feat: Update testing documentation and introduce comprehensive report for IPSExtractor campaign. Add sections for executive summary, objectives, environment details, and methodology. Enhance README with compilation instructions and data sources for metrics analysis.

services/services-count-facturas-medicas-serving/internal/pipeline/processor.go

@@ -32,28 +32,31 @@ func NewProcessor(predictor *model.Predictor, ocrEngine *ocr.Engine, opts Option
 func (p *Processor) ProcessPDF(filename string, pdfData []byte) domain.PDFResult {
 	start := time.Now()
 	result := domain.PDFResult{Filename: filename, Status: domain.StatusOK}
+	var renderDur, ocrDur, clusterDur, inferDur time.Duration
 
 	tRender := time.Now()
 	img, pageW, pageH, err := ocr.RenderPage(pdfData, p.opts.DPI)
-	renderDur := time.Since(tRender)
+	renderDur = time.Since(tRender)
 	if err != nil {
 		result.Status = domain.StatusError
 		result.Error = err.Error()
+		p.finishTiming(&result, renderDur, 0, 0, 0, time.Since(start))
 		return result
 	}
 
 	tOCR := time.Now()
 	boxes, err := p.ocrEngine.OCRImage(img)
-	ocrDur := time.Since(tOCR)
+	ocrDur = time.Since(tOCR)
 	if err != nil {
 		result.Status = domain.StatusError
 		result.Error = err.Error()
+		p.finishTiming(&result, renderDur, ocrDur, 0, 0, time.Since(start))
 		return result
 	}
 	if len(boxes) == 0 {
 		result.ServiceCount = 0
 		result.Services = []domain.ServiceRow{}
-		p.logTiming(filename, renderDur, ocrDur, 0, 0, time.Since(start))
+		p.finishTiming(&result, renderDur, ocrDur, 0, 0, time.Since(start))
 		return result
 	}
 
@@ -65,11 +68,11 @@ func (p *Processor) ProcessPDF(filename string, pdfData []byte) domain.PDFResult
 	ClusterRows(indexed, p.opts.DBSCANEps, p.opts.DBSCANMinSamples)
 
 	rows := buildFeatureRows(indexed, pageW, pageH)
-	clusterDur := time.Since(tCluster)
+	clusterDur = time.Since(tCluster)
 	if len(rows) == 0 {
 		result.ServiceCount = 0
 		result.Services = []domain.ServiceRow{}
-		p.logTiming(filename, renderDur, ocrDur, clusterDur, 0, time.Since(start))
+		p.finishTiming(&result, renderDur, ocrDur, clusterDur, 0, time.Since(start))
 		return result
 	}
 
@@ -84,7 +87,7 @@ func (p *Processor) ProcessPDF(filename string, pdfData []byte) domain.PDFResult
 		}
 		result.Services = services
 		result.ServiceCount = len(rows)
-		p.logTiming(filename, renderDur, ocrDur, clusterDur, 0, time.Since(start))
+		p.finishTiming(&result, renderDur, ocrDur, clusterDur, 0, time.Since(start))
 		return result
 	}
 
@@ -95,10 +98,11 @@ func (p *Processor) ProcessPDF(filename string, pdfData []byte) domain.PDFResult
 
 	tInfer := time.Now()
 	probs, err := p.predictor.PredictProba(features)
-	inferDur := time.Since(tInfer)
+	inferDur = time.Since(tInfer)
 	if err != nil {
 		result.Status = domain.StatusError
 		result.Error = fmt.Sprintf("inference: %v", err)
+		p.finishTiming(&result, renderDur, ocrDur, clusterDur, inferDur, time.Since(start))
 		return result
 	}
 
@@ -128,16 +132,21 @@ func (p *Processor) ProcessPDF(filename string, pdfData []byte) domain.PDFResult
 
 	result.Services = services
 	result.ServiceCount = serviceCount
-	p.logTiming(filename, renderDur, ocrDur, clusterDur, inferDur, time.Since(start))
+	p.finishTiming(&result, renderDur, ocrDur, clusterDur, inferDur, time.Since(start))
 	return result
 }
 
-func (p *Processor) logTiming(filename string, render, ocr, cluster, infer, total time.Duration) {
+func (p *Processor) finishTiming(result *domain.PDFResult, render, ocrD, cluster, infer, total time.Duration) {
+	result.RenderMs = render.Milliseconds()
+	result.OCRMs = ocrD.Milliseconds()
+	result.ClusterMs = cluster.Milliseconds()
+	result.InferMs = infer.Milliseconds()
+	result.TotalMs = total.Milliseconds()
 	if !p.opts.LogPipelineTiming {
 		return
 	}
 	log.Printf(
 		"pipeline filename=%s ocr_render=%s ocr_tesseract=%s cluster=%s infer=%s total=%s",
-		filename, render, ocr, cluster, infer, total,
+		result.Filename, render, ocrD, cluster, infer, total,
 	)
 }