updated master parquet script.

tools/build_master_parquet.py

@@ -36,30 +36,50 @@ def main():
     ap.add_argument("--src_root", required=True, help="dir with metadata.csv world_annotations.json world_images/")
     ap.add_argument("--out_dir", required=True, help="hf_repo/data/master")
     ap.add_argument("--rows_per_shard", type=int, default=256)
+    ap.add_argument(
+        "--strict_filenames",
+        action="store_true",
+        help="if set, fail when metadata filenames are missing from COCO/images; default is to skip",
+    )
     args = ap.parse_args()
 
     src = Path(args.src_root)
     out = Path(args.out_dir)
     out.mkdir(parents=True, exist_ok=True)
 
-    meta = pd.read_csv(src / "metadata.csv")
+    meta = pd.read_csv(src / "metadata.csv", low_memory=False)
     coco = read_json(src / "world_annotations.json")
     images_by_filename, annos_by_image_id, categories = build_indices(coco)
 
     missing = set(meta["filename"]) - set(images_by_filename.keys())
     if missing:
         ex = next(iter(missing))
-        raise RuntimeError(f"{len(missing)} filenames in metadata missing from COCO images[]. Example: {ex}")
+        if args.strict_filenames:
+            raise RuntimeError(f"{len(missing)} filenames in metadata missing from COCO images[]. Example: {ex}")
+        print(
+            f"Warning: {len(missing)} filenames in metadata missing from COCO images[]; "
+            f"they will be skipped. Example: {ex}"
+        )
 
     rows = []
     shard_idx = 0
 
+    skipped_not_in_coco = 0
+    skipped_missing_image_file = 0
+    written_rows = 0
+
     for r in tqdm(meta.itertuples(index=False), total=len(meta), desc="Building master parquet"):
         filename = r.filename
+        if filename in missing:
+            skipped_not_in_coco += 1
+            continue
         img_info = images_by_filename[filename]
         image_id = int(img_info["id"])
         img_path = src / "world_images" / filename
 
+        if not img_path.exists():
+            skipped_missing_image_file += 1
+            continue
         b = img_path.read_bytes()
         annos = annos_by_image_id.get(image_id, [])
 
@@ -76,6 +96,7 @@ def main():
             "coco_annotations": orjson.dumps(annos).decode("utf-8"),
             "coco_categories": orjson.dumps(categories).decode("utf-8"),
         })
+        written_rows += 1
 
         if len(rows) >= args.rows_per_shard:
             shard_path = out / f"all-{shard_idx:05d}.parquet"
@@ -87,7 +108,10 @@ def main():
         shard_path = out / f"all-{shard_idx:05d}.parquet"
         write_shard(rows, shard_path)
 
-    print(f"Done. Wrote master shards to {out}")
+    print(
+        f"Done. Wrote master shards to {out} | written={written_rows} "
+        f"skipped_not_in_coco={skipped_not_in_coco} skipped_missing_image_file={skipped_missing_image_file}"
+    )
 
 
 if __name__ == "__main__":