Spaces:
Sleeping
Sleeping
Download old/services/document_loader.py from harshvisualz/vgecbot: direct link, hf CLI and curl.
- Browser
- Download file 899 Bytes
-
https://huggingface.co/spaces/harshvisualz/vgecbot/resolve/main/old/services/document_loader.py
- Command line
-
hf download hf://spaces/harshvisualz/vgecbot/old/services/document_loader.py
-
curl -L -o document_loader.py https://huggingface.co/spaces/harshvisualz/vgecbot/resolve/main/old/services/document_loader.py
899 Bytes
| from langchain_core.documents import Document | |
| from langchain_community.document_loaders import PyPDFLoader, DirectoryLoader | |
| from langchain_community.document_loaders import UnstructuredMarkdownLoader | |
| from pathlib import Path | |
| class document_loader: | |
| def __init__(self, filepath, glob: str = "*.pdf"): | |
| self.filepath = filepath | |
| self.glob = glob | |
| self.loader = PyPDFLoader | |
| def load(self): | |
| doc_loader = PyPDFLoader(self.filepath) | |
| return doc_loader.load() | |
| def load_md(self): | |
| return UnstructuredMarkdownLoader(self.filepath).load() | |
| def lazy_load(self): | |
| doc_loader = PyPDFLoader(self.filepath) | |
| return doc_loader.lazy_load() | |
| def load_multiple(self): | |
| doc_loader = DirectoryLoader( | |
| self.filepath, | |
| glob=self.glob, | |
| loader_cls=PyPDFLoader | |
| ) | |
| return doc_loader.load() | |