/desktop_app/src/renderer/App.tsx

@@ -5,6 +5,7 @@ import {
   Cpu,
   Download,
   Hash,
+  Keyboard,
   Link2,
   Package,
   Settings,
@@ -35,6 +36,7 @@ import "./modules/extraction/ExtractionModule";
 import "./modules/cleaning/CleaningModule";
 import "./modules/poblacion_apta/PoblacionAptaModule";
 import "./modules/downloads/DownloadsModule";
+import "./modules/digitador/DigitadorModule";
 
 const BOT_PACKAGE_URL =
   "http://s3.builker.xyz/apps-versions/BotDescargaPYSCuentasMedicas/latest/bot_descarga.zip";
@@ -403,6 +405,14 @@ function App() {
                 <Users size={17} />
                 <span>Población apta</span>
               </button>
+              <button
+                type="button"
+                className={jobSection === "digitador" ? "active" : ""}
+                onClick={() => handleActivateSection("digitador")}
+              >
+                <Keyboard size={17} />
+                <span>Bot digitador</span>
+              </button>
               <button
                 type="button"
                 className={jobSection === "results" ? "active" : ""}

services/desktop_app/src/renderer/components/MinioDownloadSourcePicker.tsx

@@ -28,7 +28,11 @@ interface Props {
 
 function jobLabel(r: DownloadJob): string {
   const title = r.title || r.original_filename || r.id.slice(0, 8);
-  return `${title} · ${r.status} · ${r.worker_id}`;
+  const bots =
+    r.shard_count && Number(r.shard_count) > 1
+      ? `${r.shard_count} bots`
+      : r.worker_id || "—";
+  return `${title} · ${r.status} · ${bots}`;
 }
 
 export function MinioDownloadSourcePicker({

services/desktop_app/src/renderer/lib/apiClient.ts

@@ -9,6 +9,10 @@ import type {
   CountSyncResponse,
   CreateCountJobPayload,
   CreateJobPayload,
+  DigitadorJob,
+  DigitadorJobEvent,
+  DigitadorJobFilesResponse,
+  DigitadorWorker,
   DownloadJob,
   DownloadJobEvent,
   DownloadJobFilesResponse,
@@ -87,7 +91,8 @@ export class ApiClient {
       path.startsWith("/cleaning/jobs/from-csv") ||
       path.startsWith("/poblacion-apta/jobs/from-csv") ||
       path.startsWith("/poblacion-apta/jobs/from-cleaning") ||
-      path === "/bot-descarga/jobs"
+      path === "/bot-descarga/jobs" ||
+      path === "/bot-digitador/jobs"
     );
   }
 
@@ -310,6 +315,83 @@ export class ApiClient {
     return this.handle<DownloadJob>(response);
   }
 
+  // ── Bot Digitador (control remoto) ─────────────────────────────────────────
+
+  async listDigitadorWorkers(): Promise<DigitadorWorker[]> {
+    const response = await fetchWithTimeout(
+      this.url("/bot-digitador/workers"),
+      { headers: this.botHeaders() },
+      this.timeoutFor("/bot-digitador/workers"),
+    );
+    return this.handle<DigitadorWorker[]>(response);
+  }
+
+  async createDigitadorJob(form: FormData): Promise<DigitadorJob> {
+    const response = await fetchWithTimeout(
+      this.url("/bot-digitador/jobs"),
+      { method: "POST", headers: this.botHeaders(), body: form },
+      this.timeoutFor("/bot-digitador/jobs"),
+    );
+    return this.handle<DigitadorJob>(response);
+  }
+
+  async listDigitadorJobs(limit = 50, parentJobId?: string): Promise<DigitadorJob[]> {
+    const qs = new URLSearchParams({ limit: String(limit) });
+    if (parentJobId) qs.set("job_id", parentJobId);
+    const response = await fetchWithTimeout(
+      this.url(`/bot-digitador/jobs?${qs.toString()}`),
+      { headers: this.botHeaders() },
+    );
+    return this.handle<DigitadorJob[]>(response);
+  }
+
+  async getDigitadorJob(jobId: string): Promise<DigitadorJob> {
+    const response = await fetchWithTimeout(
+      this.url(`/bot-digitador/jobs/${encodeURIComponent(jobId)}`),
+      { headers: this.botHeaders() },
+    );
+    return this.handle<DigitadorJob>(response);
+  }
+
+  async listDigitadorJobFiles(jobId: string): Promise<DigitadorJobFilesResponse> {
+    const response = await fetchWithTimeout(
+      this.url(`/bot-digitador/jobs/${encodeURIComponent(jobId)}/files`),
+      { headers: this.botHeaders() },
+    );
+    return this.handle<DigitadorJobFilesResponse>(response);
+  }
+
+  /** Descarga bytes de un reporte CSV/log/json del resultado MinIO (proxy bot-digitador). */
+  async fetchDigitadorObject(jobId: string, key: string): Promise<ArrayBuffer> {
+    const qs = new URLSearchParams({ key });
+    const response = await fetchWithTimeout(
+      this.url(`/bot-digitador/jobs/${encodeURIComponent(jobId)}/object?${qs.toString()}`),
+      { headers: this.botHeaders() },
+      120_000,
+    );
+    if (!response.ok) {
+      const text = await response.text().catch(() => "");
+      throw new Error(text || `Error ${response.status} al leer objeto MinIO`);
+    }
+    return response.arrayBuffer();
+  }
+
+  async getDigitadorEvents(jobId: string, since = 0): Promise<{ events: DigitadorJobEvent[] }> {
+    const response = await fetchWithTimeout(
+      this.url(`/bot-digitador/jobs/${encodeURIComponent(jobId)}/events?since=${since}`),
+      { headers: this.botHeaders() },
+    );
+    return this.handle<{ events: DigitadorJobEvent[] }>(response);
+  }
+
+  async cancelDigitadorJob(jobId: string): Promise<DigitadorJob> {
+    const response = await fetchWithTimeout(
+      this.url(`/bot-digitador/jobs/${encodeURIComponent(jobId)}/cancel`),
+      { method: "POST", headers: this.botHeaders() },
+    );
+    return this.handle<DigitadorJob>(response);
+  }
+
   // ── Count Jobs ─────────────────────────────────────────────────────────────
 
   /** Envía PDFs al servicio de conteo (síncrono). */
@@ -456,7 +538,7 @@ export class ApiClient {
 
   async poblacionAptaExportCsv(
     jobId: string,
-    subset: "aptos" | "no_aptos" = "aptos",
+    subset: "aptos" | "no_aptos" | "todos" = "aptos",
   ): Promise<Blob> {
     const params = new URLSearchParams({ subset });
     const path = `/poblacion-apta/jobs/${encodeURIComponent(jobId)}/export.csv?${params}`;

services/desktop_app/src/renderer/modules/digitador/DigitadorModule.ts

@@ -0,0 +1,18 @@
+import type { ModuleDescriptor } from "../registry";
+import { registerModule } from "../registry";
+import type { Task } from "../../../shared/types";
+import { DigitadorWizardModule } from "./DigitadorWizardModule";
+import { DigitadorResultsPanel } from "./DigitadorResultsPanel";
+
+const DigitadorModule: ModuleDescriptor = {
+  type: "digitador",
+  label: "Bot digitador",
+  iconName: "Keyboard",
+  hasTask: (tasks: Task[]) => tasks.some((t) => t.type === "digitador"),
+  Wizard: DigitadorWizardModule as unknown as ModuleDescriptor["Wizard"],
+  Results: DigitadorResultsPanel as unknown as ModuleDescriptor["Results"],
+};
+
+registerModule(DigitadorModule);
+
+export { DigitadorModule };

