Fix file upload UI refresh and add Metadata Filters UI

Dockerfile

@@ -0,0 +1,37 @@
+# Stage 1: Build frontend
+FROM node:20-alpine AS frontend-builder
+WORKDIR /frontend
+COPY RAG_FULL_APPLICATION_FRONTEND/package*.json ./
+RUN npm install
+COPY RAG_FULL_APPLICATION_FRONTEND/ ./
+RUN npm run build
+
+# Stage 2: Build backend & final monolithic image
+FROM python:3.12-slim
+WORKDIR /app
+
+# System dependencies for python-docx, tiktoken, Tesseract OCR, etc.
+RUN apt-get update && apt-get install -y \
+    build-essential libpq-dev tesseract-ocr libmagic1 libgl1 && \
+    rm -rf /var/lib/apt/lists/*
+
+COPY RAG_FULL_APPLICATION_BACKEND/requirements.txt ./
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Download cross-encoder model at build time to avoid slow cold start
+RUN python -c "from sentence_transformers import CrossEncoder; \
+    CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2')"
+
+# Copy backend files
+COPY RAG_FULL_APPLICATION_BACKEND/ ./
+
+# Copy built frontend assets to the static directory
+COPY --from=frontend-builder /frontend/dist ./static
+
+# Create data directories and set permissions for Hugging Face non-root user (UID 1000)
+RUN mkdir -p data/uploads data/bm25_indexes data/cache && \
+    chown -R 1000:1000 /app
+
+EXPOSE 7860
+
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "2"]

RAG_FULL_APPLICATION_BACKEND/Dockerfile

@@ -19,6 +19,6 @@ COPY . .
 # Create data dirs
 RUN mkdir -p data/uploads data/bm25_indexes data/cache
 
-EXPOSE 8000
+EXPOSE 7860
 
-CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000", "--workers", "2"]
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "2"]

RAG_FULL_APPLICATION_BACKEND/README.md

@@ -0,0 +1,23 @@
+---
+title: RAG Backend
+emoji: 🚀
+colorFrom: blue
+colorTo: indigo
+sdk: docker
+app_port: 7860
+pinned: false
+---
+
+# RAG Multimodal Pipeline Backend
+This is the backend API for the JEE/NEET RAG pipeline, hosted on Hugging Face Spaces.
+
+## Configuration
+All environment variables are managed via Hugging Face Secrets.
+
+## Tech Stack
+- FastAPI
+- Uvicorn
+- Docker
+- Sentence Transformers
+- Redis (External via Upstash)
+- Supabase (External)

RAG_FULL_APPLICATION_BACKEND/app/config.py

@@ -31,7 +31,7 @@ class Settings(BaseSettings):
     VISION_SPACE_URL: str = "Qwen/Qwen3-VL-30B-A3B-Demo"
 
     # Redis
-    REDIS_URL: str
+    REDIS_URL: str = ""
     CACHE_TTL_SECONDS: int = 3600
 
     # Auth
@@ -53,7 +53,7 @@ class Settings(BaseSettings):
     MAX_FILE_SIZE_MB: int = 50
 
     # CORS
-    CORS_ORIGINS: str
+    CORS_ORIGINS: str = ""
 
     class Config:
         env_file = ".env"

RAG_FULL_APPLICATION_BACKEND/app/main.py

@@ -1,8 +1,11 @@
 from fastapi import FastAPI, WebSocket, Depends
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.staticfiles import StaticFiles
+from fastapi.responses import FileResponse
 from .config import settings
 from .utils.ws_manager import ws_manager
 import logging
+import os
 
 # Setup Logger
 logging.basicConfig(level=logging.INFO)
@@ -11,7 +14,8 @@ logger = logging.getLogger(__name__)
 app = FastAPI(title="RAG Pipeline API", version="3.0.0")
 
 # CORS
-origins = settings.CORS_ORIGINS.split(",")
+# origins = settings.CORS_ORIGINS.split(",")
+origins = settings.CORS_ORIGINS.split(",") if settings.CORS_ORIGINS else ["*"]
 app.add_middleware(
     CORSMiddleware,
     allow_origins=origins,
@@ -44,3 +48,17 @@ async def pipeline_ws(websocket: WebSocket, job_id: str, token: str):
         logger.error(f"WebSocket error for job {job_id}: {e}")
     finally:
         await ws_manager.disconnect(job_id, "anonymous")
+
+# Serve frontend static files in production monolith
+static_dir = os.path.join(os.path.dirname(os.path.abspath(__file__)), "..", "static")
+if os.path.exists(static_dir):
+    app.mount("/assets", StaticFiles(directory=os.path.join(static_dir, "assets")), name="assets")
+
+    @app.get("/{catchall:path}")
+    async def serve_frontend(catchall: str):
+        # Prevent catching API calls
+        if catchall.startswith(("auth", "ingest", "query", "health", "ws")):
+            return None
+        index_file = os.path.join(static_dir, "index.html")
+        if os.path.exists(index_file):
+            return FileResponse(index_file)

