fix(indexer): correct batch progress reporting

Change progress yield to return batch size instead of total processed,
providing more accurate progress tracking during batch processing.

Co-authored-by: Bob <[email protected]>

Author: ErikBjare

gptme_rag/indexing/indexer.py

@@ -188,7 +188,7 @@ def add_documents_progress(
             batch = documents[processed : processed + batch_size]
             self._add_documents(batch)
             processed += len(batch)
-            yield processed
+            yield len(batch)
 
     def _add_documents(self, documents: list[Document]) -> None:
         try: