mirror of
https://github.com/run-llama/llama-hub.git
synced 2025-08-13 11:11:48 +00:00
48 lines
1.2 KiB
Python
48 lines
1.2 KiB
Python
"""Simple Web scraper."""
|
|
from typing import List
|
|
|
|
from langchain.requests import RequestsWrapper
|
|
from llama_index.readers.base import BaseReader
|
|
from llama_index.readers.schema.base import Document
|
|
|
|
|
|
class SimpleWebPageReader(BaseReader):
|
|
"""Simple web page reader.
|
|
|
|
Reads pages from the web.
|
|
|
|
Args:
|
|
html_to_text (bool): Whether to convert HTML to text.
|
|
Requires `html2text` package.
|
|
|
|
"""
|
|
|
|
def __init__(self, html_to_text: bool = False) -> None:
|
|
"""Initialize with parameters."""
|
|
self._html_to_text = html_to_text
|
|
|
|
def load_data(self, urls: List[str]) -> List[Document]:
|
|
"""Load data from the input directory.
|
|
|
|
Args:
|
|
urls (List[str]): List of URLs to scrape.
|
|
|
|
Returns:
|
|
List[Document]: List of documents.
|
|
|
|
"""
|
|
if not isinstance(urls, list):
|
|
raise ValueError("urls must be a list of strings.")
|
|
requests = RequestsWrapper()
|
|
documents = []
|
|
for url in urls:
|
|
response = requests.get(url)
|
|
if self._html_to_text:
|
|
import html2text
|
|
|
|
response = html2text.html2text(response)
|
|
|
|
documents.append(Document(response))
|
|
|
|
return documents
|