.

2025-11-02 21:45:44 +00:00
parent 3ea85d360c
commit 97898e14d2
413 changed files with 12229 additions and 465 deletions
@@ -529,7 +529,11 @@ def process_file(path: str, out_dir: str, overwrite: bool = False, verbose: bool
    # deterministic doc id (if present in filename)
    docid = None

-    docid = base.split("_document_content_")[1].split("_")[0]
+    try:
+        docid = base.split("_document_content_")[1].split("_")[0]
+    except Exception as e:
+        logger.error(f"Failed to extract docid from filename {base}: {e}")
+        return False, "invalid-filename"

    logger.debug(f"Extracted docid: {docid}")
    # logger.debug(f"DOCUMENT_SUMMARYS keys: {DOCUMENT_SUMMARYS.get(docid)}")