services/desktop_app/src/renderer/modules/digitador/DigitadorResultsPanel.tsx

@@ -0,0 +1,400 @@
+import {
+  AlertCircle,
+  CheckCircle2,
+  Download,
+  Keyboard,
+  Loader2,
+  Plus,
+  RefreshCw,
+} from "lucide-react";
+import { useCallback, useEffect, useState } from "react";
+import type {
+  DigitadorJob,
+  DigitadorJobEvent,
+  DigitadorJobFile,
+  Job,
+  Task,
+} from "../../../shared/types";
+import type { ApiClient } from "../../lib/apiClient";
+
+declare const desktopApi: import("../../../shared/types").DesktopBridge | undefined;
+
+interface Props {
+  client: ApiClient;
+  job: Job;
+  tasks: Task[];
+  onStartDigitador: () => void;
+}
+
+function reportBasename(file: DigitadorJobFile): string {
+  const named = (file.name || "").trim();
+  if (named) return named;
+  return file.key.split("/").pop() || "reporte.csv";
+}
+
+function extensionFilters(filename: string): Array<{ name: string; extensions: string[] }> {
+  const lower = filename.toLowerCase();
+  if (lower.endsWith(".csv")) return [{ name: "CSV", extensions: ["csv"] }];
+  if (lower.endsWith(".json")) return [{ name: "JSON", extensions: ["json"] }];
+  if (lower.endsWith(".log")) return [{ name: "Log", extensions: ["log"] }];
+  return [{ name: "Archivo", extensions: ["*"] }];
+}
+
+async function saveReportAs(filename: string, data: ArrayBuffer): Promise<boolean> {
+  const safeName = reportBasename({ key: filename, name: filename });
+  if (typeof desktopApi?.saveFileDialog === "function" && typeof desktopApi?.writeBytesFile === "function") {
+    const dest = await desktopApi.saveFileDialog(safeName, extensionFilters(safeName));
+    if (!dest) return false;
+    await desktopApi.writeBytesFile(dest, data);
+    return true;
+  }
+  const blob = new Blob([data]);
+  const url = URL.createObjectURL(blob);
+  const a = document.createElement("a");
+  a.href = url;
+  a.download = safeName;
+  a.click();
+  URL.revokeObjectURL(url);
+  return true;
+}
+
+export function DigitadorResultsPanel({ client, job, tasks, onStartDigitador }: Props) {
+  const [runs, setRuns] = useState<DigitadorJob[]>([]);
+  const [selectedId, setSelectedId] = useState<string>("");
+  const [events, setEvents] = useState<DigitadorJobEvent[]>([]);
+  const [files, setFiles] = useState<DigitadorJobFile[]>([]);
+  const [filesError, setFilesError] = useState<string | null>(null);
+  const [loading, setLoading] = useState(false);
+  const [error, setError] = useState<string | null>(null);
+  const [dlBusyKey, setDlBusyKey] = useState<string | null>(null);
+  const [dlStatus, setDlStatus] = useState<string | null>(null);
+
+  const digitadorTasks = tasks
+    .filter((t) => t.type === "digitador")
+    .sort((a, b) => new Date(b.created_at).getTime() - new Date(a.created_at).getTime());
+
+  const loadRuns = useCallback(async () => {
+    setLoading(true);
+    setError(null);
+    try {
+      const list = await client.listDigitadorJobs(100, job.id);
+      setRuns(list);
+      setSelectedId((prev) => {
+        if (prev && list.some((r) => r.id === prev)) return prev;
+        return list[0]?.id ?? "";
+      });
+    } catch (e) {
+      setError(e instanceof Error ? e.message : String(e));
+    } finally {
+      setLoading(false);
+    }
+  }, [client, job.id]);
+
+  useEffect(() => {
+    void loadRuns();
+  }, [loadRuns, digitadorTasks.length]);
+
+  const selected = runs.find((r) => r.id === selectedId) ?? null;
+  const running = Boolean(
+    selected && !["completed", "failed", "cancelled"].includes(selected.status),
+  );
+
+  useEffect(() => {
+    if (!selectedId) {
+      setEvents([]);
+      setFiles([]);
+      setFilesError(null);
+      return;
+    }
+
+    let cancelled = false;
+    async function loadDetail() {
+      try {
+        const [{ events: ev }, fresh] = await Promise.all([
+          client.getDigitadorEvents(selectedId, 0),
+          client.getDigitadorJob(selectedId),
+        ]);
+        if (cancelled) return;
+        setEvents(ev);
+        setRuns((prev) => prev.map((r) => (r.id === fresh.id ? fresh : r)));
+
+        try {
+          const fileResp = await client.listDigitadorJobFiles(selectedId);
+          if (cancelled) return;
+          setFiles(fileResp.files);
+          setFilesError(null);
+        } catch (e) {
+          if (cancelled) return;
+          setFiles([]);
+          setFilesError(e instanceof Error ? e.message : String(e));
+        }
+      } catch (e) {
+        if (!cancelled) setError(e instanceof Error ? e.message : String(e));
+      }
+    }
+
+    void loadDetail();
+    if (!running) return;
+
+    const t = setInterval(() => void loadDetail(), 3000);
+    return () => {
+      cancelled = true;
+      clearInterval(t);
+    };
+  }, [client, selectedId, running]);
+
+  const result = selected?.result ?? {};
+  const exitosos = Number(result.exitosos ?? 0);
+  const fallidos = Number(result.fallidos ?? 0);
+  const existentes = Number(result.existentes ?? 0);
+  const timeouts = Number(result.timeouts ?? 0);
+  const total = Number(result.total ?? 0);
+  const bulkBusy = Boolean(dlBusyKey);
+
+  async function downloadOne(file: DigitadorJobFile) {
+    if (!selectedId || bulkBusy) return;
+    const busyId = `file:${file.key}`;
+    setDlBusyKey(busyId);
+    setDlStatus(null);
+    setError(null);
+    try {
+      const data = await client.fetchDigitadorObject(selectedId, file.key);
+      const name = reportBasename(file);
+      const ok = await saveReportAs(name, data);
+      if (ok) setDlStatus(`Guardado: ${name}`);
+    } catch (e) {
+      setError(e instanceof Error ? e.message : "No se pudo descargar el reporte.");
+    } finally {
+      setDlBusyKey(null);
+    }
+  }
+
+  if (runs.length === 0 && !loading) {
+    return (
+      <div className="results-empty">
+        <Keyboard size={28} />
+        <p>Aún no hay corridas de bot digitador en este trabajo.</p>
+        <button type="button" className="btn btn-primary" onClick={onStartDigitador}>
+          <Plus size={15} /> Nueva digitación
+        </button>
+      </div>
+    );
+  }
+
+  return (
+    <div className="results-section">
+      <div className="results-section-header">
+        <div className="results-section-header-left">
+          <label className="wiz-field-label">Corrida</label>
+          <select
+            className="form-input"
+            value={selectedId}
+            onChange={(e) => {
+              setSelectedId(e.target.value);
+              setDlStatus(null);
+            }}
+            style={{ minWidth: 280 }}
+          >
+            {runs.map((r) => (
+              <option key={r.id} value={r.id}>
+                {r.title || r.original_filename || r.id.slice(0, 8)} · {r.status}
+                {r.shard_count && Number(r.shard_count) > 1
+                  ? ` · ${r.shard_count} bots`
+                  : r.worker_id
+                    ? ` · ${r.worker_id}`
+                    : ""}
+              </option>
+            ))}
+          </select>
+          {selected && (
+            <span className={`status-badge status-badge--${selected.status}`}>
+              {selected.status}
+            </span>
+          )}
+        </div>
+        <div className="results-section-header-right">
+          <button type="button" className="btn btn-sm btn-ghost" onClick={() => void loadRuns()} disabled={loading}>
+            {loading ? <Loader2 size={13} className="spin" /> : <RefreshCw size={13} />}
+            Actualizar
+          </button>
+          <button type="button" className="btn btn-sm btn-primary" onClick={onStartDigitador}>
+            <Plus size={13} /> Nueva digitación
+          </button>
+        </div>
+      </div>
+
+      {error && (
+        <div className="wiz-toast wiz-toast--error" style={{ position: "static", marginBottom: 12 }}>
+          <AlertCircle size={15} />
+          <span>{error}</span>
+        </div>
+      )}
+
+      {selected?.error && (
+        <div className="wiz-toast wiz-toast--error" style={{ position: "static", marginBottom: 12 }}>
+          <AlertCircle size={15} />
+          <span>{selected.error}</span>
+        </div>
+      )}
+
+      {dlStatus && (
+        <div className="wiz-toast" style={{ position: "static", marginBottom: 12 }}>
+          <CheckCircle2 size={15} />
+          <span>{dlStatus}</span>
+        </div>
+      )}
+
+      {selected && (
+        <>
+          <div className="wiz-stats-row" style={{ marginBottom: 16, flexShrink: 0 }}>
+            <div className="wiz-stat">
+              <span className="wiz-stat-value">{exitosos}</span>
+              <span className="wiz-stat-label">OK</span>
+            </div>
+            <div className="wiz-stat">
+              <span className="wiz-stat-value">{fallidos}</span>
+              <span className="wiz-stat-label">Fallidos</span>
+            </div>
+            <div className="wiz-stat">
+              <span className="wiz-stat-value">{existentes}</span>
+              <span className="wiz-stat-label">Existentes</span>
+            </div>
+            <div className="wiz-stat">
+              <span className="wiz-stat-value">{timeouts}</span>
+              <span className="wiz-stat-label">Timeouts</span>
+            </div>
+            <div className="wiz-stat">
+              <span className="wiz-stat-value">{total || "—"}</span>
+              <span className="wiz-stat-label">Total</span>
+            </div>
+          </div>
+
+          <p className="wiz-hint" style={{ marginBottom: 12, flexShrink: 0 }}>
+            {selected.shard_count && Number(selected.shard_count) > 1
+              ? `${selected.shard_count} bots (distribuido)`
+              : (
+                <>
+                  Worker <strong>{selected.worker_id ?? "—"}</strong>
+                </>
+              )}
+            {selected.original_filename ? ` · ${selected.original_filename}` : ""}
+            {selected.reports_prefix ? (
+              <>
+                {" "}
+                · Prefijo <code className="dl-prefix-code">{selected.reports_prefix}</code>
+              </>
+            ) : null}
+          </p>
+
+          {(() => {
+            const shardList = Array.isArray(selected.shards)
+              ? selected.shards
+              : Array.isArray(result.shards)
+                ? (result.shards as Array<{
+                    id?: string;
+                    worker_id?: string | null;
+                    status?: string;
+                    shard_index?: number | null;
+                  }>)
+                : [];
+            if (shardList.length === 0) return null;
+            return (
+              <div className="wiz-section" style={{ marginBottom: 12, flexShrink: 0 }}>
+                <h3 className="wiz-section-title">Estado por bot</h3>
+                <div className="wiz-review-grid" style={{ marginTop: 8 }}>
+                  {shardList.map((s, i) => (
+                    <div key={s.id || String(i)} className="wiz-review-item">
+                      <div>
+                        <span className="wiz-review-label">
+                          Shard {(s.shard_index ?? i) + 1}
+                        </span>
+                        <span className="wiz-review-value">{s.worker_id || "—"}</span>
+                      </div>
+                      <div>
+                        <span className="wiz-review-label">Estado</span>
+                        <span className="wiz-review-value">{s.status || "—"}</span>
+                      </div>
+                    </div>
+                  ))}
+                </div>
+              </div>
+            );
+          })()}
+
+          <div className="wiz-section dl-files-section">
+            <div className="dl-files-header">
+              <h3 className="wiz-section-title" style={{ margin: 0 }}>
+                Reportes en MinIO
+                {files.length > 0 ? ` · ${files.length} archivo${files.length === 1 ? "" : "s"}` : ""}
+              </h3>
+            </div>
+
+            {filesError && (
+              <div className="wiz-toast wiz-toast--error" style={{ position: "static", marginBottom: 12 }}>
+                <AlertCircle size={15} />
+                <span>No se pudo listar MinIO: {filesError}</span>
+              </div>
+            )}
+
+            {!filesError && files.length === 0 ? (
+              <p className="wiz-hint">Sin reportes listados aún (RESULTADOS / OMITIDOS).</p>
+            ) : (
+              <table className="dl-files-table">
+                <thead>
+                  <tr>
+                    <th>Archivo</th>
+                    <th className="dl-files-table-actions"> </th>
+                  </tr>
+                </thead>
+                <tbody>
+                  {files.map((f) => {
+                    const fileBusy = dlBusyKey === `file:${f.key}`;
+                    const name = reportBasename(f);
+                    return (
+                      <tr key={f.key}>
+                        <td className="dl-col-file" title={f.key}>
+                          {name}
+                        </td>
+                        <td className="dl-files-table-actions">
+                          <button
+                            type="button"
+                            className="btn btn-sm btn-ghost"
+                            disabled={bulkBusy}
+                            onClick={() => void downloadOne(f)}
+                            title="Descargar reporte"
+                          >
+                            {fileBusy ? (
+                              <Loader2 size={13} className="spin" />
+                            ) : (
+                              <Download size={13} />
+                            )}
+                            Descargar
+                          </button>
+                        </td>
+                      </tr>
+                    );
+                  })}
+                </tbody>
+              </table>
+            )}
+          </div>
+
+          <div className="wiz-section">
+            <h3 className="wiz-section-title">Log</h3>
+            <div className="wiz-estimate-box dl-log-box">
+              {events.length === 0 ? (
+                <span className="wiz-hint">Sin eventos.</span>
+              ) : (
+                events.map((ev) => (
+                  <div key={ev.id}>
+                    [{ev.timestamp}] {ev.message}
+                  </div>
+                ))
+              )}
+            </div>
+          </div>
+        </>
+      )}
+    </div>
+  );
+}

