Update app.py

app.py

@@ -1,16 +1,29 @@
 from fastapi import FastAPI, UploadFile, File
-import json, re, io
+import json, re, io, os, requests
 from llama_cpp import Llama
 from PyPDF2 import PdfReader
 from docx import Document
-import os
 
-# ✅ Load Mistral 7B Model from Hugging Face Model Hub
-MODEL_PATH = "TheBloke/CapybaraHermes-2.5-Mistral-7B-GGUF"
-
-print(f"🔹 Loading Mistral 7B from Hugging Face Model Hub: {MODEL_PATH} (This may take a while)")
-
-llm = Llama.from_pretrained(MODEL_PATH, n_ctx=4096, n_gpu_layers=-1)  # Use GPU if available
+# ✅ Define model URL and path
+MODEL_URL = "https://huggingface.co/TheBloke/CapybaraHermes-2.5-Mistral-7B-GGUF/resolve/main/capybarahermes-2.5-mistral-7b.Q5_K_M.gguf"
+MODEL_DIR = "/app/models"
+MODEL_PATH = os.path.join(MODEL_DIR, "mistral-7b.Q5_K_M.gguf")
+
+# ✅ Ensure models directory exists
+os.makedirs(MODEL_DIR, exist_ok=True)
+
+# ✅ Download model if not already available
+if not os.path.exists(MODEL_PATH):
+    print(f"🔹 Downloading model from: {MODEL_URL}")
+    response = requests.get(MODEL_URL, stream=True)
+    with open(MODEL_PATH, "wb") as file:
+        for chunk in response.iter_content(chunk_size=8192):
+            file.write(chunk)
+    print("✅ Model downloaded successfully!")
+
+# ✅ Load Mistral 7B using llama_cpp
+print(f"🔹 Loading Mistral 7B from {MODEL_PATH} (This may take a while)")
+llm = Llama(model_path=MODEL_PATH, n_ctx=4096, n_gpu_layers=-1)  # Use GPU if available
 print("✅ Model loaded successfully!")
 
 app = FastAPI(title="Resume Parsing API", description="Extracts key details from resumes using Mistral 7B")