Upload 4 files

Dockerfile

@@ -0,0 +1,29 @@
+FROM python:3.11
+
+# 1. Install Ollama
+RUN curl -fsSL https://ollama.com/install.sh | sh
+
+# 2. User Setup
+RUN useradd -m -u 1000 user
+ENV USER=user
+ENV PATH="/home/user/.local/bin:$PATH"
+ENV HOME=/home/user
+ENV OLLAMA_KEEP_ALIVE=5m
+
+# 3. Workdir
+WORKDIR $HOME/app
+
+# 4. Switch User
+USER user
+
+# 5. Install Python Libs
+RUN pip install --no-cache-dir fastapi uvicorn ollama 
+# 6. Copy Files
+COPY --chown=user . .
+
+# 7. Start Script Permission
+RUN chmod +x entrypoint.sh
+
+# 8. Ports
+EXPOSE 7860
+CMD ["./entrypoint.sh"]

entrypoint.sh

@@ -0,0 +1,16 @@
+#!/bin/bash
+pip install -r requirements.txt
+# 1. Ollama Server Start
+ollama serve &
+
+echo "Waiting for Ollama server..."
+sleep 5
+
+# 2. Qwen Model Pull
+# Note: Ye model bada hai, download hone mein 2-3 minute lag sakte hain
+echo "Pulling qwen2.5:3b..."
+ollama pull qwen2.5:3b
+
+# 3. FastAPI Start
+echo "Starting Public API..."
+uvicorn main:app --host 0.0.0.0 --port 7860

