dsProject/dsLightRag/Util/LightRagUtil.py

import logging
import logging.config
import os
import numpy as np
from lightrag import LightRAG
from lightrag.kg.shared_storage import initialize_pipeline_status
from lightrag.llm.openai import openai_complete_if_cache, openai_embed
from lightrag.rerank import jina_rerank, custom_rerank
from lightrag.utils import EmbeddingFunc, logger, set_verbose_debug

import Config.Config
from Config.Config import *


async def print_stream(stream):
    async for chunk in stream:
        if chunk:
            print(chunk, end="", flush=True)


def configure_logging():
    for logger_name in ["uvicorn", "uvicorn.access", "uvicorn.error", "lightrag"]:
        logger_instance = logging.getLogger(logger_name)
        logger_instance.handlers = []
        logger_instance.filters = []

    log_dir = os.getenv("LOG_DIR", os.getcwd())
    log_file_path = os.path.abspath(
        os.path.join(log_dir, "./Logs/lightrag.log")
    )

    print(f"\nLightRAG log file: {log_file_path}\n")
    os.makedirs(os.path.dirname(log_dir), exist_ok=True)

    log_max_bytes = int(os.getenv("LOG_MAX_BYTES", 10485760))
    log_backup_count = int(os.getenv("LOG_BACKUP_COUNT", 5))

    logging.config.dictConfig(
        {
            "version": 1,
            "disable_existing_loggers": False,
            "formatters": {
                "default": {
                    "format": "%(levelname)s: %(message)s",
                },
                "detailed": {
                    "format": "%(asctime)s - %(name)s - %(levelname)s - %(message)s",
                },
            },
            "handlers": {
                "console": {
                    "formatter": "default",
                    "class": "logging.StreamHandler",
                    "stream": "ext://sys.stderr",
                },
                "file": {
                    "formatter": "detailed",
                    "class": "logging.handlers.RotatingFileHandler",
                    "filename": log_file_path,
                    "maxBytes": log_max_bytes,
                    "backupCount": log_backup_count,
                    "encoding": "utf-8",
                },
            },
            "loggers": {
                "lightrag": {
                    "handlers": ["console", "file"],
                    "level": "INFO",
                    "propagate": False,
                },
            },
        }
    )
    logger.setLevel(logging.INFO)
    set_verbose_debug(os.getenv("VERBOSE_DEBUG", "false").lower() == "true")


async def llm_model_func(
        prompt, system_prompt=None, history_messages=None, **kwargs
) -> str:
    return await openai_complete_if_cache(
        os.getenv("LLM_MODEL", LLM_MODEL_NAME),
        prompt,
        system_prompt=system_prompt,
        history_messages=history_messages,
        api_key=LLM_API_KEY,
        base_url=LLM_BASE_URL,
        **kwargs,
    )


async def embedding_func(texts: list[str]) -> np.ndarray:
    return await openai_embed(
        texts,
        model=EMBED_MODEL_NAME,
        api_key=EMBED_API_KEY,
        base_url=EMBED_BASE_URL
    )

async def rerank_func(query: str, documents: list, top_k: int = None, **kwargs):
    return await custom_rerank(
        query=query,
        documents=documents,
        model=Config.Config.RERANK_MODEL,
        base_url=Config.Config.RERANK_BASE_URL,
        api_key=Config.Config.RERANK_BINDING_API_KEY,
        top_k=top_k or 10,
        **kwargs,
    )

async def initialize_rag(working_dir, graph_storage=None):
    if graph_storage is None:
        graph_storage = 'NetworkXStorage'
    rag = LightRAG(
        working_dir=working_dir,
        llm_model_func=llm_model_func,
        graph_storage=graph_storage,
        embedding_func=EmbeddingFunc(
            embedding_dim=EMBED_DIM,
            max_token_size=EMBED_MAX_TOKEN_SIZE,
            func=embedding_func
        ),
        rerank_model_func=rerank_func,
    )
    await rag.initialize_storages()
    await initialize_pipeline_status()
    return rag


def create_llm_model_func():
    def llm_model_func(prompt, system_prompt=None, history_messages=[], **kwargs):
        return openai_complete_if_cache(
            LLM_MODEL_NAME,
            prompt,
            system_prompt=system_prompt,
            history_messages=history_messages,
            api_key=LLM_API_KEY,
            base_url=LLM_BASE_URL,
            **kwargs,
        )

    return llm_model_func


def create_embedding_func():
    return EmbeddingFunc(
        embedding_dim=1024,
        max_token_size=8192,
        func=lambda texts: openai_embed(
            texts,
            model=EMBED_MODEL_NAME,
            api_key=EMBED_API_KEY,
            base_url=EMBED_BASE_URL,
        ),
    )



# AGE
os.environ["POSTGRES_HOST"] = POSTGRES_HOST
os.environ["POSTGRES_PORT"] = str(POSTGRES_PORT)
os.environ["POSTGRES_USER"] = POSTGRES_USER
os.environ["POSTGRES_PASSWORD"] = POSTGRES_PASSWORD
os.environ["POSTGRES_DATABASE"] = POSTGRES_DATABASE


async def initialize_pg_rag(WORKING_DIR, workspace='default'):
    rag = LightRAG(
        working_dir=WORKING_DIR,
        llm_model_func=llm_model_func,
        llm_model_name=LLM_MODEL_NAME,
        llm_model_max_async=4,
        llm_model_max_token_size=32768,
        enable_llm_cache_for_entity_extract=True,
        embedding_func=EmbeddingFunc(
            embedding_dim=EMBED_DIM,
            max_token_size=EMBED_MAX_TOKEN_SIZE,
            func=embedding_func
        ),
        rerank_model_func=rerank_func,
        kv_storage="PGKVStorage",
        doc_status_storage="PGDocStatusStorage",
        graph_storage="PGGraphStorage",
        vector_storage="PGVectorStorage",
        auto_manage_storages_states=False,
        vector_db_storage_cls_kwargs={"workspace": workspace}
    )

    await rag.initialize_storages()
    await initialize_pipeline_status()

    return rag