Simplify falcon-ocr.py to plain mode only, add --progress flag

Layout mode is currently broken upstream (PPDocLayoutV3ImageProcessorFast
import fails with latest transformers). Removing it for now; can re-enable
once falcon-perception updates.

Also adds --progress flag to expose the engine's per-image tqdm bar,
and replaces script-level batch tqdm with logger.info messages.

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

falcon-ocr.py

@@ -6,8 +6,7 @@
 #     "pillow",
 #     "torch>=2.5",
 #     "torchvision",
-#     "falcon-perception[ocr]",
-#     "tqdm",
+#     "falcon-perception",
 # ]
 # ///
 
@@ -31,9 +30,6 @@ Examples:
     # Basic text OCR
     uv run falcon-ocr.py input-dataset output-dataset
 
-    # Layout-aware OCR
-    uv run falcon-ocr.py dense-docs output --task-mode layout
-
     # Test with small sample
     uv run falcon-ocr.py dataset test --max-samples 5 --shuffle
 
@@ -58,7 +54,6 @@ import torch
 from datasets import load_dataset
 from huggingface_hub import DatasetCard, login
 from PIL import Image
-from tqdm.auto import tqdm
 
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
@@ -67,7 +62,6 @@ MODEL_ID = "tiiuae/Falcon-OCR"
 
 TASK_MODES = {
     "plain": "Full-page text extraction",
-    "layout": "Layout-aware OCR (region detection + per-region extraction)",
 }
 
 
@@ -155,6 +149,7 @@ def main(
     create_pr: bool = False,
     compile: bool = True,
     cudagraph: bool = True,
+    progress: bool = False,
     verbose: bool = False,
 ):
     check_cuda_availability()
@@ -212,42 +207,30 @@ def main(
     logger.info(f"Processing {len(dataset)} images...")
     all_outputs = []
 
-    if task_mode == "layout":
-        # Process one at a time for layout (returns structured regions)
-        for i in tqdm(range(len(dataset)), desc="Falcon OCR (layout)"):
+    # Batch plain OCR for better throughput
+    batch_size = 8
+    total_batches = (len(dataset) + batch_size - 1) // batch_size
+    for batch_idx, batch_start in enumerate(range(0, len(dataset), batch_size), 1):
+        batch_end = min(batch_start + batch_size, len(dataset))
+        logger.info(f"Batch {batch_idx}/{total_batches} ({batch_start}/{len(dataset)} done)")
+        batch_images = []
+        for i in range(batch_start, batch_end):
             try:
-                pil_image = prepare_image(dataset[i][image_column])
-                results = engine.generate_with_layout(images=[pil_image], use_tqdm=False)
-                regions = results[0] if results else []
-                all_outputs.append(json.dumps(regions, ensure_ascii=False))
+                batch_images.append(prepare_image(dataset[i][image_column]))
             except Exception as e:
-                logger.error(f"Error processing image {i}: {e}")
-                all_outputs.append(f"[OCR ERROR: {str(e)[:200]}]")
-    else:
-        # Batch plain OCR for better throughput
-        batch_size = 8
-        for batch_start in tqdm(
-            range(0, len(dataset), batch_size), desc="Falcon OCR (plain)"
-        ):
-            batch_end = min(batch_start + batch_size, len(dataset))
-            batch_images = []
-            for i in range(batch_start, batch_end):
-                try:
-                    batch_images.append(prepare_image(dataset[i][image_column]))
-                except Exception as e:
-                    logger.error(f"Error preparing image {i}: {e}")
-                    batch_images.append(Image.new("RGB", (100, 100)))
+                logger.error(f"Error preparing image {i}: {e}")
+                batch_images.append(Image.new("RGB", (100, 100)))
 
-            try:
-                texts = engine.generate_plain(
-                    images=batch_images, use_tqdm=False
-                )
-                all_outputs.extend(texts)
-            except Exception as e:
-                logger.error(f"Error processing batch {batch_start}-{batch_end}: {e}")
-                all_outputs.extend(
-                    [f"[OCR ERROR: {str(e)[:200]}]"] * len(batch_images)
-                )
+        try:
+            texts = engine.generate_plain(
+                images=batch_images, use_tqdm=progress
+            )
+            all_outputs.extend(texts)
+        except Exception as e:
+            logger.error(f"Error processing batch {batch_start}-{batch_end}: {e}")
+            all_outputs.extend(
+                [f"[OCR ERROR: {str(e)[:200]}]"] * len(batch_images)
+            )
 
     # Calculate processing time
     processing_duration = datetime.now() - start_time
@@ -419,6 +402,10 @@ if __name__ == "__main__":
         "--no-cudagraph", action="store_true",
         help="Disable CUDA graph capture",
     )
+    parser.add_argument(
+        "--progress", action="store_true",
+        help="Show per-image progress bar from the inference engine",
+    )
     parser.add_argument(
         "--verbose", action="store_true", help="Log resolved package versions",
     )
@@ -441,5 +428,6 @@ if __name__ == "__main__":
         create_pr=args.create_pr,
         compile=not args.no_compile,
         cudagraph=not args.no_cudagraph,
+        progress=args.progress,
         verbose=args.verbose,
     )