main.py

@@ -0,0 +1,194 @@
+import os
+import logging
+import asyncio
+import base64
+import io
+from typing import Annotated, List
+from contextlib import asynccontextmanager
+
+from fastapi import FastAPI
+from fastapi.responses import StreamingResponse
+from pydantic import BaseModel, Field
+
+import httpx
+from duckduckgo_search import DDGS 
+from bs4 import BeautifulSoup
+from PIL import Image # Image save karne ke liye
+
+# --- LangChain / AI Core ---
+from langchain_ollama import ChatOllama
+from langchain_core.messages import HumanMessage, SystemMessage, BaseMessage
+from langchain_core.tools import tool
+from langgraph.graph import StateGraph, END, START
+from langgraph.prebuilt import ToolNode
+from langgraph.checkpoint.memory import MemorySaver
+
+# --------------------------------------------------------------------------------------
+# 1. Configuration
+# --------------------------------------------------------------------------------------
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger("GenAI-Agent")
+
+MODEL_NAME = "qwen2.5:3b" 
+BASE_URL = "http://localhost:11434"
+
+# --- IMAGE JUGAAD CONFIG ---
+# HuggingFace ka Free API use karenge (No GPU needed locally for image gen)
+HF_API_URL = "https://api-inference.huggingface.co/models/black-forest-labs/FLUX.1-dev"
+# Agar tumhara apna Token hai toh yahan daalo, warna kai models free open access dete hain
+# Best practice: Get a free token from huggingface.co/settings/tokens
+HF_TOKEN = os.getenv("HF_TOKEN", "") 
+headers = {"Authorization": f"Bearer {HF_TOKEN}"} if HF_TOKEN else {}
+
+http_client = httpx.AsyncClient(timeout=30.0, follow_redirects=True)
+
+@asynccontextmanager
+async def lifespan(app: FastAPI):
+    # Create static folder for images
+    os.makedirs("static/images", exist_ok=True)
+    yield
+    await http_client.aclose()
+
+app = FastAPI(title="GenAI Text & Image Agent", lifespan=lifespan)
+
+# --------------------------------------------------------------------------------------
+# 2. Tools (Web + Image Jugaad)
+# --------------------------------------------------------------------------------------
+
+@tool
+async def web_search(query: str) -> str:
+    """Search the web for information."""
+    def run_sync_search(q):
+        try:
+            with DDGS() as ddgs:
+                return list(ddgs.text(q, max_results=4))
+        except Exception as e:
+            return str(e)
+
+    try:
+        results = await asyncio.to_thread(run_sync_search, query)
+        if isinstance(results, str) or not results:
+            return "No results found."
+        
+        output = []
+        for r in results:
+            output.append(f"Title: {r.get('title')}\nLink: {r.get('href')}\nSnippet: {r.get('body')}\n---")
+        return "\n".join(output)
+    except Exception as e:
+        return f"Error: {str(e)}"
+
+# --- YE HAI NAYA IMAGE JUGAAD TOOL ---
+@tool
+async def generate_image(prompt: str) -> str:
+    """
+    Generates an image based on the user's prompt using AI.
+    Use this when user asks to 'draw', 'create image', or 'paint'.
+    Returns the file path of the generated image.
+    """
+    logger.info(f"🎨 Generating Image for: {prompt}")
+    
+    # Fallback payload
+    payload = {"inputs": prompt}
+    
+    try:
+        # HuggingFace API Call
+        response = await http_client.post(HF_API_URL, headers=headers, json=payload)
+        
+        if response.status_code != 200:
+            return f"Image Gen Failed: {response.text}"
+        
+        # Image Bytes ko file mein save karna
+        image_bytes = response.content
+        image = Image.open(io.BytesIO(image_bytes))
+        
+        # Unique filename banate hain
+        filename = f"static/images/gen_{asyncio.get_event_loop().time()}.png"
+        
+        # Thread mein save karo taaki server freeze na ho
+        await asyncio.to_thread(image.save, filename)
+        
+        return f"Image generated successfully! View at: {filename}"
+    except Exception as e:
+        return f"Image System Error: {str(e)}"
+
+tools = [web_search, generate_image]
+
+# --------------------------------------------------------------------------------------
+# 3. Agent Setup
+# --------------------------------------------------------------------------------------
+
+class AgentState(TypedDict):
+    messages: Annotated[List[BaseMessage], "add_messages"]
+
+llm = ChatOllama(
+    model=MODEL_NAME,
+    base_url=BASE_URL,
+    temperature=0.3,
+).bind_tools(tools)
+
+SYSTEM_PROMPT = """You are a smart AI assistant capable of Web Search and Image Generation.
+
+RULES:
+1. If user asks to DRAW, PAINT, or GENERATE an image, use 'generate_image'.
+2. If user asks for Info, use 'web_search'.
+3. If you generate an image, tell the user the file path returned by the tool.
+"""
+
+async def agent_node(state: AgentState):
+    messages = [SystemMessage(content=SYSTEM_PROMPT)] + state["messages"]
+    response = await llm.ainvoke(messages)
+    return {"messages": [response]}
+
+workflow = StateGraph(AgentState)
+workflow.add_node("agent", agent_node)
+workflow.add_node("tools", ToolNode(tools))
+workflow.add_edge(START, "agent")
+workflow.add_conditional_edges("agent", lambda s: "tools" if s["messages"][-1].tool_calls else END)
+workflow.add_edge("tools", "agent")
+
+memory = MemorySaver()
+app_graph = workflow.compile(checkpointer=memory)
+
+# --------------------------------------------------------------------------------------
+# 4. API Endpoints
+# --------------------------------------------------------------------------------------
+
+class ChatRequest(BaseModel):
+    query: str
+    thread_id: str
+
+# Serve static files so you can see images in browser
+from fastapi.staticfiles import StaticFiles
+app.mount("/static", StaticFiles(directory="static"), name="static")
+
+async def event_generator(query: str, thread_id: str):
+    config = {"configurable": {"thread_id": thread_id}}
+    inputs = {"messages": [HumanMessage(content=query)]}
+    
+    yield "🤖 **Agent Active...**\n\n"
+    
+    async for event in app_graph.astream_events(inputs, config=config, version="v1"):
+        event_type = event["event"]
+        
+        if event_type == "on_chat_model_stream":
+            chunk = event["data"]["chunk"].content
+            if chunk: yield chunk
+
+        elif event_type == "on_tool_start":
+            tool_name = event['name']
+            if tool_name == "generate_image":
+                yield f"\n\n🎨 **Artist Mode On:** Painting '{event['data'].get('input')}'. Please wait...\n\n"
+            else:
+                yield f"\n\n🔎 **Searching:** {tool_name}...\n\n"
+
+        elif event_type == "on_tool_end":
+            output = str(event['data'].get('output'))
+            if "static/images" in output:
+                 # Markdown image syntax for frontend
+                yield f"\n\n🖼️ **Image Ready:**\n![]({output})\n\n"
+            else:
+                yield f"✅ **Data:** {output[:100]}...\n\n"
+
+@app.post("/chat")
+async def chat_endpoint(req: ChatRequest):
+    return StreamingResponse(event_generator(req.query, req.thread_id), media_type="text/plain")

requirements.txt

@@ -0,0 +1,11 @@
+fastapi 
+uvicorn 
+httpx 
+duckduckgo-search 
+langchain-ollama 
+langchain-core 
+langgraph 
+beautifulsoup4 
+async-lru
+pillow
+huggingface_hub