Upload 4 files

Dockerfile

@@ -0,0 +1,32 @@
+# File : Dockerfile
+# Purpose : Combined FastAPI + Streamlit in one container for HF Spaces
+
+FROM python:3.11-slim
+
+WORKDIR /app
+
+# System dependencies
+RUN apt-get update && apt-get install -y \
+    build-essential \
+    libpq-dev \
+    curl \
+    && rm -rf /var/lib/apt/lists/*
+
+# Install Python dependencies
+COPY requirements.txt .
+RUN pip install --no-cache-dir --upgrade pip && \
+    pip install --no-cache-dir -r requirements.txt && \
+    pip install --no-cache-dir streamlit
+
+# Copy application code
+COPY app/ ./app/
+COPY streamlit_app.py .
+
+# Hugging Face Spaces runs as non-root user 1000
+RUN useradd -m -u 1000 appuser && chown -R appuser:appuser /app
+USER appuser
+
+# Expose port 7860 (HF Spaces requirement)
+EXPOSE 7860
+
+CMD ["/bin/bash", "-c", "uvicorn app.main:app --host 0.0.0.0 --port 8000 & sleep 5 && streamlit run streamlit_app.py --server.port 7860 --server.address 0.0.0.0"]

README.md

@@ -1,11 +1,25 @@
----
-title: Banking Rag Ui
-emoji: ⚡
-colorFrom: gray
-colorTo: indigo
-sdk: docker
-pinned: false
-license: apache-2.0
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+---
+title: Banking RAG Platform
+emoji: 🏦
+colorFrom: blue
+colorTo: green
+sdk: docker
+pinned: false
+---
+
+# Banking RAG Platform
+
+Enterprise RAG system built on LangChain + Groq for RBI and Finance documents.
+
+## Stack
+
+| Layer | Technology |
+|---|---|
+| Backend | FastAPI |
+| AI Framework | LangChain + Groq (llama-3.3-70b-versatile) |
+| Observability | LangSmith |
+| Vector DB | ChromaDB |
+| Cache | Upstash Redis |
+| Database | Neon PostgreSQL |
+| Embeddings | BAAI/bge-small-en |
+| Deployment | Hugging Face Spaces (Docker) |

requirements.txt

@@ -0,0 +1,53 @@
+# File : requirements.txt
+# Purpose : All pinned dependencies for Banking RAG Platform
+
+# --- Web Framework ---
+fastapi==0.111.0
+uvicorn[standard]==0.29.0
+python-multipart==0.0.9
+
+# --- Auth ---
+python-jose[cryptography]==3.3.0
+passlib[bcrypt]==1.7.4
+
+# --- Database ---
+sqlalchemy==2.0.30
+asyncpg==0.29.0
+psycopg2-binary==2.9.9
+alembic==1.13.1
+
+# --- Vector Store ---
+chromadb==0.5.3
+
+# --- LangChain + Groq ---
+langchain==0.2.5
+langchain-groq==0.1.6
+langchain-community==0.2.5
+langchain-huggingface==0.0.3
+langsmith==0.1.77
+
+# --- Embeddings ---
+sentence-transformers==3.0.1
+torch==2.3.0
+transformers==4.41.2
+
+# --- Document Parsing ---
+pypdf==4.2.0
+python-docx==1.1.2
+pdfplumber==0.11.0
+
+# --- Retrieval ---
+rank-bm25==0.2.2
+
+# --- Caching ---
+redis==5.0.4
+hiredis==2.3.2
+
+# --- Async + Utils ---
+httpx==0.27.0
+python-dotenv==1.0.1
+pydantic==2.7.1
+pydantic-settings==2.3.0
+aiofiles==23.2.1
+tenacity==8.3.0
+numpy==1.26.4