RAG_FULL_APPLICATION_BACKEND/app/routers/auth.py

@@ -46,12 +46,10 @@ async def login(form_data: OAuth2PasswordRequestForm = Depends()):
 
 @router.post("/seed_admin")
 async def seed_admin():
-    """Utility to pre-create admin user for local testing. Forced clean sync."""
-    hashed = get_password_hash("admin123")
+    import traceback
     try:
-        # Delete existing to ensure fresh hash if environment changed
+        hashed = get_password_hash("admin123")
         supabase_service.client.table("users").delete().eq("username", "admin").execute()
-        
         supabase_service.client.table("users").insert({
             "username": "admin",
             "password_hash": hashed
@@ -60,4 +58,4 @@ async def seed_admin():
         return {"msg": "Admin user created/reset (admin / admin123)"}
     except Exception as e:
         logger.error(f"Seeding failed: {e}")
-        return {"msg": f"Seeding failed: {str(e)}"}
+        return {"error": str(e), "traceback": traceback.format_exc()}

RAG_FULL_APPLICATION_BACKEND/app/techniques/metadata_filter.py

@@ -6,6 +6,10 @@ class MetadataFilter(BaseRAGTechnique):
     async def retrieve(self, query: str, document_id: str, top_k: int, **kwargs) -> List[Dict[str, Any]]:
         filters = kwargs.get("filters", {})
         
+        if not filters:
+            await self.emit("DONE", "#EF4444", "Metadata Filter selected but no filters provided.")
+            return []
+            
         # 1. SQL Pre-filtering
         await self.emit("FILTER", "#D97706", f"SQL filter: {filters}...")
         matching_ids = await self.supabase.filter_chunk_ids(document_id, self.user_id, filters)

RAG_FULL_APPLICATION_BACKEND/requirements.txt

@@ -7,6 +7,7 @@ asyncpg==0.29.0
 redis==5.0.6
 python-jose[cryptography]==3.3.0
 passlib[bcrypt]==1.7.4
+bcrypt==4.0.1
 python-multipart==0.0.9
 httpx==0.27.0
 gradio_client>=1.0.0

RAG_FULL_APPLICATION_FRONTEND/src/App.jsx

@@ -15,7 +15,7 @@ function App() {
   const [password, setPassword] = useState('admin123');
   const [query, setQuery] = useState('');
   const [technique, setTechnique] = useState('hybrid');
-  const [activeJob, setActiveJob] = useState(null);
+  const [metadataFilters, setMetadataFilters] = useState('{}');
   
   const { 
     documents, setDocuments, 
@@ -23,7 +23,8 @@ function App() {
     currentAnswer, setAnswer,
     sources, setSources,
     steps, addStep, clearSteps,
-    setQuerying, isQuerying
+    setQuerying, isQuerying,
+    activeJob, setActiveJob
   } = usePipelineStore();
 
   const ws = useRef(null);
@@ -39,7 +40,7 @@ function App() {
   useEffect(() => {
     if (activeJob && isAuthenticated) {
       const token = sessionStorage.getItem('token');
-      const apiBase = import.meta.env.VITE_API_BASE_URL || 'http://localhost:8001';
+      const apiBase = import.meta.env.VITE_API_BASE_URL || window.location.origin;
       const wsProtocol = apiBase.startsWith('https') ? 'wss' : 'ws';
       const wsHost = apiBase.replace(/^https?:\/\//, '');
       const wsUrl = `${wsProtocol}://${wsHost}/ws/pipeline/${activeJob}?token=${token}`;
@@ -82,17 +83,29 @@ function App() {
     setShowSettings(false);
     
     try {
-      const { data } = await api.post('/query/search', {
+      const payload = {
         query,
         document_id: selectedDoc.id,
         technique
-      });
+      };
+      
+      if (technique === 'meta') {
+        try {
+          payload.filters = JSON.parse(metadataFilters);
+        } catch (e) {
+          alert("Invalid JSON format for Metadata Filters. Please check your input.");
+          setQuerying(false);
+          return;
+        }
+      }
+
+      const { data } = await api.post('/query/search', payload);
       setAnswer(data.answer);
       setSources(data.sources);
       setActiveJob(data.job_id);
     } catch (error) {
       console.error('Search failed', error);
-      addStep({ step: 'ERROR', detail: 'Search failed. Please try again.', color: '#EF4444' });
+      addStep({ step: 'ERROR', detail: error?.response?.data?.detail || 'Search failed. Please try again.', color: '#EF4444' });
     } finally {
       setQuerying(false);
     }
@@ -271,6 +284,19 @@ function App() {
                     <label className="text-[10px] text-gray-600 font-bold uppercase">LLM Temp</label>
                     <input type="number" defaultValue={0.1} step={0.1} className="w-full bg-surface-800 border border-surface-700 rounded px-2 py-1 text-xs outline-none focus:border-accent-500" />
                   </div>
+                  <div className="col-span-2 space-y-1 pt-2 border-t border-surface-800">
+                    <label className="text-[10px] text-gray-600 font-bold uppercase flex justify-between">
+                      <span>Metadata Filters (JSON)</span>
+                      {technique === 'meta' && <span className="text-orange-500">Required for Meta search</span>}
+                    </label>
+                    <input 
+                      type="text" 
+                      value={metadataFilters}
+                      onChange={e => setMetadataFilters(e.target.value)}
+                      placeholder='e.g., {"page": 1, "section": "Introduction"}' 
+                      className={`w-full bg-surface-800 border ${technique === 'meta' ? 'border-orange-500/50 focus:border-orange-500' : 'border-surface-700 focus:border-accent-500'} rounded px-2 py-1 text-xs outline-none font-mono text-gray-400`} 
+                    />
+                  </div>
                 </div>
               </motion.div>
             )}

