langserve integration; chatUI adapter

Dockerfile

@@ -7,11 +7,12 @@ WORKDIR /app
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
-# copy app
+# copy app and config files
 COPY app ./app
+COPY params.cfg .
 
-# expose Gradio default port
-EXPOSE 7860
+# expose both FastAPI and Gradio ports
+EXPOSE 7860 7861
 
 # launch with unbuffered output
 CMD ["python", "-u", "app/main.py"]

app/main.py

@@ -1,14 +1,10 @@
-#!/usr/bin/env python3
-"""
-Hybrid ChatFed Orchestrator with both Gradio MCP endpoints and LangServe API.
-Provides MCP compatibility while adding enhanced observability through LangServe.
-"""
-
 import gradio as gr
 from fastapi import FastAPI
 from langserve import add_routes
 from langgraph.graph import StateGraph, START, END
-from typing import TypedDict, Optional, Dict, Any
+from typing import Optional, Dict, Any, List, Literal
+from typing_extensions import TypedDict
+from pydantic import BaseModel
 from gradio_client import Client
 import uvicorn
 import os
@@ -18,15 +14,23 @@ from contextlib import asynccontextmanager
 import io
 from PIL import Image
 import threading
+from langchain_core.runnables import RunnableLambda
+
+# Local imports
+from utils import getconfig
+
+config = getconfig("params.cfg")
+
+RETRIEVER = config.get("retriever", "RETRIEVER")
+GENERATOR = config.get("generator", "GENERATOR")
 
-# Configure logging for observability
 logging.basicConfig(
     level=logging.INFO,
     format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
 )
 logger = logging.getLogger(__name__)
 
-# Define the state schema
+# Define langgraph state schema
 class GraphState(TypedDict):
     query: str
     context: str
@@ -51,13 +55,13 @@ class ChatFedOutput(TypedDict):
     result: str
     metadata: Dict[str, Any]
 
-# Enhanced retriever node with logging
+# Retriever 
 def retrieve_node(state: GraphState) -> GraphState:
     start_time = datetime.now()
     logger.info(f"Starting retrieval for query: {state['query'][:100]}...")
     
     try:
-        client = Client("giz/chatfed_retriever")
+        client = Client(RETRIEVER)
         context = client.predict(
             query=state["query"],
             reports_filter=state.get("reports_filter", ""),
@@ -90,13 +94,13 @@ def retrieve_node(state: GraphState) -> GraphState:
         })
         return {"context": "", "metadata": metadata}
 
-# Enhanced generator node with logging
+# Generator
 def generate_node(state: GraphState) -> GraphState:
     start_time = datetime.now()
     logger.info(f"Starting generation for query: {state['query'][:100]}...")
     
     try:
-        client = Client("giz/chatfed_generator")
+        client = Client(GENERATOR)
         result = client.predict(
             query=state["query"],
             context=state["context"],
@@ -126,7 +130,7 @@ def generate_node(state: GraphState) -> GraphState:
         })
         return {"result": f"Error generating response: {str(e)}", "metadata": metadata}
 
-# Build the graph
+# Build graph
 workflow = StateGraph(GraphState)
 workflow.add_node("retrieve", retrieve_node)
 workflow.add_node("generate", generate_node)
@@ -309,7 +313,7 @@ def create_gradio_interface():
     return demo
 
 # =============================================================================
-# LANGSERVE API (ENHANCED OBSERVABILITY)
+# LANGSERVE API (TELEMETRY)
 # =============================================================================
 
 def process_chatfed_query_langserve(input_data: ChatFedInput) -> ChatFedOutput:
@@ -335,12 +339,14 @@ async def lifespan(app: FastAPI):
     yield
     logger.info("🛑 Orchestrator shutting down...")
 
-# Create FastAPI app
+# Create FastAPI app with docs disabled
 app = FastAPI(
     title="ChatFed Orchestrator - Enhanced API",
     version="1.0.0",
     description="Enhanced API with observability. MCP endpoints available via Gradio interface.",
-    lifespan=lifespan
+    lifespan=lifespan,
+    docs_url=None,  # Disable /docs endpoint
+    redoc_url=None  # Disable /redoc endpoint
 )
 
 # Health check
@@ -352,9 +358,28 @@ async def health_check():
         "enhanced_api": "available_via_langserve"
     }
 
-# NEW: ChatUI-compatible input schema
-from pydantic import BaseModel
-from typing import List, Literal
+# Add root endpoint
+@app.get("/")
+async def root():
+    return {
+        "message": "ChatFed Orchestrator API",
+        "version": "1.0.0",
+        "endpoints": {
+            "health": "/health",
+            "chatfed": "/chatfed",
+            "chatfed-chatui": "/chatfed-chatui",
+            "process_query": "/process_query"
+        },
+        "gradio_interface": "http://localhost:7861/",
+        "mcp_endpoints": "http://localhost:7861/gradio_api/mcp/sse",
+        "note": "LangServe telemetry enabled - use /chatfed and /chatfed-chatui for enhanced observability"
+    }
+
+# =============================================================================
+# CHATUI ADAPTER
+# =============================================================================
+
+
 
 class ChatMessage(BaseModel):
     role: Literal["system", "user", "assistant"]
@@ -372,10 +397,15 @@ def chatui_adapter(data: ChatUIInput):
     result = process_chatfed_query_core(query=last_user_msg)
     return {"result": result, "metadata": {"source": "chatfed-langserve-adapter"}}
 
-# Add LangServe routes
+# Add LangServe routes with explicit schema definitions
+# Convert functions to Runnables
+process_chatfed_query_runnable = RunnableLambda(process_chatfed_query_langserve)
+chatui_adapter_runnable = RunnableLambda(chatui_adapter)
+
+# Add routes with explicit input/output schemas
 add_routes(
     app,
-    process_chatfed_query_langserve,
+    process_chatfed_query_runnable,
     path="/chatfed",
     input_type=ChatFedInput,
     output_type=ChatFedOutput
@@ -384,7 +414,7 @@ add_routes(
 # NEW: ChatUI-compatible LangServe route
 add_routes(
     app,
-    chatui_adapter,
+    chatui_adapter_runnable,
     path="/chatfed-chatui",
     input_type=ChatUIInput
 )
@@ -422,7 +452,7 @@ def run_gradio_server():
     demo.launch(
         server_name="0.0.0.0",
         server_port=7861,  # Different port from FastAPI
-        mcp_server=True,   # Enable MCP endpoints!
+        mcp_server=True,   
         show_error=True,
         share=False,
         quiet=True
@@ -439,7 +469,7 @@ if __name__ == "__main__":
     port = int(os.getenv("PORT", "7860"))
     
     logger.info(f"🚀 Starting FastAPI server on {host}:{port}")
-    logger.info("📊 Enhanced API with observability available at /docs")
+    logger.info("📊 Enhanced API with LangServe telemetry available")
     logger.info("🔗 MCP endpoints available via Gradio on port 7861")
     
     uvicorn.run(

params.cfg

@@ -1 +1,5 @@
+[retriever]
+RETRIEVER = giz/chatfed_retriever_old
 
+[generator]
+GENERATOR = giz/chatfed_generator