streamlit_app.py

@@ -0,0 +1,241 @@
+# File : streamlit_app.py
+# Purpose : Streamlit frontend for Banking RAG Platform
+
+import streamlit as st
+import requests
+from pathlib import Path
+
+# ─── Config ───────────────────────────────────────────────
+API_URL = "http://localhost:8000"
+
+st.set_page_config(
+    page_title="Banking RAG Platform",
+    page_icon="🏦",
+    layout="wide",
+    initial_sidebar_state="expanded",
+)
+
+# ─── Session State ────────────────────────────────────────
+if "token" not in st.session_state:
+    st.session_state.token = None
+if "username" not in st.session_state:
+    st.session_state.username = None
+if "chat_history" not in st.session_state:
+    st.session_state.chat_history = []
+
+# ─── Helpers ──────────────────────────────────────────────
+def auth_headers():
+    return {"Authorization": f"Bearer {st.session_state.token}"}
+
+
+def api_post(endpoint, payload=None, files=None, auth=False):
+    headers = auth_headers() if auth else {}
+    try:
+        if files:
+            r = requests.post(f"{API_URL}{endpoint}", headers=headers, files=files, timeout=120)
+        else:
+            r = requests.post(f"{API_URL}{endpoint}", headers=headers, json=payload, timeout=120)
+        return r
+    except requests.exceptions.ConnectionError:
+        st.error("Cannot connect to backend.")
+        return None
+    except requests.exceptions.Timeout:
+        st.error("Request timed out. Please try again.")
+        return None
+
+
+def api_get(endpoint, auth=False):
+    headers = auth_headers() if auth else {}
+    try:
+        r = requests.get(f"{API_URL}{endpoint}", headers=headers, timeout=30)
+        return r
+    except requests.exceptions.ConnectionError:
+        st.error("Cannot connect to backend.")
+        return None
+
+
+# ─── Sidebar ──────────────────────────────────────────────
+with st.sidebar:
+    st.title("🏦 Banking RAG")
+    st.markdown("---")
+
+    if st.session_state.token:
+        st.success(f"Logged in as **{st.session_state.username}**")
+        if st.button("Logout", use_container_width=True):
+            st.session_state.token = None
+            st.session_state.username = None
+            st.session_state.chat_history = []
+            st.rerun()
+        st.markdown("---")
+        page = st.radio(
+            "Navigate",
+            ["💬 Ask Documents", "📁 My Documents", "📤 Upload"],
+            label_visibility="collapsed",
+        )
+    else:
+        page = "🔐 Login"
+
+    st.markdown("---")
+    st.caption("Built with LangChain + Groq + ChromaDB")
+    st.caption(f"Backend: {API_URL}")
+
+
+# ─── Login / Register Page ────────────────────────────────
+if not st.session_state.token:
+    st.title("🏦 Banking RAG Platform")
+    st.markdown("Enterprise AI assistant for RBI and Finance documents.")
+    st.markdown("---")
+
+    tab1, tab2 = st.tabs(["Login", "Register"])
+
+    with tab1:
+        st.subheader("Login")
+        username = st.text_input("Username", key="login_user")
+        password = st.text_input("Password", type="password", key="login_pass")
+        if st.button("Login", use_container_width=True, type="primary"):
+            if username and password:
+                r = api_post("/auth/login", {"username": username, "password": password})
+                if r and r.status_code == 200:
+                    data = r.json()
+                    st.session_state.token = data["access_token"]
+                    st.session_state.username = username
+                    st.success("Login successful!")
+                    st.rerun()
+                elif r:
+                    st.error(r.json().get("detail", "Login failed"))
+            else:
+                st.warning("Enter username and password")
+
+    with tab2:
+        st.subheader("Register")
+        reg_email = st.text_input("Email", key="reg_email")
+        reg_username = st.text_input("Username", key="reg_user")
+        reg_password = st.text_input("Password", type="password", key="reg_pass")
+        if st.button("Register", use_container_width=True, type="primary"):
+            if reg_email and reg_username and reg_password:
+                r = api_post("/auth/register", {
+                    "email": reg_email,
+                    "username": reg_username,
+                    "password": reg_password,
+                })
+                if r and r.status_code == 201:
+                    st.success("Account created! Please login.")
+                elif r:
+                    st.error(r.json().get("detail", "Registration failed"))
+            else:
+                st.warning("Fill all fields")
+
+
+# ─── Ask Documents Page ───────────────────────────────────
+elif page == "💬 Ask Documents":
