main.py

import boto3
from dotenv import load_dotenv
from flask import Flask, jsonify, request
from llama_index.vector_stores.opensearch import (
    OpensearchVectorClient,
    OpensearchVectorStore,
)
from opensearchpy import OpenSearch, RequestsHttpConnection

from config import Config
from core.controller.vector import VectorController
from core.service.llama_index_service import LlamaIndexService
from core.service.opensearch_service import OpensearchService
from core.service.s3_service import S3Service
from core.usecase.vectorizer import VectorizerUsecase
from core.utils.definitions import MAPPINGS
from core.utils.logger import logger

load_dotenv()

app = Flask(__name__)


def initialize_app():
    cfg = Config()
    app.config.from_object(cfg)

    s3_client = boto3.client("s3")
    s3_service = S3Service(s3_client, logger)

    host = {"host": cfg.OPENSEARCH_HOST, "port": cfg.OPENSEARCH_PORT}

    # Opensearch initialization
    try:
        opensearch_client = OpenSearch(
            hosts=[host],
            http_auth=(cfg.OPENSEARCH_USER, cfg.OPENSEARCH_PASS),
            use_ssl=True,
            verify_certs=True,
            connection_class=RequestsHttpConnection,
        )
        if not opensearch_client.indices.exists(index=cfg.OPENSEARCH_INDEX):
            opensearch_client.indices.create(index=cfg.OPENSEARCH_INDEX, body=MAPPINGS)
    except Exception as e:
        logger.error(f"Failed to connect to OpenSearch: {e}")
        raise

    opensearch_service = OpensearchService(
        opensearch_client, cfg.OPENSEARCH_INDEX, logger
    )

    text_field = "content"
    embedding_field = "embedding"
    try:
        os_vector_client = OpensearchVectorClient(
            cfg.OPENSEARCH_HOST,
            cfg.OPENSEARCH_INDEX,
            384,
            embedding_field=embedding_field,
            text_field=text_field,
            http_auth=(cfg.OPENSEARCH_USER, cfg.OPENSEARCH_PASS),
            use_ssl=True,
            verify_certs=True,
            connection_class=RequestsHttpConnection,
            timeout=10,
            port="",
        )
        vector_store = OpensearchVectorStore(os_vector_client)
    except Exception as e:
        logger.error(f"Failed to initialize OpensearchVectorClient: {e}")
        raise

    llama_service = LlamaIndexService(vector_store, logger)
    usecase = VectorizerUsecase(s3_service, llama_service, opensearch_service, logger)
    controller = VectorController(usecase, logger)

    return controller


controller = initialize_app()


@app.route("/v1/api/vectorize", methods=["POST"])
def vectorize():
    try:
        request_data = request.get_json()
        return controller.vectoring(request_data)
    except Exception as e:
        return jsonify({"error": "Failed to decode JSON object: " + str(e)}), 400


@app.route("/v1/api/search", methods=["POST"])
def search():
    try:
        request_data = request.get_json()
        return controller.search(request_data)
    except Exception as e:
        return jsonify({"error": "Failed to decode JSON object: " + str(e)}), 400


@app.route("/health", methods=["GET"])
def health():
    return jsonify({"status": "healthy"}), 200


if __name__ == "__main__":
    app.run(debug=True)