Update app.py

app.py

@@ -1,6 +1,7 @@
 import json
 import os
 import shutil
+import uuid
 import zipfile
 from pathlib import Path
 from typing import Any, Dict, List, Optional, Tuple
@@ -20,19 +21,23 @@ DEFAULT_CONFIG: Dict[str, Any] = {
 
 TMP_ROOT = Path("/tmp/resume_evaluator").resolve()
 UPLOAD_DIR = TMP_ROOT / "input_uploads"
-OUTPUT_DIR = TMP_ROOT / "output"          # IMPORTANT: do not delete if you want dedupe persistence
+
+# IMPORTANT:
+# Treat OUTPUT_ROOT as a persistent root (for dedupe manifest/state).
+# Each run should write into OUTPUT_ROOT/jobs/{job_id}/...
+OUTPUT_ROOT = TMP_ROOT / "output"
 RESULTS_ZIP = TMP_ROOT / "results.zip"
 
 
 def _ensure_dirs() -> None:
     TMP_ROOT.mkdir(parents=True, exist_ok=True)
     UPLOAD_DIR.mkdir(parents=True, exist_ok=True)
-    OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
+    OUTPUT_ROOT.mkdir(parents=True, exist_ok=True)
 
 
 def _clean_upload_staging_only() -> None:
     """
-    DO NOT delete OUTPUT_DIR if you want index dedupe to persist across runs.
+    DO NOT delete OUTPUT_ROOT if you want dedupe state to persist across runs.
     Only clear uploads staging.
     """
     _ensure_dirs()
@@ -86,7 +91,7 @@ def _stage_inputs(uploaded_files: Optional[List[str]]) -> List[str]:
             shutil.copy2(src, dst)
             staged.append(str(dst.resolve()))
 
-    # dedupe while preserving order
+    # Dedupe while preserving order
     seen = set()
     out: List[str] = []
     for p in staged:
@@ -118,21 +123,38 @@ def process(uploaded_files: Optional[List[str]], config_text: str) -> Tuple[str,
     if not pdf_paths:
         return ("No PDFs found. Upload PDFs or a ZIP containing PDFs.", None)
 
+    # Job boundary for future FastAPI/DB/SFTP integration
+    job_id = uuid.uuid4().hex[:12]
+
     try:
-        run_pipeline(
+        # IMPORTANT: pipeline must create a job-scoped folder under OUTPUT_ROOT and return job_dir.
+        result = run_pipeline(
             input_files=pdf_paths,
             config=cfg,
-            output_dir=str(OUTPUT_DIR),
+            output_dir=str(OUTPUT_ROOT),
+            job_id=job_id,
         )
     except Exception as e:
         return (f"Pipeline failed: {type(e).__name__}: {e}", None)
 
+    # We only zip the job folder, NOT the entire OUTPUT_ROOT.
     try:
-        _zip_dir(OUTPUT_DIR, RESULTS_ZIP)
+        job_dir = Path(result.get("job_dir") or "").resolve()
+        if not job_dir.exists() or not job_dir.is_dir():
+            return (f"Pipeline failed: invalid job_dir returned: {result.get('job_dir')!r}", None)
+
+        _zip_dir(job_dir, RESULTS_ZIP)
     except Exception as e:
         return (f"Failed to package results: {type(e).__name__}: {e}", None)
 
-    return (f"Processed {len(pdf_paths)} file(s). Results are ready.", str(RESULTS_ZIP))
+    counts = result.get("counts") or {}
+    total = counts.get("total", len(pdf_paths))
+    success = counts.get("success", "?")
+    skipped = counts.get("skipped", "?")
+    failed = counts.get("failed", "?")
+
+    status_msg = f"Job {job_id} done. total={total} success={success} skipped={skipped} failed={failed}"
+    return (status_msg, str(RESULTS_ZIP))
 
 
 def build_ui() -> gr.Blocks: