Add A/B testing toggles for OCR quality steps

Each quality improvement step can now be toggled independently:
- CLAHE checkbox (Step 3: image enhancement on/off)
- MaxCols dropdown (Step 2: 0=unlimited, 2-5)
- MinConf dropdown (Step 1: auto/20/30/40/50/60)

Backend: Query params enhance, max_cols, min_conf on process-single-page.
Response includes active_steps dict showing which steps are enabled.
Frontend: Toggle controls in VocabularyTab above the table.

This allows empirical A/B testing of each step on the same scan.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

klausur-service/backend/vocab_worksheet_api.py

@@ -1290,6 +1290,9 @@ async def process_single_page(
     page_number: int,
     ipa_mode: str = Query("none", pattern="^(auto|all|de|en|none)$"),
     syllable_mode: str = Query("none", pattern="^(auto|all|de|en|none)$"),
+    enhance: bool = Query(True, description="Step 3: CLAHE + Denoise for degraded scans"),
+    max_cols: int = Query(3, description="Step 2: Max column count (0=unlimited)"),
+    min_conf: int = Query(0, description="Step 1: Min OCR confidence (0=auto from quality score)"),
 ):
     """
     Process a SINGLE page of an uploaded PDF using the Kombi OCR pipeline.
@@ -1300,6 +1303,9 @@ async def process_single_page(
     Query params:
         ipa_mode: "none" (default), "auto", "all", "en", "de"
         syllable_mode: "none" (default), "auto", "all", "en", "de"
+        enhance: true (default) — apply CLAHE/denoise for degraded scans
+        max_cols: 3 (default) — max column count (0=unlimited)
+        min_conf: 0 (default=auto) — min OCR confidence (0=from quality score)
 
     The frontend should call this sequentially for each page.
     Returns the vocabulary for just this one page.
@@ -1332,6 +1338,9 @@ async def process_single_page(
             page_vocabulary, rotation_deg, quality_report = await _run_ocr_pipeline_for_page(
                 img_bgr, page_number, session_id,
                 ipa_mode=ipa_mode, syllable_mode=syllable_mode,
+                enable_enhance=enhance,
+                max_columns=max_cols if max_cols > 0 else None,
+                override_min_conf=min_conf if min_conf > 0 else None,
             )
         except Exception as e:
             logger.error(f"OCR pipeline failed for page {page_number + 1}: {e}", exc_info=True)
@@ -1395,11 +1404,17 @@ async def process_single_page(
         "rotation": rotation_deg,
     }
 
-    # Add scan quality report if available
+    # Add scan quality report + active steps info
     if quality_report:
-        result["scan_quality"] = quality_report.to_dict()
+        sq = quality_report.to_dict()
+        sq["active_steps"] = {
+            "step1_confidence": f"min_conf={min_ocr_conf}" if not override_min_conf else f"min_conf={override_min_conf} (override)",
+            "step2_max_columns": f"max_cols={max_columns}" if max_columns else "unlimited",
+            "step3_enhance": "on" if enable_enhance and quality_report.is_degraded else "off",
+        }
+        result["scan_quality"] = sq
     else:
-        quality_report = None  # ensure variable exists for non-pipeline path
+        quality_report = None
 
     return result
 
@@ -1411,6 +1426,9 @@ async def _run_ocr_pipeline_for_page(
     *,
     ipa_mode: str = "none",
     syllable_mode: str = "none",
+    enable_enhance: bool = True,
+    max_columns: Optional[int] = 3,
+    override_min_conf: Optional[int] = None,
 ) -> tuple:
     """Run the full Kombi OCR pipeline on a single page and return vocab entries.
 
@@ -1488,11 +1506,14 @@ async def _run_ocr_pipeline_for_page(
     except Exception as e:
         logger.warning(f"  scan quality: failed ({e})")
 
-    min_ocr_conf = scan_quality_report.recommended_min_conf if scan_quality_report else 40
+    if override_min_conf:
+        min_ocr_conf = override_min_conf
+    else:
+        min_ocr_conf = scan_quality_report.recommended_min_conf if scan_quality_report else 40
 
     # 5c. Image enhancement for degraded scans
     is_degraded = scan_quality_report.is_degraded if scan_quality_report else False
-    if is_degraded:
+    if is_degraded and enable_enhance:
         try:
             from ocr_image_enhance import enhance_for_ocr
             dewarped_bgr = enhance_for_ocr(dewarped_bgr, is_degraded=True)
@@ -1547,8 +1568,8 @@ async def _run_ocr_pipeline_for_page(
     else:
         merged_words = tess_words  # fallback to Tesseract only
 
-    # Build initial grid from merged words (limit to 3 columns for vocab tables)
-    cells, columns_meta = build_grid_from_words(merged_words, img_w, img_h, max_columns=3)
+    # Build initial grid from merged words
+    cells, columns_meta = build_grid_from_words(merged_words, img_w, img_h, max_columns=max_columns)
     for cell in cells:
         cell["ocr_engine"] = "rapid_kombi"