Source code for langchain_community.document_loaders.gcs_directory

import logging
from typing import Callable, List, Optional

from langchain_core._api.deprecation import deprecated
from langchain_core.documents import Document

from langchain_community.document_loaders.base import BaseLoader
from langchain_community.document_loaders.gcs_file import GCSFileLoader
from langchain_community.utilities.vertexai import get_client_info

logger = logging.getLogger(__name__)


@deprecated(
    since="0.0.32",
    removal="0.3.0",
    alternative_import="langchain_google_community.GCSDirectoryLoader",
)
class GCSDirectoryLoader(BaseLoader):
    """从GCS目录加载。"""

    def __init__(
        self,
        project_name: str,
        bucket: str,
        prefix: str = "",
        loader_func: Optional[Callable[[str], BaseLoader]] = None,
        continue_on_failure: bool = False,
    ):
        """初始化使用存储桶和键名。

参数：
    project_name：GCS存储桶所属项目的ID。
    bucket：GCS存储桶的名称。
    prefix：GCS存储桶的前缀。
    loader_func：一个加载器函数，根据file_path参数实例化一个加载器。如果未提供任何内容，GCSFileLoader将使用其默认加载器。
    continue_on_failure：对GCS目录中的每个文件使用try-except块。如果设置为`True`，则无法处理文件不会导致错误。
"""
        self.project_name = project_name
        self.bucket = bucket
        self.prefix = prefix
        self._loader_func = loader_func
        self.continue_on_failure = continue_on_failure

    def load(self) -> List[Document]:
        """加载文档。"""
        try:
            from google.cloud import storage
        except ImportError:
            raise ImportError(
                "Could not import google-cloud-storage python package. "
                "Please install it with `pip install google-cloud-storage`."
            )
        client = storage.Client(
            project=self.project_name,
            client_info=get_client_info(module="google-cloud-storage"),
        )
        docs = []
        for blob in client.list_blobs(self.bucket, prefix=self.prefix):
            # we shall just skip directories since GCSFileLoader creates
            # intermediate directories on the fly
            if blob.name.endswith("/"):
                continue
            # Use the try-except block here
            try:
                loader = GCSFileLoader(
                    self.project_name,
                    self.bucket,
                    blob.name,
                    loader_func=self._loader_func,
                )
                docs.extend(loader.load())
            except Exception as e:
                if self.continue_on_failure:
                    logger.warning(f"Problem processing blob {blob.name}, message: {e}")
                    continue
                else:
                    raise e
        return docs