Update app.py

app.py

@@ -1,35 +1,22 @@
 from fastapi import FastAPI
-from pydantic import BaseModel
-from llama_cpp import Llama
-from huggingface_hub import hf_hub_download
+import subprocess
+import sys
 
 app = FastAPI()
 
-# 1. Define your specific model details
-REPO_ID = "SarmaHighOnAI/physics-tutor-gguf"
-FILENAME = "llama-3.2-3b-instruct.Q4_K_M.gguf"
-
-print("Downloading your fine-tuned model...")
-# This downloads the file LOCALLY to the container
-model_path = hf_hub_download(repo_id=REPO_ID, filename=FILENAME)
-
-print("Loading model into memory...")
-# This loads the 'brain' locally. n_threads=2 is safe for the free tier.
-llm = Llama(model_path=model_path, n_ctx=2048, n_threads=2)
-
-class Request(BaseModel):
-    prompt: str
-
 @app.get("/")
 def home():
-    return {"status": "Running", "message": "Physics Tutor API is Live (Local Inference)"}
-
-@app.post("/generate")
-def generate(request: Request):
-    # Standard prompt format
-    formatted_prompt = f"<|start_header_id|>user<|end_header_id|>\n\n{request.prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
-    
-    # Run inference LOCALLY (No API Key needed)
-    output = llm(formatted_prompt, max_tokens=256, stop=["<|eot_id|>"], echo=False)
+    # This forces the server to list all installed packages
+    # Check your "Logs" tab after visiting this page!
+    result = subprocess.run([sys.executable, "-m", "pip", "freeze"], capture_output=True, text=True)
+    print("=== INSTALLED PACKAGES ===")
+    print(result.stdout)
+    print("==========================")
     
-    return {"response": output["choices"][0]["text"]}
+    try:
+        import llama_cpp
+        status = "llama_cpp is INSTALLED!"
+    except ImportError as e:
+        status = f"CRITICAL ERROR: {e}"
+        
+    return {"status": status, "packages_printed_to_logs": True}