Update app.py

app.py

@@ -1,23 +1,31 @@
 from fastapi import FastAPI, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel
+from pathlib import Path
+import logging
+import traceback
 
 from agent.graph import AgentGraph
 from agent.tools import FetchTools
 from ingest.sec import fetch_recent_filings_by_cik
 from settings import INDEX_DIR
 
+log = logging.getLogger("uvicorn.error")
+
 app = FastAPI(title="DeepDive IR Agent")
 
-# CORS so the Next.js frontend can call this later
+# CORS — tighten later once your Vercel URL is final
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["*"],          # DEV: allow all
+    allow_origins=["*"],       # for debugging; later: ["https://<your-vercel>.vercel.app", "http://localhost:3000"]
     allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
 )
 
+# Ensure index path exists (and is persistent on HF: set INDEX_DIR="/data/index" in settings)
+Path(INDEX_DIR).mkdir(parents=True, exist_ok=True)
+
 graph = AgentGraph(index_dir=INDEX_DIR)
 tools = FetchTools()
 
@@ -32,25 +40,60 @@ class AskRequest(BaseModel):
 def root():
     return {"ok": True, "msg": "DeepDive IR Agent API"}
 
+@app.get("/healthz")
+def healthz():
+    return {"ok": True}
+
 @app.post("/ingest")
 async def ingest(req: IngestRequest):
-    filings = await fetch_recent_filings_by_cik(req.cik)
-    docs = []
-    for form, url, title in filings:
-        text = await tools.get_text_from_url(url)
-        docs.append({"title": title, "url": url, "text": text})
-    if req.ir_url:
-        ir_text = await tools.get_text_from_url(req.ir_url)
-        docs.append({"title": "IR site", "url": req.ir_url, "text": ir_text})
-    if not docs:
-        raise HTTPException(400, "No documents fetched.")
-    graph.build_index(docs)
-    return {"ok": True, "num_docs": len(docs)}
+    try:
+        # SEC wants a *10-digit, zero-padded* CIK string. Keep as string to preserve leading zeros.
+        cik = req.cik.strip()
+        if not cik.isdigit():
+            raise HTTPException(status_code=400, detail="CIK must be digits only.")
+        if len(cik) > 10:
+            raise HTTPException(status_code=400, detail="CIK too long; use 10 digits.")
+        cik = cik.zfill(10)
+
+        filings = await fetch_recent_filings_by_cik(cik)  # must send a real User-Agent inside!
+        docs = []
+        for form, url, title in filings:
+            text = await tools.get_text_from_url(url)     # must send a real User-Agent inside!
+            if not text:
+                log.warning(f"Empty text for {url}")
+                continue
+            docs.append({"title": title, "url": url, "text": text})
+
+        if req.ir_url:
+            ir_text = await tools.get_text_from_url(req.ir_url)
+            if ir_text:
+                docs.append({"title": "IR site", "url": req.ir_url, "text": ir_text})
+
+        if not docs:
+            raise HTTPException(status_code=400, detail="No documents fetched.")
+
+        graph.build_index(docs)
+        return {"ok": True, "num_docs": len(docs)}
+
+    except HTTPException:
+        raise
+    except Exception as e:
+        # Log full traceback for debugging in HF "Logs"
+        log.error("Ingest failed: %s\n%s", e, traceback.format_exc())
+        raise HTTPException(status_code=502, detail=f"Ingest failed: {type(e).__name__}: {e}")
 
 @app.post("/ask")
 def ask(req: AskRequest):
-    return graph.answer(req.question)
+    try:
+        return graph.answer(req.question)
+    except Exception as e:
+        log.error("Ask failed: %s\n%s", e, traceback.format_exc())
+        raise HTTPException(status_code=500, detail=str(e))
 
 @app.get("/brief")
 def brief():
-    return graph.brief()
+    try:
+        return graph.brief()
+    except Exception as e:
+        log.error("Brief failed: %s\n%s", e, traceback.format_exc())
+        raise HTTPException(status_code=500, detail=str(e))