imagequerying.py

# import cv2
# import torch
# import ollama
# import base64
# import os
# import time
# from sentence_transformers import SentenceTransformer, util
# import chromadb
# import os
# from langchain.schema import Document  # Import the Document class from LangChain
# import re
# import fitz
# from langchain_chroma import Chroma
# from chromadb.config import Settings, DEFAULT_DATABASE, DEFAULT_TENANT
# from chromadb.utils import embedding_functions
# from langchain.text_splitter import RecursiveCharacterTextSplitter
# from langchain.chains.qa_with_sources.retrieval import RetrievalQAWithSourcesChain
# from langchain_huggingface import HuggingFaceEmbeddings
# from langchain_core.prompts import PromptTemplate
# from langchain_core.output_parsers import StrOutputParser
# from langchain_ollama import ChatOllama


# def vision_model(file_path, query):
#     """Processes an image and queries the LLaMA vision model."""
#     print("<<<<< VISION MODEL STARTED >>>>>")

#     image = cv2.imread(file_path)
#     if image is None:
#         return "Error: Failed to load image."

#     _, buffer = cv2.imencode(".jpg", image)
#     image_base64 = base64.b64encode(buffer).decode("utf-8")

#     prompt = f"""
#     Please describe the following image based on the given query.
#     If the query is not relevant, respond with: 
#     "Sorry, I don't have enough information from this specific image."
    
#     Query: {query}
#     """

#     try:
#         response = ollama.chat(
#             model="llama3.2-vision",
#             messages=[{"role": "user", "content": prompt, "images": [image_base64]}],
#         )
#         return response.get("message", {}).get("content", "").strip()
#     except Exception as e:
#         return f"Error: {str(e)}"