Create app.py

app.py

@@ -0,0 +1,104 @@
+import os
+import glob
+import json
+import uuid
+from fastapi import FastAPI, Request, HTTPException
+from fastapi.responses import HTMLResponse, StreamingResponse
+from fastapi.staticfiles import StaticFiles
+from fastapi.templating import Jinja2Templates
+from llama_cpp import Llama
+
+app = FastAPI()
+
+# --- Configuration ---
+MODEL_DIR = "."  # Looks for models in the root
+current_model = None
+current_model_name = ""
+
+# Serve static files
+app.mount("/static", StaticFiles(directory="static"), name="static")
+templates = Jinja2Templates(directory="templates")
+
+# --- Model Logic ---
+def get_model(model_name):
+    global current_model, current_model_name
+    
+    if not model_name:
+        raise HTTPException(status_code=400, detail="No model selected")
+        
+    if current_model_name == model_name and current_model is not None:
+        return current_model
+
+    print(f"Loading new model: {model_name}...")
+    try:
+        # Unload previous model to free RAM
+        if current_model is not None:
+            del current_model
+        
+        # Load new model (Optimized for Free Tier)
+        current_model = Llama(
+            model_path=model_name,
+            n_ctx=2048,      # Context window
+            n_threads=2,     # CPU threads (Free tier limit)
+            n_batch=512,
+            verbose=False
+        )
+        current_model_name = model_name
+        return current_model
+    except Exception as e:
+        print(f"Load Error: {e}")
+        raise HTTPException(status_code=500, detail=f"Failed to load {model_name}")
+
+# --- Routes ---
+
+@app.get("/", response_class=HTMLResponse)
+async def read_root(request: Request):
+    return templates.TemplateResponse("index.html", {"request": request})
+
+@app.get("/api/models")
+async def list_models():
+    # Scans for .gguf files
+    models = glob.glob("*.gguf")
+    return {"models": models}
+
+@app.post("/api/chat")
+async def chat(request: Request):
+    data = await request.json()
+    user_input = data.get("message")
+    model_file = data.get("model")
+    history = data.get("history", []) # Receive conversation history if needed
+
+    llm = get_model(model_file)
+
+    # Stream Generator
+    def iter_response():
+        # System Prompt for Hannah
+        prompt = f"""<|im_start|>system
+You are Hannah, a highly intelligent and helpful AI assistant similar to Gemini and ChatGPT.
+<|im_end|>
+<|im_start|>user
+{user_input}<|im_end|>
+<|im_start|>assistant
+"""
+        stream = llm(
+            prompt,
+            max_tokens=1024,
+            stop=["<|im_end|>", "User:", "System:"],
+            stream=True,
+            temperature=0.7
+        )
+        for output in stream:
+            text = output['choices'][0]['text']
+            yield json.dumps({"text": text}) + "\n"
+
+    return StreamingResponse(iter_response(), media_type="application/x-ndjson")
+
+@app.post("/api/gen_title")
+async def gen_title(request: Request):
+    # Simple logic to generate a 3-4 word title from the first message
+    data = await request.json()
+    message = data.get("message", "")
+    # In a real app, we'd ask the AI to summarize this. For speed:
+    words = message.split()[:4]
+    title = " ".join(words).capitalize() + "..."
+    return {"title": title}