app.py

from fastapi import FastAPI, File, UploadFile
import torch
from dotenv import load_dotenv
import logging
import os
from PIL import Image
from transformers import CLIPProcessor, CLIPModel


load_dotenv()
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

app = FastAPI(title='CLIP API',
              description='Returns CLIP embedding for text and image')

HF_TOKEN = os.getenv('hf_token')

logger.info("Loading CLIP processor and model")
try:
    processor = CLIPProcessor.from_pretrained(
        "openai/clip-vit-large-patch14", use_auth_token=HF_TOKEN
    )
    clip_model = CLIPModel.from_pretrained(
        "openai/clip-vit-large-patch14", use_auth_token=HF_TOKEN)
    clip_model.eval()
    logger.info("CLIP model loaded successfully")
except Exception as e:
    logger.error(f"Failed to load CLIP model : {e}")
    raise


def get_text_embedding(text: str):
    try:
        inputs = processor(text=[text], return_tensors="pt",
                           padding=True, truncation=True)
        with torch.no_grad():
            text_embeddings = clip_model.get_text_features(**inputs)
        logger.info("Text embedding generated")
        return text_embeddings.squeeze(0).tolist()
    except Exception as e:
        logger.error(f"Error while generating embedding : {e}")
        raise


@app.get("/")
async def root():
    logger.info("Root endpoint accessed")
    return {"message": "Welcome to the CLIP embedding API."}


@app.get("/embedding")
async def get_embedding_text(text: str):
    logger.info(f"Embedding endpoint called with text")
    embedding = get_text_embedding(text)
    return {"embedding": embedding, "dimension": len(embedding)}


@app.post("/clip/process")
async def process_image(file: UploadFile = File(...)):
    logger.info("Processing image")
    image = Image.open(file.file).convert("RGB")
    inputs = processor(images=image, return_tensors="pt")
    with torch.no_grad():
        embeddings = clip_model.get_image_features(**inputs)
    return {"embedding": embeddings.tolist()}