functions.py

from langchain.chains import TransformChain
from langchain_core.messages import HumanMessage
from langchain_openai import ChatOpenAI
from langchain import globals
from langchain_core.runnables import chain
import base64
from typing import Dict,List,Union



def load_image(inputs: dict) -> dict:
  """Load image from file and encode it as base64."""
  image_path = inputs["image_path"]

  def encode_image(image_path):
      with open(image_path, "rb") as image_file:
          return base64.b64encode(image_file.read()).decode('utf-8')
  image_base64 = encode_image(image_path)
  return {"image": image_base64}

load_image_chain = TransformChain(
    input_variables=["image_path"],
    output_variables=["image"],
    transform=load_image
)


@chain
def image_model(inputs: dict) -> Union[str, List[str], dict]:
 """Invoke model with image and prompt."""
 model = ChatOpenAI(temperature=0.1, model="gpt-4o", max_tokens=1024)
 parser = inputs["parser"]
 msg = model.invoke(
             [HumanMessage(
             content=[
             {"type": "text", "text": inputs["prompt"]},
             {"type": "text", "text": parser.get_format_instructions()},
             {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{inputs['image']}"}},
             ])]
             )
 return msg.content

def get_image_informations(image_path: str,prompt,parser) -> dict:
   vision_chain = load_image_chain | image_model | parser
   return vision_chain.invoke({'image_path': f'{image_path}', 
                               'prompt': prompt,
                               'parser': parser
                               })