RAG_FULL_APPLICATION_FRONTEND/src/api/client.js

@@ -1,7 +1,7 @@
 import axios from 'axios';
 
 const api = axios.create({
-  baseURL: import.meta.env.VITE_API_BASE_URL || 'http://localhost:8001',
+  baseURL: import.meta.env.VITE_API_BASE_URL || window.location.origin,
 });
 
 // Interceptor for JWT

RAG_FULL_APPLICATION_FRONTEND/src/components/FileUpload.jsx

@@ -14,7 +14,7 @@ const STRATEGIES = [
 
 export default function FileUpload() {
   const [file, setFile] = useState(null);
-  const { setIngesting, isIngesting } = usePipelineStore();
+  const { setIngesting, isIngesting, setActiveJob, clearSteps } = usePipelineStore();
   const [status, setStatus] = useState('idle'); // idle | loading | success | error
   const [errorMsg, setErrorMsg] = useState('');
 
@@ -36,7 +36,9 @@ export default function FileUpload() {
     formData.append('strategy', strategy);
     
     try {
+      clearSteps();
       const { data } = await api.post('/ingest/upload', formData);
+      setActiveJob(data.job_id);
       setStatus('success');
       setTimeout(() => {
         setStatus('idle');

RAG_FULL_APPLICATION_FRONTEND/src/store/pipelineStore.js

@@ -8,6 +8,7 @@ export const usePipelineStore = create((set) => ({
   steps: [],
   isQuerying: false,
   isIngesting: false,
+  activeJob: null,
   
   setDocuments: (documents) => set({ documents }),
   setSelectedDoc: (selectedDoc) => set({ selectedDoc }),
@@ -17,4 +18,5 @@ export const usePipelineStore = create((set) => ({
   clearSteps: () => set({ steps: [] }),
   setQuerying: (isQuerying) => set({ isQuerying }),
   setIngesting: (isIngesting) => set({ isIngesting }),
+  setActiveJob: (activeJob) => set({ activeJob }),
 }));

README.md

@@ -0,0 +1,23 @@
+---
+title: RAG Monolith
+emoji: 🚀
+colorFrom: blue
+colorTo: indigo
+sdk: docker
+app_port: 7860
+pinned: false
+---
+
+# RAG Multimodal Pipeline
+
+This repository hosts a monolithic full-stack deployment of the RAG Multimodal Pipeline on Hugging Face Spaces.
+
+## Setup & Secrets
+
+Make sure to configure the following Secrets in your Space settings:
+1. `SUPABASE_URL`
+2. `SUPABASE_KEY`
+3. `SUPABASE_DB_URL`
+4. `REDIS_URL`
+5. `JWT_SECRET_KEY`
+6. `CORS_ORIGINS` (value can be `*` or the final Space URL)

patch_auth.py

@@ -0,0 +1,26 @@
+import re
+
+with open("/Content/AI-PROJECTS/RAG_FULL_APPLICATION/RAG_FULL_APPLICATION_BACKEND/app/routers/auth.py", "r") as f:
+    content = f.read()
+
+new_seed_admin = """@router.post("/seed_admin")
+async def seed_admin():
+    import traceback
+    try:
+        hashed = get_password_hash("admin123")
+        supabase_service.client.table("users").delete().eq("username", "admin").execute()
+        supabase_service.client.table("users").insert({
+            "username": "admin",
+            "password_hash": hashed
+        }).execute()
+        logger.info("Admin user seeded successfully.")
+        return {"msg": "Admin user created/reset (admin / admin123)"}
+    except Exception as e:
+        logger.error(f"Seeding failed: {e}")
+        return {"error": str(e), "traceback": traceback.format_exc()}
+"""
+
+content = re.sub(r'@router\.post\("/seed_admin"\).*?(?=\n@|\Z)', new_seed_admin, content, flags=re.DOTALL)
+
+with open("/Content/AI-PROJECTS/RAG_FULL_APPLICATION/RAG_FULL_APPLICATION_BACKEND/app/routers/auth.py", "w") as f:
+    f.write(content)