fix(api): cleanup file garantito tramite finally block dopo upload su Qdrant o fallimento pipeline

Author: strawberry-code

CHANGELOG.md

@@ -7,6 +7,9 @@ and the project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.
 
 ## [Unreleased]
 
+### Fixed
+- File cleanup now guaranteed via finally block: uploaded files are deleted after Qdrant upload or on pipeline failure
+
 ## [1.2.0] - 2025-12-02
 
 ### Added

api/routes/upload.py

@@ -170,6 +170,20 @@ def update_progress(stage: str, progress: float):
         jobs[job_id]["message"] = error_msg
         jobs[job_id]["completed_at"] = datetime.utcnow().isoformat()
 
+    finally:
+        # Cleanup: delete uploaded files after processing (success or failure)
+        cleanup_count = 0
+        for filename in filenames:
+            file_path = collection_dir / filename
+            try:
+                if file_path.exists():
+                    file_path.unlink()
+                    cleanup_count += 1
+            except Exception as cleanup_err:
+                logger.warning(f"[{job_id}] Failed to cleanup {file_path}: {cleanup_err}")
+
+        logger.info(f"[{job_id}] Cleanup: removed {cleanup_count}/{len(filenames)} uploaded files")
+
 
 @router.post("/upload")
 async def upload_file(
@@ -326,14 +340,15 @@ def run_zip_indexing(job_id: str, zip_path: Path, collection_dir: Path, collecti
     """
     logger.info(f"[{job_id}] Starting ZIP extraction from {zip_path}")
 
+    extracted_files = []  # Track files for cleanup in finally block
+
     try:
         jobs[job_id]["status"] = "running"
         jobs[job_id]["stage"] = "extracting_zip"
         jobs[job_id]["message"] = "Extracting ZIP archive"
         jobs[job_id]["progress"] = 0.05
 
         # Extract ZIP, filtering out macOS metadata and hidden files
-        extracted_files = []
         with zipfile.ZipFile(zip_path, 'r') as zf:
             for name in zf.namelist():
                 # Skip macOS metadata, hidden files, and directories
@@ -403,18 +418,42 @@ def update_progress(stage: str, progress: float):
         logger.error(f"[{job_id}] ZIP indexing FAILED: {error_msg}")
         logger.error(f"[{job_id}] Stack trace:\n{traceback.format_exc()}")
 
-        # Cleanup ZIP if still exists
-        if zip_path.exists():
-            try:
-                zip_path.unlink()
-            except Exception:
-                pass
-
         jobs[job_id]["status"] = "failed"
         jobs[job_id]["stage"] = "failed"
         jobs[job_id]["message"] = error_msg
         jobs[job_id]["completed_at"] = datetime.utcnow().isoformat()
 
+    finally:
+        # Cleanup: delete ZIP if still exists
+        if zip_path.exists():
+            try:
+                zip_path.unlink()
+                logger.debug(f"[{job_id}] Cleanup: removed ZIP file")
+            except Exception as cleanup_err:
+                logger.warning(f"[{job_id}] Failed to cleanup ZIP {zip_path}: {cleanup_err}")
+
+        # Cleanup: delete extracted files after processing (success or failure)
+        cleanup_count = 0
+        for filename in extracted_files:
+            file_path = collection_dir / filename
+            try:
+                if file_path.exists():
+                    file_path.unlink()
+                    cleanup_count += 1
+                    # Also try to remove parent dirs if empty (for nested paths)
+                    parent = file_path.parent
+                    while parent != collection_dir:
+                        try:
+                            parent.rmdir()  # Only removes if empty
+                            parent = parent.parent
+                        except OSError:
+                            break
+            except Exception as cleanup_err:
+                logger.warning(f"[{job_id}] Failed to cleanup {file_path}: {cleanup_err}")
+
+        if extracted_files:
+            logger.info(f"[{job_id}] Cleanup: removed {cleanup_count}/{len(extracted_files)} extracted files")
+
 
 @router.post("/upload-zip")
 async def upload_zip(