Upload folder using huggingface_hub

run_all_samples.py

@@ -6,82 +6,86 @@ from pathlib import Path
 from fastapi.testclient import TestClient
 from app import app   # import your FastAPI app
 
+
 # -----------------------------
 # 1. Paths
 # -----------------------------
 BASE_DIR = Path(__file__).parent
-SAMPLES_DIR = BASE_DIR / "samples"      # <-- UPDATED
+SAMPLES_DIR = BASE_DIR / "samples"      # <-- Your uploaded folder
 RESULTS_DIR = BASE_DIR / "results"
 
 RESULTS_DIR.mkdir(exist_ok=True)
 
-# -----------------------------
-# 2. Init FastAPI client
-# -----------------------------
-client = TestClient(app)
 
 # -----------------------------
-# 3. File filter
+# 2. File filter
 # -----------------------------
 def is_bill_file(path: Path):
     return path.suffix.lower() in [".pdf", ".png", ".jpg", ".jpeg"]
 
-print("\n🚀 Running extractor on local sample files...\n")
-print(f"📁 Using samples from: {SAMPLES_DIR}\n")
-
-if not SAMPLES_DIR.exists():
-    raise RuntimeError("❌ ERROR: 'samples/' folder does not exist in your Space.")
-
-summary = {}
 
 # -----------------------------
-# 4. Process all files
+# 3. MAIN EXTRACTION FUNCTION
 # -----------------------------
-for fp in sorted(SAMPLES_DIR.rglob("*")):
-    if not fp.is_file() or not is_bill_file(fp):
-        continue
-
-    print(f"🔍 Processing: {fp.name}")
-
-    # Build a usable "file://" URL for extractor
-    file_url = f"file://{fp.resolve()}"
-
-    # Call extractor API
-    try:
-        res = client.post(
-            "/extract-bill-data",
-            json={"document": file_url},
-            timeout=120
-        )
-    except Exception as e:
-        print(f"❌ Error calling extractor: {e}")
-        summary[fp.name] = "EXTRACTOR_ERROR"
-        continue
-
-    if res.status_code != 200:
-        print(f"❌ API error {res.status_code}")
-        summary[fp.name] = "API_ERROR"
-        continue
-
-    data = res.json()
-
-    # Save JSON
-    out_path = RESULTS_DIR / (fp.stem + ".json")
-    with open(out_path, "w", encoding="utf-8") as f:
-        json.dump(data, f, indent=2, ensure_ascii=False)
-
-    # Count items
-    try:
-        total_items = data["data"]["total_item_count"]
-    except:
-        total_items = "N/A"
-
-    summary[fp.name] = total_items
-    print(f"   → Saved: {out_path}")
-    print(f"   → Items Extracted: {total_items}\n")
-
-print("\n🎉 DONE! All sample bills processed.\n")
-
-print("📊 Summary:")
-for fn, items in summary.items():
-    print(f"  {fn}: {items}")
+def main():
+    print("\n🚀 Running extractor on local sample files...\n")
+    print(f"📁 Using samples from: {SAMPLES_DIR}\n")
+
+    if not SAMPLES_DIR.exists():
+        raise RuntimeError("❌ ERROR: 'samples/' folder does not exist in your Space.")
+
+    client = TestClient(app)
+
+    summary = {}
+
+    # -----------------------------
+    # 4. Process all sample files
+    # -----------------------------
+    for fp in sorted(SAMPLES_DIR.rglob("*")):
+        if not fp.is_file() or not is_bill_file(fp):
+            continue
+
+        print(f"🔍 Processing: {fp.name}")
+
+        file_url = f"file://{fp.resolve()}"
+
+        try:
+            res = client.post(
+                "/extract-bill-data",
+                json={"document": file_url},
+                timeout=120
+            )
+        except Exception as e:
+            print(f"❌ Error calling extractor: {e}")
+            summary[fp.name] = "EXTRACTOR_ERROR"
+            continue
+
+        if res.status_code != 200:
+            print(f"❌ API error {res.status_code}")
+            summary[fp.name] = "API_ERROR"
+            continue
+
+        data = res.json()
+
+        out_path = RESULTS_DIR / (fp.stem + ".json")
+        with open(out_path, "w", encoding="utf-8") as f:
+            json.dump(data, f, indent=2, ensure_ascii=False)
+
+        try:
+            total_items = data["data"]["total_item_count"]
+        except:
+            total_items = "N/A"
+
+        summary[fp.name] = total_items
+        print(f"   → Saved: {out_path}")
+        print(f"   → Items Extracted: {total_items}\n")
+
+    print("\n🎉 DONE! All sample bills processed.\n")
+    print("📊 Summary:")
+    for fn, items in summary.items():
+        print(f"  {fn}: {items}")
+
+
+# Required so FastAPI background task can call main()
+if __name__ == "__main__":
+    main()