services/desktop_app/src/renderer/modules/digitador/DigitadorWizardModule.tsx

@@ -0,0 +1,760 @@
+import {
+  AlertCircle,
+  Bot,
+  CheckCircle2,
+  FileSpreadsheet,
+  Keyboard,
+  Loader2,
+  Play,
+  RefreshCw,
+  Square,
+  Upload,
+  Users,
+} from "lucide-react";
+import { useCallback, useEffect, useMemo, useRef, useState } from "react";
+import type { ApiClient } from "../../lib/apiClient";
+import type {
+  DigitadorJob,
+  DigitadorJobEvent,
+  DigitadorWorker,
+  Job,
+  PoblacionAptaJob,
+} from "../../../shared/types";
+
+export type DigitadorSourceMode = "aptos" | "archivo";
+
+export type DigitadorInitialSource = {
+  poblacionAptaJobId: string;
+};
+
+interface Props {
+  client: ApiClient;
+  job: Job;
+  initialSource?: DigitadorInitialSource | null;
+}
+
+type DigitadorStep = 1 | 2;
+
+type ShardInfo = {
+  id?: string;
+  worker_id?: string | null;
+  status?: string;
+  shard_index?: number | null;
+  result?: Record<string, unknown>;
+  error?: string | null;
+};
+
+export function DigitadorWizardModule({ client, job, initialSource = null }: Props) {
+  const [step, setStep] = useState<DigitadorStep>(1);
+  const [workers, setWorkers] = useState<DigitadorWorker[]>([]);
+  const [selectedIds, setSelectedIds] = useState<string[]>([]);
+  const [useAllOnline, setUseAllOnline] = useState(true);
+  const [sourceMode, setSourceMode] = useState<DigitadorSourceMode>(
+    initialSource?.poblacionAptaJobId ? "aptos" : "aptos",
+  );
+  const [paJobs, setPaJobs] = useState<PoblacionAptaJob[]>([]);
+  const [selectedPaId, setSelectedPaId] = useState(initialSource?.poblacionAptaJobId ?? "");
+  const [loadingPa, setLoadingPa] = useState(false);
+  const [file, setFile] = useState<File | null>(null);
+  const [title, setTitle] = useState("");
+  const [busy, setBusy] = useState(false);
+  const [refreshing, setRefreshing] = useState(false);
+  const [error, setError] = useState<string | null>(null);
+  const [activeJob, setActiveJob] = useState<DigitadorJob | null>(null);
+  const [events, setEvents] = useState<DigitadorJobEvent[]>([]);
+  const sinceRef = useRef(0);
+  const pollRef = useRef<ReturnType<typeof setInterval> | null>(null);
+  const fileInputRef = useRef<HTMLInputElement | null>(null);
+  const initializedSelectRef = useRef(false);
+  const appliedInitialRef = useRef<string | null>(null);
+
+  const onlineWorkers = useMemo(
+    () => workers.filter((w) => w.status === "online" || w.status === "busy"),
+    [workers],
+  );
+
+  useEffect(() => {
+    if (!initialSource?.poblacionAptaJobId) return;
+    const key = initialSource.poblacionAptaJobId;
+    if (appliedInitialRef.current === key) return;
+    appliedInitialRef.current = key;
+    setSourceMode("aptos");
+    setSelectedPaId(initialSource.poblacionAptaJobId);
+    setStep(1);
+  }, [initialSource?.poblacionAptaJobId]);
+
+  const refreshWorkers = useCallback(async () => {
+    setRefreshing(true);
+    try {
+      const list = await client.listDigitadorWorkers();
+      setWorkers(list);
+      setError(null);
+      const online = list.filter((w) => w.status === "online" || w.status === "busy");
+      if (!initializedSelectRef.current && online.length > 0) {
+        initializedSelectRef.current = true;
+        setUseAllOnline(online.length >= 2);
+        setSelectedIds(online.map((w) => w.id));
+      } else if (useAllOnline) {
+        setSelectedIds(online.map((w) => w.id));
+      } else {
+        setSelectedIds((prev) => prev.filter((id) => online.some((w) => w.id === id)));
+      }
+    } catch (e) {
+      setError(e instanceof Error ? e.message : String(e));
+    } finally {
+      setRefreshing(false);
+    }
+  }, [client, useAllOnline]);
+
+  useEffect(() => {
+    void refreshWorkers();
+    const t = setInterval(() => void refreshWorkers(), 5000);
+    return () => clearInterval(t);
+  }, [refreshWorkers]);
+
+  const loadPaJobs = useCallback(async () => {
+    setLoadingPa(true);
+    try {
+      const list = await client.poblacionAptaJobs(job.id);
+      const completed = list
+        .filter((j) => j.status === "completed")
+        .sort((a, b) => new Date(b.created_at).getTime() - new Date(a.created_at).getTime());
+      setPaJobs(completed);
+      setSelectedPaId((prev) => {
+        if (prev && completed.some((j) => j.id === prev)) return prev;
+        return completed[0]?.id ?? "";
+      });
+    } catch (e) {
+      setError(e instanceof Error ? e.message : String(e));
+    } finally {
+      setLoadingPa(false);
+    }
+  }, [client, job.id]);
+
+  useEffect(() => {
+    if (sourceMode !== "aptos") return;
+    void loadPaJobs();
+  }, [sourceMode, loadPaJobs]);
+
+  useEffect(() => {
+    if (!activeJob || ["completed", "failed", "cancelled"].includes(activeJob.status)) {
+      if (pollRef.current) {
+        clearInterval(pollRef.current);
+        pollRef.current = null;
+      }
+      return;
+    }
+
+    async function poll() {
+      if (!activeJob) return;
+      try {
+        const nextJob = await client.getDigitadorJob(activeJob.id);
+        setActiveJob(nextJob);
+        const { events: next } = await client.getDigitadorEvents(activeJob.id, sinceRef.current);
+        if (next.length) {
+          setEvents((prev) => [...prev, ...next]);
+          sinceRef.current = next[next.length - 1].id;
+        }
+      } catch (e) {
+        setError(e instanceof Error ? e.message : String(e));
+      }
+    }
+
+    void poll();
+    pollRef.current = setInterval(() => void poll(), 3000);
+    return () => {
+      if (pollRef.current) clearInterval(pollRef.current);
+    };
+  }, [activeJob?.id, activeJob?.status, client]);
+
+  function toggleWorker(id: string) {
+    if (useAllOnline) return;
+    setSelectedIds((prev) =>
+      prev.includes(id) ? prev.filter((x) => x !== id) : [...prev, id],
+    );
+  }
+
+  const selectedPa = paJobs.find((j) => j.id === selectedPaId) ?? null;
+  const aptosCount = Number(selectedPa?.summary?.paso_bot ?? 0);
+
+  async function startJob() {
+    const targets = useAllOnline ? onlineWorkers.map((w) => w.id) : selectedIds;
+    if (targets.length === 0) {
+      setError("Necesitas al menos un bot online.");
+      return;
+    }
+    if (sourceMode === "archivo") {
+      if (!file) {
+        setError("Selecciona un CSV.");
+        return;
+      }
+    } else {
+      if (!selectedPaId) {
+        setError("Selecciona una corrida de población apta.");
+        return;
+      }
+      if (aptosCount <= 0) {
+        setError("Esa corrida no tiene aptos (PASO_BOT) para digitar.");
+        return;
+      }
+    }
+
+    setBusy(true);
+    setError(null);
+    try {
+      const form = new FormData();
+      form.append("job_id", job.id);
+      if (sourceMode === "archivo" && file) {
+        form.append("file", file);
+      } else {
+        form.append("poblacion_apta_job_id", selectedPaId);
+        form.append("subset", "aptos");
+      }
+      if (title.trim()) form.append("title", title.trim());
+      if (targets.length === 1) {
+        form.append("worker_id", targets[0]);
+      } else {
+        form.append("worker_ids", JSON.stringify(targets));
+      }
+      const created = await client.createDigitadorJob(form);
+      sinceRef.current = 0;
+      setEvents([]);
+      setActiveJob(created);
+      setStep(2);
+    } catch (e) {
+      setError(e instanceof Error ? e.message : String(e));
+    } finally {
+      setBusy(false);
+    }
+  }
+
+  async function cancelJob() {
+    if (!activeJob) return;
+    setBusy(true);
+    try {
+      const updated = await client.cancelDigitadorJob(activeJob.id);
+      setActiveJob(updated);
+    } catch (e) {
+      setError(e instanceof Error ? e.message : String(e));
+    } finally {
+      setBusy(false);
+    }
+  }
+
+  function resetWizard() {
+    setStep(1);
+    setActiveJob(null);
+    setEvents([]);
+    setFile(null);
+    setTitle("");
+    setError(null);
+    sinceRef.current = 0;
+    if (fileInputRef.current) fileInputRef.current.value = "";
+  }
+
+  const result = activeJob?.result ?? {};
+  const processed = Number(result.processed ?? result.exitosos ?? 0);
+  const total = Number(result.total ?? result.total_rows ?? 0);
+  const pct = total > 0 ? Math.min(100, Math.round((processed / total) * 100)) : 0;
+  const running = Boolean(
+    activeJob && !["completed", "failed", "cancelled"].includes(activeJob.status),
+  );
+  const isDistributed =
+    Boolean(activeJob?.shard_count && Number(activeJob.shard_count) > 1) ||
+    result.mode === "distributed";
+  const shards: ShardInfo[] = Array.isArray(activeJob?.shards)
+    ? (activeJob!.shards as ShardInfo[])
+    : Array.isArray(result.shards)
+      ? (result.shards as ShardInfo[])
+      : [];
+  const effectiveTargets = useAllOnline ? onlineWorkers.map((w) => w.id) : selectedIds;
+  const sourceReady =
+    sourceMode === "archivo" ? Boolean(file) : Boolean(selectedPaId) && aptosCount > 0;
+  const canStart = effectiveTargets.length > 0 && sourceReady && !busy;
+
+  return (
+    <div className="wiz">
+      <nav className="stepper" aria-label="Pasos del bot digitador">
+        <div className="stepper-item">
+          <button
+            type="button"
+            className={`stepper-dot ${step === 1 ? "active" : step > 1 ? "done" : "todo"}`}
+            onClick={() => !running && setStep(1)}
+            disabled={running}
+          >
+            1
+          </button>
+          <span className={`stepper-label ${step === 1 ? "active" : step > 1 ? "done" : "todo"}`}>
+            Configurar
+          </span>
+          <div className={`stepper-connector ${step > 1 ? "done" : ""}`} />
+        </div>
+        <div className="stepper-item">
+          <button
+            type="button"
+            className={`stepper-dot ${step === 2 ? "active" : "todo"}`}
+            onClick={() => activeJob && setStep(2)}
+            disabled={!activeJob}
+          >
+            2
+          </button>
+          <span className={`stepper-label ${step === 2 ? "active" : "todo"}`}>Ejecución</span>
+        </div>
+      </nav>
+
+      <div className="wiz-body">
+        {step === 1 && (
+          <div className="wiz-step">
+            <div className="wiz-header" style={{ marginBottom: 4 }}>
+              <h2 style={{ display: "flex", alignItems: "center", gap: 8, margin: 0 }}>
+                <Keyboard size={18} />
+                Bot digitador
+              </h2>
+              <p className="wiz-desc" style={{ marginTop: 6 }}>
+                Digita en Integración Total solo los <strong>aptos (PASO_BOT)</strong>, o un CSV
+                propio con el mismo formato.
+              </p>
+            </div>
+
+            {/* 1. Origen */}
+            <div className="wiz-section">
+              <h3 className="wiz-section-title">1. Origen</h3>
+              <div style={{ display: "flex", flexWrap: "wrap", gap: 8, marginBottom: 14 }}>
+                <button
+                  type="button"
+                  className={`btn${sourceMode === "aptos" ? " btn-primary" : ""}`}
+                  onClick={() => setSourceMode("aptos")}
+                >
+                  <Users size={14} /> Aptos de población apta
+                </button>
+                <button
+                  type="button"
+                  className={`btn${sourceMode === "archivo" ? " btn-primary" : ""}`}
+                  onClick={() => setSourceMode("archivo")}
+                >
+                  <Upload size={14} /> Subir CSV
+                </button>
+              </div>
+
+              <label className="wiz-field-label">Título (opcional)</label>
+              <input
+                className="form-input"
+                value={title}
+                onChange={(e) => setTitle(e.target.value)}
+                placeholder="Ej. Digitación lote mayo 2026"
+                style={{ width: "100%", marginBottom: 14 }}
+              />
+
+              {sourceMode === "aptos" ? (
+                <>
+                  <div className="wiz-section-header" style={{ marginBottom: 8 }}>
+                    <label className="wiz-field-label" style={{ margin: 0 }}>
+                      Corrida de población apta
+                    </label>
+                    <button
+                      type="button"
+                      className="btn btn-sm btn-ghost"
+                      onClick={() => void loadPaJobs()}
+                      disabled={loadingPa}
+                    >
+                      {loadingPa ? <Loader2 size={13} className="spin" /> : <RefreshCw size={13} />}
+                      Actualizar
+                    </button>
+                  </div>
+
+                  {loadingPa && paJobs.length === 0 ? (
+                    <p className="wiz-hint">
+                      <Loader2 size={13} className="spin" /> Cargando corridas…
+                    </p>
+                  ) : paJobs.length === 0 ? (
+                    <div className="wiz-estimate-box" style={{ textAlign: "left" }}>
+                      <p className="wiz-desc" style={{ margin: 0 }}>
+                        No hay corridas de población apta completadas en este trabajo.
+                      </p>
+                      <p className="wiz-hint" style={{ marginTop: 6, marginBottom: 0 }}>
+                        Ejecuta primero Población apta, o cambia a «Subir CSV».
+                      </p>
+                    </div>
+                  ) : (
+                    <>
+                      <select
+                        className="form-input"
+                        value={selectedPaId}
+                        onChange={(e) => setSelectedPaId(e.target.value)}
+                        style={{ width: "100%" }}
+                      >
+                        {paJobs.map((pa) => {
+                          const n = Number(pa.summary?.paso_bot ?? 0);
+                          return (
+                            <option key={pa.id} value={pa.id}>
+                              {pa.id.slice(0, 8)} · {n} aptos · {pa.summary?.total ?? 0} total
+                            </option>
+                          );
+                        })}
+                      </select>
+
+                      {selectedPa && (
+                        <div
+                          className="wiz-estimate-box"
+                          style={{
+                            marginTop: 12,
+                            display: "flex",
+                            alignItems: "center",
+                            gap: 12,
+                            textAlign: "left",
+                          }}
+                        >
+                          {aptosCount > 0 ? (
+                            <CheckCircle2 size={20} style={{ color: "var(--success, #16a34a)", flexShrink: 0 }} />
+                          ) : (
+                            <AlertCircle size={20} style={{ color: "var(--danger, #dc2626)", flexShrink: 0 }} />
+                          )}
+                          <div>
+                            <div style={{ fontWeight: 600 }}>
+                              {aptosCount > 0
+                                ? `${aptosCount} registros aptos listos para digitar`
+                                : "Sin aptos en esta corrida"}
+                            </div>
+                            <div className="wiz-hint" style={{ margin: 0 }}>
+                              Solo se envían filas PASO_BOT. Los no aptos no se digitan.
+                            </div>
+                          </div>
+                        </div>
+                      )}
+                    </>
+                  )}
+                </>
+              ) : (
+                <>
+                  <label className="wiz-field-label">Archivo CSV</label>
+                  <div className="folder-picker">
+                    <div className="folder-picker-display">
+                      {file ? (
+                        <>
+                          <FileSpreadsheet size={16} className="folder-picker-icon" />
+                          <span className="folder-picker-path">{file.name}</span>
+                        </>
+                      ) : (
+                        <span className="folder-picker-placeholder">Ningún archivo seleccionado</span>
+                      )}
+                    </div>
+                    <button
+                      type="button"
+                      className="btn btn-primary"
+                      onClick={() => fileInputRef.current?.click()}
+                    >
+                      <FileSpreadsheet size={15} />
+                      Seleccionar CSV
+                    </button>
+                    <input
+                      ref={fileInputRef}
+                      type="file"
+                      accept=".csv,text/csv"
+                      hidden
+                      onChange={(e) => setFile(e.target.files?.[0] ?? null)}
+                    />
+                  </div>
+                  <p className="wiz-hint" style={{ marginTop: 8 }}>
+                    Columnas: radicado, documento, ingreso, total_cargos, autorizacion,
+                    tipo_documento, tipo_servicio, copago, cuota_moderadora (también acepta
+                    headers IT).
+                  </p>
+                </>
+              )}
+            </div>
+
+            {/* 2. Bots */}
+            <div className="wiz-section">
+              <div className="wiz-section-header">
+                <h3 className="wiz-section-title">
+                  2. Bots
+                  {onlineWorkers.length > 0 && (
+                    <span className="wiz-count-badge">{onlineWorkers.length} online</span>
+                  )}
+                </h3>
+                <button
+                  type="button"
+                  className="btn btn-sm btn-ghost"
+                  onClick={() => void refreshWorkers()}
+                  disabled={refreshing}
+                >
+                  {refreshing ? <Loader2 size={13} className="spin" /> : <RefreshCw size={13} />}
+                  Actualizar
+                </button>
+              </div>
+              <p className="wiz-desc">
+                El agent debe estar encendido en la PC con Integración Total. Con varios bots, las
+                filas se reparte automáticamente.
+              </p>
+
+              {onlineWorkers.length >= 2 && (
+                <label
+                  style={{
+                    display: "flex",
+                    alignItems: "center",
+                    gap: 8,
+                    marginTop: 12,
+                    marginBottom: 8,
+                    cursor: "pointer",
+                  }}
+                >
+                  <input
+                    type="checkbox"
+                    checked={useAllOnline}
+                    onChange={(e) => {
+                      const on = e.target.checked;
+                      setUseAllOnline(on);
+                      if (on) setSelectedIds(onlineWorkers.map((w) => w.id));
+                    }}
+                  />
+                  <span>Usar todos los bots online ({onlineWorkers.length})</span>
+                </label>
+              )}
+
+              {workers.length === 0 ? (
+                <div className="wiz-estimate-box" style={{ marginTop: 12, textAlign: "left" }}>
+                  <div style={{ display: "flex", alignItems: "flex-start", gap: 10 }}>
+                    <Bot size={18} style={{ flexShrink: 0, marginTop: 2 }} />
+                    <div>
+                      <div style={{ fontWeight: 600 }}>Ningún bot conectado</div>
