repositories
loading repo index
repositories
loading repo index
repository
loading code, commits, and activity
public Clawd ADK gateway launch mirror
stars
latest
clone command
git clone gitlawb://did:key:z6Mkq5mY...iFZ5/my-project-publ...git clone gitlawb://did:key:z6Mkq5mY.../my-project-publ...2fa351d6docs: add automaton and perps launch sources16d ago| #1 | import hashlib |
| #2 | import re |
| #3 | |
| #4 | try: |
| #5 | from googleapiclient.errors import HttpError |
| #6 | except ImportError: |
| #7 | raise ImportError( |
| #8 | "Google Drive requires extra dependencies. Install with `pip install embedchain[googledrive]`" |
| #9 | ) from None |
| #10 | |
| #11 | from langchain_community.document_loaders import GoogleDriveLoader as Loader |
| #12 | |
| #13 | try: |
| #14 | import unstructured # noqa: F401 |
| #15 | from langchain_community.document_loaders import UnstructuredFileIOLoader |
| #16 | except ImportError: |
| #17 | raise ImportError( |
| #18 | 'Unstructured file requires extra dependencies. Install with `pip install "unstructured[local-inference, all-docs]"`' # noqa: E501 |
| #19 | ) from None |
| #20 | |
| #21 | from embedchain.helpers.json_serializable import register_deserializable |
| #22 | from embedchain.loaders.base_loader import BaseLoader |
| #23 | |
| #24 | |
| #25 | @register_deserializable |
| #26 | class GoogleDriveLoader(BaseLoader): |
| #27 | @staticmethod |
| #28 | def _get_drive_id_from_url(url: str): |
| #29 | regex = r"^https:\/\/drive\.google\.com\/drive\/(?:u\/\d+\/)folders\/([a-zA-Z0-9_-]+)$" |
| #30 | if re.match(regex, url): |
| #31 | return url.split("/")[-1] |
| #32 | raise ValueError( |
| #33 | f"The url provided {url} does not match a google drive folder url. Example drive url: " |
| #34 | f"https://drive.google.com/drive/u/0/folders/xxxx" |
| #35 | ) |
| #36 | |
| #37 | def load_data(self, url: str): |
| #38 | """Load data from a Google drive folder.""" |
| #39 | folder_id: str = self._get_drive_id_from_url(url) |
| #40 | |
| #41 | try: |
| #42 | loader = Loader( |
| #43 | folder_id=folder_id, |
| #44 | recursive=True, |
| #45 | file_loader_cls=UnstructuredFileIOLoader, |
| #46 | ) |
| #47 | |
| #48 | data = [] |
| #49 | all_content = [] |
| #50 | |
| #51 | docs = loader.load() |
| #52 | for doc in docs: |
| #53 | all_content.append(doc.page_content) |
| #54 | # renames source to url for later use. |
| #55 | doc.metadata["url"] = doc.metadata.pop("source") |
| #56 | data.append({"content": doc.page_content, "meta_data": doc.metadata}) |
| #57 | |
| #58 | doc_id = hashlib.sha256((" ".join(all_content) + url).encode()).hexdigest() |
| #59 | return {"doc_id": doc_id, "data": data} |
| #60 | |
| #61 | except HttpError: |
| #62 | raise FileNotFoundError("Unable to locate folder or files, check provided drive URL and try again") |
| #63 |