+    st.title("💬 Ask Your Documents")
+    st.markdown("Ask questions about your uploaded RBI and Finance documents.")
+    st.markdown("---")
+
+    # Chat history
+    for msg in st.session_state.chat_history:
+        with st.chat_message(msg["role"]):
+            st.markdown(msg["content"])
+            if msg["role"] == "assistant" and "citations" in msg:
+                if msg["citations"]:
+                    with st.expander("📄 Sources"):
+                        for i, cite in enumerate(msg["citations"], 1):
+                            st.markdown(f"**[{i}]** Page {cite.get('page_number', 'N/A')} — {cite.get('content_preview', '')}")
+
+    # Query input
+    query = st.chat_input("Ask something about your documents...")
+    if query:
+        st.session_state.chat_history.append({"role": "user", "content": query})
+        with st.chat_message("user"):
+            st.markdown(query)
+
+        with st.chat_message("assistant"):
+            with st.spinner("Searching documents..."):
+                r = api_post("/rag/query", {"query": query}, auth=True)
+                if r and r.status_code == 200:
+                    data = r.json()
+                    answer = data.get("answer", "No answer found.")
+                    citations = data.get("citations", [])
+                    from_cache = data.get("from_cache", False)
+
+                    st.markdown(answer)
+
+                    if from_cache:
+                        st.caption("⚡ From cache")
+
+                    if citations:
+                        with st.expander("📄 Sources"):
+                            for i, cite in enumerate(citations, 1):
+                                st.markdown(f"**[{i}]** Page {cite.get('page_number', 'N/A')} — {cite.get('content_preview', '')}")
+
+                    st.session_state.chat_history.append({
+                        "role": "assistant",
+                        "content": answer,
+                        "citations": citations,
+                    })
+                elif r:
+                    err = r.json().get("detail", "Query failed")
+                    st.error(err)
+                    st.session_state.chat_history.append({
+                        "role": "assistant",
+                        "content": f"Error: {err}",
+                        "citations": [],
+                    })
+
+    if st.session_state.chat_history:
+        if st.button("Clear Chat"):
+            st.session_state.chat_history = []
+            st.rerun()
+
+
+# ─── Upload Page ──────────────────────────────────────────
+elif page == "📤 Upload":
+    st.title("📤 Upload Documents")
+    st.markdown("Upload PDF, DOCX, or TXT files to query with AI.")
+    st.markdown("---")
+
+    uploaded_file = st.file_uploader(
+        "Choose a file",
+        type=["pdf", "docx", "txt"],
+        help="Max 50MB per file",
+    )
+
+    if uploaded_file:
+        st.info(f"Selected: **{uploaded_file.name}** ({round(uploaded_file.size / 1024, 1)} KB)")
+        if st.button("Upload & Process", type="primary", use_container_width=True):
+            with st.spinner("Uploading and processing document... this may take a minute."):
+                files = {"file": (uploaded_file.name, uploaded_file.getvalue(), uploaded_file.type)}
+                r = api_post("/documents/upload", files=files, auth=True)
+                if r and r.status_code == 201:
+                    data = r.json()
+                    st.success("Document uploaded successfully!")
+                    st.json({
+                        "filename": data["filename"],
+                        "total_chunks": data["total_chunks"],
+                        "status": data["status"],
+                    })
+                elif r:
+                    st.error(r.json().get("detail", "Upload failed"))
+
+
+# ─── My Documents Page ────────────────────────────────────
+elif page == "📁 My Documents":
+    st.title("📁 My Documents")
+    st.markdown("---")
+
+    r = api_get("/documents/list", auth=True)
+    if r and r.status_code == 200:
+        docs = r.json()
+        if not docs:
+            st.info("No documents uploaded yet. Go to Upload to add documents.")
+        else:
+            st.success(f"{len(docs)} document(s) found")
+            for doc in docs:
+                with st.expander(f"📄 {doc['filename']}"):
+                    col1, col2, col3 = st.columns(3)
+                    col1.metric("Chunks", doc["total_chunks"])
+                    col2.metric("Type", doc["file_type"].upper())
+                    col3.metric("Status", doc["status"].capitalize())
+                    st.caption(f"Uploaded: {doc['created_at']}")
+    elif r:
+        st.error("Failed to load documents")