Update app.py

Key fix: Gradio File type must be filepath, not file.
Also: support ZIP upload OR multiple PDFs.

app.py

@@ -1,8 +1,9 @@
-import os
 import json
+import os
 import shutil
 import zipfile
 from pathlib import Path
+from typing import List, Optional, Dict, Any
 
 import gradio as gr
 
@@ -12,71 +13,88 @@ from pipeline import run_pipeline
 DEFAULT_CONFIG = {
     "model": os.getenv("OPENAI_MODEL", "gpt-4o-mini"),
     "rewrite": False,
-    # later: projects, thresholds, etc.
+    # Later: WordPress will send projects in here
+    "projects": [
+        {"name": "STANDARD", "description": "Generic scoring"}
+    ],
 }
 
 
-def _ensure_clean_dir(p: Path):
-    if p.exists():
-        shutil.rmtree(p)
-    p.mkdir(parents=True, exist_ok=True)
+def _stage_inputs(uploaded_files: List[str]) -> List[str]:
+    """
+    Gradio gives us file paths (type='filepath').
+    We copy them into /tmp/input_uploads so pipeline has stable paths.
+    If ZIP is uploaded, extract PDFs.
+    """
+    input_dir = Path("/tmp/input_uploads").resolve()
+    if input_dir.exists():
+        shutil.rmtree(input_dir)
+    input_dir.mkdir(parents=True, exist_ok=True)
 
+    staged: List[str] = []
 
-def process(files, config_json):
-    # 1) parse config safely
-    cfg = DEFAULT_CONFIG.copy()
+    for f in uploaded_files or []:
+        src = Path(f)
+        if not src.exists():
+            continue
+
+        # ZIP case
+        if src.suffix.lower() == ".zip":
+            with zipfile.ZipFile(src, "r") as z:
+                z.extractall(input_dir)
+            # collect PDFs
+            for p in sorted(input_dir.rglob("*.pdf")):
+                staged.append(str(p.resolve()))
+            continue
+
+        # PDF case
+        if src.suffix.lower() == ".pdf":
+            dst = input_dir / src.name
+            shutil.copy2(src, dst)
+            staged.append(str(dst.resolve()))
+
+    return staged
+
+
+def process(files, config_json: str):
+    # files can be None / single / list depending on gradio version; normalize.
+    if files is None:
+        return None
+
+    if isinstance(files, str):
+        uploaded_files = [files]
+    else:
+        uploaded_files = [f for f in files]
+
+    staged_pdfs = _stage_inputs(uploaded_files)
+    if not staged_pdfs:
+        raise gr.Error("No PDFs found. Upload PDFs or a ZIP containing PDFs.")
+
+    # config
+    cfg = DEFAULT_CONFIG
     if config_json and config_json.strip():
         try:
-            user_cfg = json.loads(config_json)
-            if isinstance(user_cfg, dict):
-                cfg.update(user_cfg)
-        except Exception as e:
-            raise gr.Error(f"Config JSON is invalid: {e}")
-
-    # 2) validate upload
-    if not files or len(files) == 0:
-        raise gr.Error("Upload at least 1 PDF.")
-
-    # 3) copy uploads into /tmp (HF safe)
-    base = Path("/tmp/resume_eval")
-    input_dir = base / "input_uploads"
-    _ensure_clean_dir(input_dir)
-
-    input_files = []
-    for f in files:
-        src = Path(f)  # because we use type="filepath"
-        if not src.exists():
-            raise gr.Error(f"Uploaded file path missing: {src}")
-        dst = input_dir / src.name
-        shutil.copy2(src, dst)
-        input_files.append(str(dst))
-
-    # 4) run pipeline => must return real zip path
-    zip_path = run_pipeline(input_files=input_files, config=cfg)
-
-    # 5) hard validation: file must exist and be non-trivial
-    z = Path(zip_path)
-    if not z.exists():
-        raise gr.Error(f"Pipeline returned zip path but file does not exist: {z}")
-    if z.stat().st_size < 200:
-        raise gr.Error(f"ZIP too small ({z.stat().st_size} bytes). Pipeline likely wrote nothing.")
+            cfg = json.loads(config_json)
+        except Exception:
+            raise gr.Error("Config JSON is invalid. Fix JSON formatting.")
 
-    return str(z)
+    zip_path = run_pipeline(input_files=staged_pdfs, config=cfg)
+    return zip_path
 
 
 with gr.Blocks() as demo:
     gr.Markdown("# Resume Evaluator")
 
     files = gr.File(
-        label="Upload PDF(s)",
+        label="Upload PDF(s) or a ZIP",
         file_count="multiple",
-        type="filepath"   # IMPORTANT: must be filepath or binary
+        type="filepath",  # <-- FIX
     )
 
     config_json = gr.Textbox(
         label="Config JSON (optional)",
         value=json.dumps(DEFAULT_CONFIG, indent=2),
-        lines=10
+        lines=10,
     )
 
     btn = gr.Button("Process")