+                      <p className="wiz-hint" style={{ margin: "4px 0 0" }}>
+                        En la PC del digitador ejecuta <code>run_gui.py</code> del agent y conéctalo
+                        al gateway.
+                      </p>
+                    </div>
+                  </div>
+                </div>
+              ) : (
+                <div className="wiz-review-grid" style={{ marginTop: 12 }}>
+                  {workers.map((w) => {
+                    const online = w.status === "online" || w.status === "busy";
+                    const selected = effectiveTargets.includes(w.id);
+                    return (
+                      <button
+                        key={w.id}
+                        type="button"
+                        className={`wiz-review-item${selected ? " wiz-review-item--selected" : ""}`}
+                        disabled={!online || useAllOnline}
+                        onClick={() => online && toggleWorker(w.id)}
+                        style={{
+                          textAlign: "left",
+                          cursor: online && !useAllOnline ? "pointer" : "not-allowed",
+                          opacity: online ? 1 : 0.45,
+                          borderColor: selected ? "var(--accent, #3b8ed0)" : undefined,
+                        }}
+                      >
+                        <div>
+                          <span className="wiz-review-label">Worker</span>
+                          <span className="wiz-review-value">{w.id}</span>
+                        </div>
+                        <div>
+                          <span className="wiz-review-label">Estado</span>
+                          <span className="wiz-review-value">
+                            {w.status}
+                            {w.hostname ? ` · ${w.hostname}` : ""}
+                            {selected ? " · seleccionado" : ""}
+                          </span>
+                        </div>
+                      </button>
+                    );
+                  })}
+                </div>
+              )}
+              {onlineWorkers.length === 0 && workers.length > 0 && (
+                <p className="wiz-hint" style={{ marginTop: 8 }}>
+                  Hay workers registrados pero ninguno online. Revisa el agent.
+                </p>
+              )}
+              {effectiveTargets.length > 1 && (
+                <p className="wiz-hint" style={{ marginTop: 8 }}>
+                  Se repartirá entre {effectiveTargets.length} bots.
+                </p>
+              )}
+            </div>
+          </div>
+        )}
+
+        {step === 2 && activeJob && (
+          <div className="wiz-step wiz-step-centered">
+            <div className="wiz-hero">
+              <div className={`wiz-hero-icon ${running ? "wiz-hero-icon--go" : ""}`}>
+                {running ? (
+                  <Loader2 size={28} className="spin" />
+                ) : activeJob.status === "completed" ? (
+                  <Keyboard size={28} />
+                ) : activeJob.status === "failed" ? (
+                  <AlertCircle size={28} />
+                ) : (
+                  <Bot size={28} />
+                )}
+              </div>
+              <h3 className="wiz-hero-title">
+                {running
+                  ? isDistributed
+                    ? "Digitación distribuida en curso"
+                    : "Digitación en curso"
+                  : activeJob.status === "completed"
+                    ? "Digitación completada"
+                    : activeJob.status === "cancelled"
+                      ? "Digitación cancelada"
+                      : activeJob.status === "failed"
+                        ? "Digitación fallida"
+                        : "Digitación finalizada"}
+              </h3>
+              <p className="wiz-desc">
+                {isDistributed
+                  ? `${Number(activeJob.shard_count) || shards.length || "?"} bots`
+                  : `Bot ${activeJob.worker_id ?? "—"}`}
+                {activeJob.original_filename ? ` · ${activeJob.original_filename}` : ""}
+              </p>
+            </div>
+
+            {activeJob.error && (
+              <div className="wiz-toast wiz-toast--error" style={{ position: "static", marginBottom: 12 }}>
+                <AlertCircle size={15} />
+                <span>{activeJob.error}</span>
+              </div>
+            )}
+
+            <div className="wiz-progress">
+              <div className="wiz-progress-header">
+                <span>
+                  {Number(result.exitosos ?? processed)}/{total || "?"} filas
+                </span>
+                <span>{pct}%</span>
+              </div>
+              <div className="wiz-progress-track">
+                <div className="wiz-progress-fill" style={{ width: `${pct}%` }} />
+              </div>
+            </div>
+
+            <div className="wiz-stats-row">
+              <div className="wiz-stat">
+                <span className="wiz-stat-value">{String(result.exitosos ?? processed)}</span>
+                <span className="wiz-stat-label">OK</span>
+              </div>
+              <div className="wiz-stat">
+                <span className="wiz-stat-value">{String(result.fallidos ?? 0)}</span>
+                <span className="wiz-stat-label">Fallidos</span>
+              </div>
+              <div className="wiz-stat">
+                <span className="wiz-stat-value">{String(result.existentes ?? 0)}</span>
+                <span className="wiz-stat-label">Existentes</span>
+              </div>
+              <div className="wiz-stat">
+                <span className="wiz-stat-value">{String(result.timeouts ?? 0)}</span>
+                <span className="wiz-stat-label">Timeouts</span>
+              </div>
+            </div>
+
+            {shards.length > 0 && (
+              <div className="wiz-section" style={{ marginTop: 16, width: "100%", maxWidth: 640 }}>
+                <h3 className="wiz-section-title">Shards por bot</h3>
+                <div className="wiz-review-grid" style={{ marginTop: 8 }}>
+                  {shards.map((s, i) => (
+                    <div key={s.id || String(i)} className="wiz-review-item">
+                      <div>
+                        <span className="wiz-review-label">
+                          Shard {(s.shard_index ?? i) + 1}
+                        </span>
+                        <span className="wiz-review-value">{s.worker_id || "—"}</span>
+                      </div>
+                      <div>
+                        <span className="wiz-review-label">Estado</span>
+                        <span className="wiz-review-value">{s.status || "—"}</span>
+                      </div>
+                    </div>
+                  ))}
+                </div>
+              </div>
+            )}
+
+            <div className="wiz-section" style={{ marginTop: 16, width: "100%", maxWidth: 640 }}>
+              <h3 className="wiz-section-title">Log del agent</h3>
+              <div
+                className="wiz-estimate-box"
+                style={{
+                  maxHeight: 220,
+                  overflow: "auto",
+                  fontFamily: "ui-monospace, SFMono-Regular, Menlo, monospace",
+                  fontSize: 12,
+                  textAlign: "left",
+                }}
+              >
+                {events.length === 0 ? (
+                  <span className="wiz-hint">Esperando eventos…</span>
+                ) : (
+                  events.map((ev) => (
+                    <div key={ev.id}>
+                      [{ev.timestamp}] {ev.message}
+                    </div>
+                  ))
+                )}
+              </div>
+            </div>
+          </div>
+        )}
+      </div>
+
+      {error && (
+        <div className="wiz-toast wiz-toast--error">
+          <AlertCircle size={15} />
+          <span>{error}</span>
+        </div>
+      )}
+
+      <footer className="wiz-nav">
+        {step === 2 ? (
+          <button type="button" className="btn" onClick={resetWizard} disabled={running}>
+            Nueva digitación
+          </button>
+        ) : (
+          <span />
+        )}
+
+        <div className="wiz-nav-dots">
+          <button
+            type="button"
+            className={`wiz-nav-dot ${step === 1 ? "active" : ""}`}
+            onClick={() => !running && setStep(1)}
+            disabled={running}
+            aria-label="Paso 1"
+          />
+          <button
+            type="button"
+            className={`wiz-nav-dot ${step === 2 ? "active" : ""}`}
+            onClick={() => activeJob && setStep(2)}
+            disabled={!activeJob}
+            aria-label="Paso 2"
+          />
+        </div>
+
+        {step === 1 && (
+          <button
+            type="button"
+            className="btn btn-primary"
+            disabled={!canStart}
+            onClick={() => void startJob()}
+          >
+            {busy ? <Loader2 size={15} className="spin" /> : <Play size={15} />}
+            {effectiveTargets.length > 1
+              ? `Digitar con ${effectiveTargets.length} bots`
+              : sourceMode === "aptos" && aptosCount > 0
+                ? `Digitar ${aptosCount} aptos`
+                : "Iniciar digitación"}
+          </button>
+        )}
+        {step === 2 && running && (
+          <button type="button" className="btn" disabled={busy} onClick={() => void cancelJob()}>
+            <Square size={14} /> Cancelar
+          </button>
+        )}
+        {step === 2 && !running && (
+          <button type="button" className="btn btn-primary" onClick={resetWizard}>
+            <Play size={15} /> Otra digitación
+          </button>
+        )}
+      </footer>
+    </div>
+  );
+}

services/desktop_app/src/renderer/modules/poblacion_apta/PoblacionAptaResults.tsx

@@ -1,4 +1,4 @@
-import { Download, Loader2, RefreshCw, Users } from "lucide-react";
+import { Download, Keyboard, Loader2, RefreshCw, Users } from "lucide-react";
 import { useEffect, useState } from "react";
 import type { PoblacionAptaJob } from "../../../shared/types";
 import { PoblacionAptaResultsTable } from "../../components/PoblacionAptaResultsTable";
@@ -7,7 +7,7 @@ import type { ModuleResultsProps } from "../registry";
 
 type Row = Record<string, unknown>;
 
-export function PoblacionAptaResults({ client, task, job }: ModuleResultsProps) {
+export function PoblacionAptaResults({ client, task, job, onActivateModule }: ModuleResultsProps) {
   const [paJob, setPaJob] = useState<PoblacionAptaJob | null>(null);
   const [rows, setRows] = useState<Row[]>([]);
   const [loading, setLoading] = useState(true);
@@ -60,6 +60,16 @@ export function PoblacionAptaResults({ client, task, job }: ModuleResultsProps)
     downloadCsv(buildExportCsvFilename("poblacion_apta_no_aptos", job.title), text);
   }
 
+  async function exportTodos() {
+    const paJobId = String(
+      (task.result as Record<string, unknown>)?.poblacion_apta_job_id ?? "",
+    );
+    if (!paJobId) return;
+    const blob = await client.poblacionAptaExportCsv(paJobId, "todos");
+    const text = await blob.text();
+    downloadCsv(buildExportCsvFilename("poblacion_apta_todos", job.title), text);
+  }
+
   if (loading) {
     return (
       <div className="module-loading">
@@ -88,6 +98,10 @@ export function PoblacionAptaResults({ client, task, job }: ModuleResultsProps)
   }
 
   const summary = paJob?.summary ?? {};
+  const paJobId = String(
+    (task.result as Record<string, unknown>)?.poblacion_apta_job_id ?? paJob?.id ?? "",
+  );
+  const canDigitador = paJob?.status === "completed" && Boolean(paJobId) && Boolean(onActivateModule);
 
   return (
     <div className="module-results">
@@ -108,6 +122,23 @@ export function PoblacionAptaResults({ client, task, job }: ModuleResultsProps)
           <button type="button" className="btn btn-sm" onClick={() => void exportNoAptos()}>
             <Download size={13} /> CSV no aptos ({summary.con_error ?? 0})
           </button>
+          <button type="button" className="btn btn-sm" onClick={() => void exportTodos()}>
+            <Download size={13} /> CSV todos ({summary.total ?? 0})
+          </button>
+          {canDigitador && (
+            <button
+              type="button"
+              className="btn btn-sm btn-primary"
+              disabled={Number(summary.paso_bot ?? 0) === 0}
+              onClick={() =>
+                onActivateModule?.("digitador", {
+                  poblacionAptaJobId: paJobId,
+                })
+              }
+            >
+              <Keyboard size={13} /> Digitar aptos
+            </button>
+          )}
         </div>
       </div>
 

services/desktop_app/src/renderer/modules/poblacion_apta/PoblacionAptaWizard.tsx

@@ -3,6 +3,7 @@ import {
   CheckSquare,
   Download,
   FileSpreadsheet,
+  Keyboard,
   Loader2,
   Square,
   Upload,
@@ -235,6 +236,13 @@ export function PoblacionAptaWizard({
     downloadCsv(buildExportCsvFilename("poblacion_apta_no_aptos", job.title), text);
   }
 
+  async function exportTodos() {
+    if (!resultJob) return;
+    const blob = await client.poblacionAptaExportCsv(resultJob.id, "todos");
+    const text = await blob.text();
+    downloadCsv(buildExportCsvFilename("poblacion_apta_todos", job.title), text);
+  }
+
   function reset() {
     setDone(false);
     setResultJob(null);
@@ -271,6 +279,23 @@ export function PoblacionAptaWizard({
             <button type="button" className="btn btn-sm" onClick={() => void exportNoAptos()}>
               <Download size={13} /> CSV no aptos ({summary.con_error ?? 0})
             </button>
+            <button type="button" className="btn btn-sm" onClick={() => void exportTodos()}>
+              <Download size={13} /> CSV todos ({summary.total ?? 0})
+            </button>
+            {onActivateModule && (
+              <button
+                type="button"
+                className="btn btn-sm btn-primary"
+                disabled={Number(summary.paso_bot ?? 0) === 0}
+                onClick={() =>
+                  onActivateModule("digitador", {
+                    poblacionAptaJobId: resultJob.id,
+                  })
+                }
+              >
+                <Keyboard size={13} /> Digitar aptos
+              </button>
+            )}
             <button type="button" className="btn btn-sm" onClick={reset}>
               Nueva corrida
             </button>

services/desktop_app/src/renderer/types/jobSection.ts

@@ -4,4 +4,5 @@ export type JobSection =
   | "cleaning"
   | "poblacion_apta"
   | "downloads"
+  | "digitador"
   | "results";

services/desktop_app/src/renderer/views/JobResultsView.tsx

@@ -5,6 +5,7 @@ import {
   Cpu,
   Download,
   Hash,
+  Keyboard,
   Loader2,
   Plus,
   RefreshCw,
@@ -29,8 +30,9 @@ import type { ApiClient } from "../lib/apiClient";
 import { downloadCsv, toCsv, buildExportCsvFilename } from "../lib/csv";
 import { taskTimingLabel, timingDuration } from "../lib/time";
 import { DownloadsResultsPanel } from "../modules/downloads/DownloadsResultsPanel";
+import { DigitadorResultsPanel } from "../modules/digitador/DigitadorResultsPanel";
 
-type ResultTab = "counting" | "extraction" | "cleaning" | "poblacion_apta" | "downloads";
+type ResultTab = "counting" | "extraction" | "cleaning" | "poblacion_apta" | "downloads" | "digitador";
 
 interface Props {
   client: ApiClient;
@@ -46,11 +48,15 @@ interface Props {
   onStartCleaning: (ctx?: Record<string, unknown>) => void;
   onStartPoblacionApta: () => void;
   onStartDownload: () => void;
+  onStartDigitador: () => void;
   onContinueToCounting?: (ctx: {
     downloadJobId: string;
     downloadsPrefix?: string;
     files?: DownloadJobFile[];
   }) => void;
+  onContinueToDigitador?: (ctx: {
+    poblacionAptaJobId: string;
+  }) => void;
 }
 
 export function JobResultsView({
@@ -67,7 +73,9 @@ export function JobResultsView({
   onStartCleaning,
   onStartPoblacionApta,
   onStartDownload,
+  onStartDigitador,
   onContinueToCounting,
+  onContinueToDigitador,
 }: Props) {
   const [activeTab, setActiveTab] = useState<ResultTab>("counting");
   const [refreshing, setRefreshing] = useState(false);
@@ -82,6 +90,8 @@ export function JobResultsView({
     .sort((a, b) => new Date(b.created_at).getTime() - new Date(a.created_at).getTime());
   const downloadTasks = tasks.filter((t) => t.type === "downloads")
     .sort((a, b) => new Date(b.created_at).getTime() - new Date(a.created_at).getTime());
+  const digitadorTasks = tasks.filter((t) => t.type === "digitador")
+    .sort((a, b) => new Date(b.created_at).getTime() - new Date(a.created_at).getTime());
 
   async function refresh() {
     setRefreshing(true);
@@ -164,6 +174,17 @@ export function JobResultsView({
             <span className="job-results-tab-count">{downloadTasks.length}</span>
           )}
         </button>
+        <button
+          type="button"
+          className={`job-results-tab${activeTab === "digitador" ? " job-results-tab--active" : ""}`}
+          onClick={() => setActiveTab("digitador")}
+        >
+          <Keyboard size={15} />
+          Bot digitador
+          {digitadorTasks.length > 0 && (
+            <span className="job-results-tab-count">{digitadorTasks.length}</span>
+          )}
+        </button>
         <div style={{ flex: 1 }} />
         <button
           type="button"
@@ -216,6 +237,7 @@ export function JobResultsView({
             tasks={poblacionAptaTasks}
             liveProgress={poblacionAptaProgress}
             onStartPoblacionApta={onStartPoblacionApta}
+            onContinueToDigitador={onContinueToDigitador}
           />
         )}
         {activeTab === "downloads" && (
@@ -227,6 +249,14 @@ export function JobResultsView({
             onContinueToCounting={onContinueToCounting}
           />
         )}
+        {activeTab === "digitador" && (
+          <DigitadorResultsPanel
+            client={client}
+            job={job}
+            tasks={digitadorTasks}
+            onStartDigitador={onStartDigitador}
+          />
+        )}
       </div>
     </div>
   );
@@ -780,12 +810,16 @@ function PoblacionAptaResultsSection({
   tasks,
   liveProgress,
   onStartPoblacionApta,
+  onContinueToDigitador,
 }: {
   client: ApiClient;
   job: Job;
   tasks: Task[];
   liveProgress?: BackgroundProgress | null;
   onStartPoblacionApta: () => void;
+  onContinueToDigitador?: (ctx: {
+    poblacionAptaJobId: string;
+  }) => void;
 }) {
   const [selectedTask, setSelectedTask] = useState<Task | null>(tasks[0] ?? null);
   const [paJob, setPaJob] = useState<PoblacionAptaJob | null>(null);
@@ -853,6 +887,16 @@ function PoblacionAptaResultsSection({
     downloadCsv(buildExportCsvFilename("poblacion_apta_no_aptos", job.title), text);
   }
 
+  async function exportTodos() {
+    const paJobId = String(
+      (selectedTask?.result as Record<string, unknown>)?.poblacion_apta_job_id ?? "",
+    );
+    if (!paJobId) return;
+    const blob = await client.poblacionAptaExportCsv(paJobId, "todos");
+    const text = await blob.text();
+    downloadCsv(buildExportCsvFilename("poblacion_apta_todos", job.title), text);
+  }
+
   if (tasks.length === 0 && !liveProgress?.running) {
     return (
       <div className="job-results-empty">
@@ -874,6 +918,13 @@ function PoblacionAptaResultsSection({
   }
 
   const summary = paJob?.summary ?? {};
+  const paJobId = String(
+    (selectedTask?.result as Record<string, unknown>)?.poblacion_apta_job_id ?? paJob?.id ?? "",
+  );
+  const canDigitador =
+    Boolean(onContinueToDigitador) &&
+    paJob?.status === "completed" &&
+    Boolean(paJobId);
 
   return (
     <div className="results-section">
@@ -917,6 +968,21 @@ function PoblacionAptaResultsSection({
           <button type="button" className="btn btn-sm" onClick={() => void exportNoAptos()}>
             <Download size={13} /> CSV no aptos ({summary.con_error ?? 0})
           </button>
+          <button type="button" className="btn btn-sm" onClick={() => void exportTodos()}>
+            <Download size={13} /> CSV todos ({summary.total ?? 0})
+          </button>
+          {canDigitador && (
+            <button
+              type="button"
+              className="btn btn-sm btn-primary"
+              onClick={() =>
+                onContinueToDigitador?.({ poblacionAptaJobId: paJobId })
+              }
+              disabled={Number(summary.paso_bot ?? 0) === 0}
+            >
+              <Keyboard size={13} /> Digitar aptos
+            </button>
+          )}
           <button type="button" className="btn btn-sm btn-ghost" onClick={onStartPoblacionApta}>
             <Plus size={13} /> Nueva corrida
           </button>

services/desktop_app/src/renderer/views/JobView.tsx

@@ -9,6 +9,7 @@ import { CleaningWizardModule } from "../modules/cleaning/CleaningWizardModule";
 import { PoblacionAptaWizardModule } from "../modules/poblacion_apta/PoblacionAptaWizardModule";
 import { JobExtractionWizard } from "../modules/extraction/JobExtractionWizard";
 import { DownloadsWizardModule } from "../modules/downloads/DownloadsWizardModule";
+import { DigitadorWizardModule } from "../modules/digitador/DigitadorWizardModule";
 import { JobResultsView } from "./JobResultsView";
 
 interface Props {
@@ -127,6 +128,7 @@ export function JobView({ client, job, section, sectionContext, onActivateSectio
           {section === "cleaning" && "Limpieza de datos"}
           {section === "poblacion_apta" && "Población apta"}
           {section === "downloads" && "Bot de descarga"}
+          {section === "digitador" && "Bot digitador"}
           {section === "results" && "Resultados del trabajo"}
         </div>
       </header>
@@ -201,6 +203,19 @@ export function JobView({ client, job, section, sectionContext, onActivateSectio
           <DownloadsWizardModule client={client} job={job} />
         </div>
 
