Fixed app v2

app.py

@@ -1,47 +1,39 @@
-from fastapi import FastAPI, File, UploadFile, Form
-from transformers import pipeline
+from fastapi import FastAPI, Form
+from llama_cpp import Llama
 from typing import List
 import json
 
 # Initialize FastAPI app
 app = FastAPI()
 
-# Initialize the text generation pipeline with the Qwen model
-pipe = pipeline("text-generation", model="Qwen/Qwen2.5-1.5B-Instruct")
+# Load the Llama model
+llm = Llama.from_pretrained(
+    repo_id="HuggingFaceTB/SmolLM2-360M-Instruct-GGUF",
+    filename="smollm2-360m-instruct-q8_0.gguf",  # Replace with the actual path to your GGUF file
+)
 
-# Endpoint to upload CV file and store CV text
-@app.post("/upload-cv/")
-async def upload_cv(file: UploadFile = File(...)):
-    content = await file.read()
-    cv_text = content.decode("utf-8")
-    return {"cv_text": cv_text}
-
-# Endpoint to compare job descriptions with the CV text
-@app.post("/compare/")
-async def compare_job_cv(job_descriptions: str = Form(...), cv_text: str = Form(...)):
-    # Split job descriptions by line
-    descriptions = job_descriptions.strip().split("\n")
-    results = []
+# Endpoint to generate response from model based on user input
+@app.post("/ask/")
+async def ask_question(prompt: str = Form(...)):
+    # Format the prompt as a chat message
+    messages = [
+        {"role": "user", "content": prompt}
+    ]
+    
+    # Generate a response using Llama
+    response = llm.create_chat_completion(messages=messages)
+    response_content = response["choices"][0]["message"]["content"]
     
-    for description in descriptions:
-        # Create prompt for the Qwen model
-        prompt = f"Compare the following job description with this resume. Job Description: {description}. Resume: {cv_text}. Give a match score and brief analysis."
-        
-        # Generate response using the text generation pipeline
-        response = pipe(prompt, max_length=100, num_return_sequences=1)
-        response_content = response[0]["generated_text"]
-        
-        # Parse response content for score and summary if JSON formatted
-        try:
-            response_data = json.loads(response_content)
-            results.append(response_data)
-        except json.JSONDecodeError:
-            results.append({
-                "Job Description": description,
-                "Analysis": response_content  # Use raw response if JSON parsing fails
-            })
+    return {"response": response_content}
+
+# Endpoint to test a simple query (optional)
+@app.get("/test/")
+async def test():
+    # Test the model with a simple question
+    messages = [{"role": "user", "content": "What is the capital of France?"}]
+    response = llm.create_chat_completion(messages=messages)
+    response_content = response["choices"][0]["message"]["content"]
     
-    return {"results": results}
+    return {"test_response": response_content}
 
-