LangChain文档加载器中的UnstructedLoader需要什么API Key？

Question

我目前正在 Node.js 项目中使用 LangChain 库，并尝试使用

langchain/document_loaders/fs/unstructured

模块中的 UnstructedLoader。这是我的代码片段：

import { UnstructuredLoader } from "langchain/document_loaders/fs/unstructured";

const options = {
  apiKey: "MY_API_KEY",
};

const loader = new UnstructuredLoader(
  "src/document_loaders/example_data/notion.md",
  options
);
const docs = await loader.load();

我不确定

apiKey

的选项对象中是否需要

UnstructuredLoader

。有人可以澄清这应该是什么密钥以及我可以在哪里获得它吗？是浪链提供的特定API密钥，还是指外部服务API密钥，例如AWS或其他云提供商？

此外，如果此密钥正常工作需要任何特定的配置步骤或权限，我将不胜感激。

我传递 Open AI API 密钥时出错：

2023-12-06T11:21:21.704Z    65bd9f86-55cc-5ddf-9fb5-c26c0eac0992    ERROR   Invoke Error    {
    "errorType": "Error",
    "errorMessage": "Failed to partition file /tmp/README.md with error 401 and message {\"detail\":\"API key is malformed, please type the API key correctly in the header.\"}",
    "stack": [
        "Error: Failed to partition file /tmp/README.md with error 401 and message {\"detail\":\"API key is malformed, please type the API key correctly in the header.\"}",
        "    at UnstructuredLoader._partition (/var/task/node_modules/langchain/dist/document_loaders/fs/unstructured.cjs:189:19)",
        "    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)",
        "    at async UnstructuredLoader.load (/var/task/node_modules/langchain/dist/document_loaders/fs/unstructured.cjs:198:26)",
        "    at async processFile (/var/task/index.js:103:21)",
        "    at async Runtime.handler (/var/task/index.js:151:5)"
    ]
}

Answer 1

这是可选的。当您从 api 加载文档时会使用它。来自Python源代码：

def get_elements_from_api(
    file_path: Union[str, List[str], None] = None,
    file: Union[IO, Sequence[IO], None] = None,
    api_url: str = "https://api.unstructured.io/general/v0/general",
    api_key: str = "",
    **unstructured_kwargs: Any,
) -> List:
    """Retrieves a list of elements from the Unstructured API."""
    if isinstance(file, collections.abc.Sequence) or isinstance(file_path, list):
        from unstructured.partition.api import partition_multiple_via_api

        _doc_elements = partition_multiple_via_api(
            filenames=file_path,
            files=file,
            api_key=api_key,
            api_url=api_url,
            **unstructured_kwargs,
        )

        elements = []
        for _elements in _doc_elements:
            elements.extend(_elements)

        return elements
    else:
        from unstructured.partition.api import partition_via_api

        return partition_via_api(
            filename=file_path,
            file=file,
            api_key=api_key,
            api_url=api_url,
            **unstructured_kwargs,
        )