fix: unconditional SPA catch-all route + diagnostic fallback for missing frontend build

backend/main.py

@@ -5,7 +5,7 @@ from pathlib import Path
 import pandas as pd
 from fastapi import FastAPI, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
-from fastapi.responses import FileResponse
+from fastapi.responses import FileResponse, JSONResponse
 from fastapi.staticfiles import StaticFiles
 
 from src.app.schemas import (
@@ -29,10 +29,11 @@ def _safe_int(val) -> int | None:
     except Exception:
         return None
 
+
 app = FastAPI(title="Traffic Incident Summarization API", version="0.4.0")
 app.add_middleware(
     CORSMiddleware,
-        allow_origins=["*"],
+    allow_origins=["*"],
     allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
@@ -67,11 +68,9 @@ def get_samples(track: str = "gcc"):
         df = df[df["dataset_track"].fillna("us") == track]
     if df.empty:
         raise HTTPException(status_code=404, detail=f"No samples found for dataset track: {track}")
-    
     if "Start_Time" in df.columns:
         df["Start_Time"] = pd.to_datetime(df["Start_Time"], errors="coerce")
         df = df.sort_values(by="Start_Time", ascending=False)
-        
     # Pick one representative sample per severity level for a compact, diverse preview
     sev_map_int = {1: "Low", 2: "Medium", 3: "High", 4: "Critical"}
     severity_order = [3, 2, 4, 1]  # High, Medium, Critical, Low — most interesting first
@@ -94,19 +93,15 @@ def get_samples(track: str = "gcc"):
             selected_rows.append(row)
             used_indices.add(row.name)
     sample_df = pd.DataFrame(selected_rows)
-
     items = []
     for idx, row in sample_df.iterrows():
         def clean(val):
             s = str(val).strip() if pd.notna(val) else ""
             return "" if s.lower() in ("nan", "none", "") else s
-
         loc_cols = ["road_name", "Street", "district", "City", "State", "emirate"]
         location_parts = [clean(row.get(col)) for col in loc_cols]
         title = " · ".join([p for p in location_parts if p][:3]) or f"Sample incident {idx + 1}"
-
         desc = clean(row.get("Description", "")) or "No description available."
-
         sev = row.get("Severity", "")
         if clean(str(sev)):
             if "severity" not in desc.lower():
@@ -116,11 +111,9 @@ def get_samples(track: str = "gcc"):
                     desc = f"{desc} Classified as {sev_str} severity."
                 except Exception:
                     pass
-
         src_lbl = clean(row.get("source_label", ""))
         if not src_lbl:
             src_lbl = "US Accidents" if track == "us" else "GCC sample"
-
         items.append(
             SampleItem(
                 id=str(idx + 1),
@@ -133,7 +126,6 @@ def get_samples(track: str = "gcc"):
     return SamplesResponse(items=items)
 
 
-
 @app.post("/summarize", response_model=SummarizeResponse)
 def summarize(request: SummarizeRequest):
     try:
@@ -163,20 +155,34 @@ def compare(request: CompareRequest):
     except Exception as exc:
         raise HTTPException(status_code=500, detail=str(exc)) from exc
 
+
 # ── Serve React Frontend (Single-Container Deployment e.g., Hugging Face) ──
 _DIST_PATH = Path(__file__).parent.parent / "frontend" / "dist"
-
-if _DIST_PATH.exists() and _DIST_PATH.is_dir():
-    # Mount static assets first (CSS/JS files inside /assets)
-    app.mount("/assets", StaticFiles(directory=str(_DIST_PATH / "assets")), name="assets")
-
-    # Catch-all route for Single Page Application (React Router)
-    @app.get("/{full_path:path}", response_class=FileResponse)
-    async def serve_frontend(full_path: str):
-        # Prevent accessing files outside dist/
-        req_path = _DIST_PATH / full_path
-        if req_path.exists() and req_path.is_file():
-            return FileResponse(req_path)
-        # Otherwise, fall back to React's index.html
-        return FileResponse(_DIST_PATH / "index.html")
-
+_INDEX_HTML = _DIST_PATH / "index.html"
+_ASSETS_PATH = _DIST_PATH / "assets"
+
+# Mount static assets if the build exists
+if _ASSETS_PATH.exists() and _ASSETS_PATH.is_dir():
+    app.mount("/assets", StaticFiles(directory=str(_ASSETS_PATH)), name="assets")
+
+
+# Catch-all SPA route — always registered so HF Spaces never gets a 404 at root
+@app.get("/{full_path:path}")
+async def serve_frontend(full_path: str):
+    # Try to serve an exact file from dist (e.g. favicon.ico, robots.txt)
+    req_path = _DIST_PATH / full_path
+    if req_path.exists() and req_path.is_file():
+        return FileResponse(req_path)
+    # SPA fallback: always return index.html so React Router handles the path
+    if _INDEX_HTML.exists():
+        return FileResponse(_INDEX_HTML)
+    # Frontend build not found — return diagnostic JSON instead of 404
+    return JSONResponse(
+        status_code=200,
+        content={
+            "status": "api_only",
+            "message": "Frontend build not found. API is running.",
+            "api_docs": "/docs",
+            "health": "/health",
+        },
+    )