+        <div className={`job-section-pane${section === "digitador" ? "" : " job-section-pane--hidden"}`}>
+          <DigitadorWizardModule
+            key={`digitador-${String(sectionContext.poblacionAptaJobId ?? "default")}`}
+            client={client}
+            job={job}
+            initialSource={
+              typeof sectionContext.poblacionAptaJobId === "string"
+                ? { poblacionAptaJobId: sectionContext.poblacionAptaJobId }
+                : null
+            }
+          />
+        </div>
+
         <div className={`job-section-pane${section === "results" ? "" : " job-section-pane--hidden"}`}>
           <JobResultsView
             client={client}
@@ -216,7 +231,9 @@ export function JobView({ client, job, section, sectionContext, onActivateSectio
             onStartCleaning={(ctx?: Record<string, unknown>) => onActivateSection("cleaning", ctx)}
             onStartPoblacionApta={() => onActivateSection("poblacion_apta")}
             onStartDownload={() => onActivateSection("downloads")}
+            onStartDigitador={() => onActivateSection("digitador")}
             onContinueToCounting={(ctx) => onActivateSection("counting", ctx)}
+            onContinueToDigitador={(ctx) => onActivateSection("digitador", ctx)}
           />
         </div>
       </div>

services/desktop_app/src/shared/types.ts

@@ -373,7 +373,7 @@ export interface DownloadWorker {
 
 export interface DownloadJob {
   id: string;
-  worker_id: string;
+  worker_id: string | null;
   status: "pending" | "running" | "completed" | "failed" | "cancelled" | string;
   title: string;
   excel_key: string;
@@ -389,6 +389,11 @@ export interface DownloadJob {
   job_id?: string | null;
   /** Task type=downloads */
   task_id?: string | null;
+  parent_download_job_id?: string | null;
+  shard_index?: number | null;
+  shard_count?: number | null;
+  /** Presente en GET de padre distribuido */
+  shards?: DownloadJob[];
 }
 
 export interface DownloadJobFile {
@@ -417,6 +422,63 @@ export interface DownloadJobEvent {
   metadata?: Record<string, unknown>;
 }
 
+/** Worker del bot digitador (control remoto). */
+export interface DigitadorWorker {
+  id: string;
+  hostname: string;
+  max_concurrent: number;
+  status: "online" | "offline" | "busy" | string;
+  last_heartbeat: string;
+  current_job_id: string | null;
+}
+
+export interface DigitadorJob {
+  id: string;
+  worker_id: string | null;
+  status: "pending" | "running" | "completed" | "failed" | "cancelled" | string;
+  title: string;
+  csv_key: string;
+  reports_prefix: string;
+  original_filename: string;
+  cancel_requested: boolean;
+  result: Record<string, unknown>;
+  error: string | null;
+  created_at: string;
+  started_at: string | null;
+  finished_at: string | null;
+  job_id?: string | null;
+  task_id?: string | null;
+  parent_digitador_job_id?: string | null;
+  shard_index?: number | null;
+  shard_count?: number | null;
+  shards?: DigitadorJob[];
+}
+
+export interface DigitadorJobFile {
+  key: string;
+  name?: string | null;
+}
+
+export interface DigitadorJobFilesResponse {
+  job_id: string;
+  prefix: string;
+  files: DigitadorJobFile[];
+  count: number;
+}
+
+export interface DigitadorJobEvent {
+  id: number;
+  job_id: string;
+  event_type: string;
+  message: string;
+  timestamp: string;
+  radicado?: string | null;
+  nit?: string | null;
+  index?: number | null;
+  total?: number | null;
+  metadata?: Record<string, unknown>;
+}
+
 export interface DesktopBridge {
   getRuntimeConfig(): Promise<RuntimeConfig>;
   selectDirectory(): Promise<string | null>;

services/poblacion_apta/api/main.py

@@ -15,6 +15,7 @@ from domain.verificador_st import (
   consolidado_to_records,
   export_no_aptos_csv,
   export_paso_bot_csv,
+  export_todos_csv,
   process_cleaning_sources,
 )
 from domain.verificador_st.core import cargar_origen_servicio_desde_bytes
@@ -483,11 +484,13 @@ def get_rows(
 @app.get("/jobs/{job_id}/export.csv")
 def export_csv(
   job_id: str,
-  subset: str = Query("aptos", description="aptos | no_aptos"),
+  subset: str = Query("aptos", description="aptos | no_aptos | todos"),
 ) -> Response:
   subset_norm = (subset or "aptos").strip().lower()
-  if subset_norm not in {"aptos", "no_aptos", "errores"}:
-    raise HTTPException(status_code=400, detail="subset debe ser aptos o no_aptos")
+  if subset_norm in {"all", "completo", "consolidado"}:
+    subset_norm = "todos"
+  if subset_norm not in {"aptos", "no_aptos", "errores", "todos"}:
+    raise HTTPException(status_code=400, detail="subset debe ser aptos, no_aptos o todos")
   if subset_norm == "errores":
     subset_norm = "no_aptos"
 
@@ -504,6 +507,9 @@ def export_csv(
     if subset_norm == "aptos":
       csv_text = export_paso_bot_csv(payload)
       filename = f"poblacion_apta_aptos_{job_id[:8]}.csv"
+    elif subset_norm == "todos":
+      csv_text = export_todos_csv(payload)
+      filename = f"poblacion_apta_todos_{job_id[:8]}.csv"
     else:
       csv_text = export_no_aptos_csv(payload)
       filename = f"poblacion_apta_no_aptos_{job_id[:8]}.csv"

services/poblacion_apta/domain/verificador_st/__init__.py

@@ -3,6 +3,7 @@ from domain.verificador_st.pipeline import (
   consolidado_to_records,
   export_no_aptos_csv,
   export_paso_bot_csv,
+  export_todos_csv,
   process_cleaning_sources,
 )
 
@@ -10,5 +11,6 @@ __all__ = [
   "consolidado_to_records",
   "export_no_aptos_csv",
   "export_paso_bot_csv",
+  "export_todos_csv",
   "process_cleaning_sources",
 ]

services/poblacion_apta/domain/verificador_st/core.py

@@ -51,6 +51,8 @@ TABLA_MODERADORA = {
     2025: {4700, 19200, 50300},
     2026: {5000, 20100, 52800},
 }
+# Todos los montos de la tabla; la clasificación no filtra por año.
+VALORES_MODERADORA = {v for vals in TABLA_MODERADORA.values() for v in vals}
 
 # Longitudes según tabla entregada por el usuario.
 # Los tipos marcados como alfanuméricos permiten letras y números.
@@ -554,14 +556,15 @@ def normalizar_autorizacion(valor):
 
 def separar_copago_moderadora(valor, anio):
     """
-    Si el valor coincide con tabla moderadora del año => cuota_moderadora.
-    Si no coincide => copago.
+    Si el valor coincide con cualquier cuota de la tabla (sin importar el año)
+    => cuota_moderadora. Si no coincide => copago.
+    El parámetro anio se conserva por compatibilidad de firma.
     """
     monto = to_int(valor, default=0)
     if monto == 0:
         return 0, 0
 
-    if anio in TABLA_MODERADORA and monto in TABLA_MODERADORA[anio]:
+    if monto in VALORES_MODERADORA:
         return 0, monto
 
     return monto, 0

services/poblacion_apta/domain/verificador_st/pipeline.py

@@ -223,3 +223,21 @@ def export_no_aptos_csv(rows: list[dict[str, Any]]) -> str:
     df = df.rename(columns=CSV_FINAL_IT_RENAME)
     df = df.reindex(columns=headers)
   return df.to_csv(sep="|", index=False, encoding="utf-8-sig")
+
+
+def export_todos_csv(rows: list[dict[str, Any]]) -> str:
+  """CSV pipe-separated con todas las filas (aptos y no aptos) e OBSERVACION."""
+  headers = [*CSV_FINAL_IT_HEADERS, "OBSERVACION"]
+  if not rows:
+    df = pd.DataFrame(columns=headers)
+  else:
+    df = pd.DataFrame(rows)
+    for c in COLUMNAS_FINAL_IT:
+      if c not in df.columns:
+        df[c] = ""
+    if "OBSERVACION" not in df.columns:
+      df["OBSERVACION"] = ""
+    df = df.reindex(columns=[*COLUMNAS_FINAL_IT, "OBSERVACION"]).fillna("")
+    df = df.rename(columns=CSV_FINAL_IT_RENAME)
+    df = df.reindex(columns=headers)
+  return df.to_csv(sep="|", index=False, encoding="utf-8-sig")