mirror of https://github.com/hwchase17/langchain
You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
209 lines
7.4 KiB
Python
209 lines
7.4 KiB
Python
9 months ago
|
"""Loader that uses Playwright to load a page, then uses unstructured to load the html.
|
||
|
"""
|
||
|
import logging
|
||
|
from abc import ABC, abstractmethod
|
||
|
from typing import TYPE_CHECKING, List, Optional
|
||
|
|
||
|
from langchain_core.documents import Document
|
||
|
|
||
|
from langchain_community.document_loaders.base import BaseLoader
|
||
|
|
||
|
if TYPE_CHECKING:
|
||
|
from playwright.async_api import Browser as AsyncBrowser
|
||
|
from playwright.async_api import Page as AsyncPage
|
||
|
from playwright.async_api import Response as AsyncResponse
|
||
|
from playwright.sync_api import Browser, Page, Response
|
||
|
|
||
|
|
||
|
logger = logging.getLogger(__name__)
|
||
|
|
||
|
|
||
|
class PlaywrightEvaluator(ABC):
|
||
|
"""Abstract base class for all evaluators.
|
||
|
|
||
|
Each evaluator should take a page, a browser instance, and a response
|
||
|
object, process the page as necessary, and return the resulting text.
|
||
|
"""
|
||
|
|
||
|
@abstractmethod
|
||
|
def evaluate(self, page: "Page", browser: "Browser", response: "Response") -> str:
|
||
|
"""Synchronously process the page and return the resulting text.
|
||
|
|
||
|
Args:
|
||
|
page: The page to process.
|
||
|
browser: The browser instance.
|
||
|
response: The response from page.goto().
|
||
|
|
||
|
Returns:
|
||
|
text: The text content of the page.
|
||
|
"""
|
||
|
pass
|
||
|
|
||
|
@abstractmethod
|
||
|
async def evaluate_async(
|
||
|
self, page: "AsyncPage", browser: "AsyncBrowser", response: "AsyncResponse"
|
||
|
) -> str:
|
||
|
"""Asynchronously process the page and return the resulting text.
|
||
|
|
||
|
Args:
|
||
|
page: The page to process.
|
||
|
browser: The browser instance.
|
||
|
response: The response from page.goto().
|
||
|
|
||
|
Returns:
|
||
|
text: The text content of the page.
|
||
|
"""
|
||
|
pass
|
||
|
|
||
|
|
||
|
class UnstructuredHtmlEvaluator(PlaywrightEvaluator):
|
||
|
"""Evaluates the page HTML content using the `unstructured` library."""
|
||
|
|
||
|
def __init__(self, remove_selectors: Optional[List[str]] = None):
|
||
|
"""Initialize UnstructuredHtmlEvaluator."""
|
||
|
try:
|
||
|
import unstructured # noqa:F401
|
||
|
except ImportError:
|
||
|
raise ImportError(
|
||
|
"unstructured package not found, please install it with "
|
||
|
"`pip install unstructured`"
|
||
|
)
|
||
|
|
||
|
self.remove_selectors = remove_selectors
|
||
|
|
||
|
def evaluate(self, page: "Page", browser: "Browser", response: "Response") -> str:
|
||
|
"""Synchronously process the HTML content of the page."""
|
||
|
from unstructured.partition.html import partition_html
|
||
|
|
||
|
for selector in self.remove_selectors or []:
|
||
|
elements = page.locator(selector).all()
|
||
|
for element in elements:
|
||
|
if element.is_visible():
|
||
|
element.evaluate("element => element.remove()")
|
||
|
|
||
|
page_source = page.content()
|
||
|
elements = partition_html(text=page_source)
|
||
|
return "\n\n".join([str(el) for el in elements])
|
||
|
|
||
|
async def evaluate_async(
|
||
|
self, page: "AsyncPage", browser: "AsyncBrowser", response: "AsyncResponse"
|
||
|
) -> str:
|
||
|
"""Asynchronously process the HTML content of the page."""
|
||
|
from unstructured.partition.html import partition_html
|
||
|
|
||
|
for selector in self.remove_selectors or []:
|
||
|
elements = await page.locator(selector).all()
|
||
|
for element in elements:
|
||
|
if await element.is_visible():
|
||
|
await element.evaluate("element => element.remove()")
|
||
|
|
||
|
page_source = await page.content()
|
||
|
elements = partition_html(text=page_source)
|
||
|
return "\n\n".join([str(el) for el in elements])
|
||
|
|
||
|
|
||
|
class PlaywrightURLLoader(BaseLoader):
|
||
|
"""Load `HTML` pages with `Playwright` and parse with `Unstructured`.
|
||
|
|
||
|
This is useful for loading pages that require javascript to render.
|
||
|
|
||
|
Attributes:
|
||
|
urls (List[str]): List of URLs to load.
|
||
|
continue_on_failure (bool): If True, continue loading other URLs on failure.
|
||
|
headless (bool): If True, the browser will run in headless mode.
|
||
|
"""
|
||
|
|
||
|
def __init__(
|
||
|
self,
|
||
|
urls: List[str],
|
||
|
continue_on_failure: bool = True,
|
||
|
headless: bool = True,
|
||
|
remove_selectors: Optional[List[str]] = None,
|
||
|
evaluator: Optional[PlaywrightEvaluator] = None,
|
||
|
):
|
||
|
"""Load a list of URLs using Playwright."""
|
||
|
try:
|
||
|
import playwright # noqa:F401
|
||
|
except ImportError:
|
||
|
raise ImportError(
|
||
|
"playwright package not found, please install it with "
|
||
|
"`pip install playwright`"
|
||
|
)
|
||
|
|
||
|
self.urls = urls
|
||
|
self.continue_on_failure = continue_on_failure
|
||
|
self.headless = headless
|
||
|
|
||
|
if remove_selectors and evaluator:
|
||
|
raise ValueError(
|
||
|
"`remove_selectors` and `evaluator` cannot be both not None"
|
||
|
)
|
||
|
|
||
|
# Use the provided evaluator, if any, otherwise, use the default.
|
||
|
self.evaluator = evaluator or UnstructuredHtmlEvaluator(remove_selectors)
|
||
|
|
||
|
def load(self) -> List[Document]:
|
||
|
"""Load the specified URLs using Playwright and create Document instances.
|
||
|
|
||
|
Returns:
|
||
|
List[Document]: A list of Document instances with loaded content.
|
||
|
"""
|
||
|
from playwright.sync_api import sync_playwright
|
||
|
|
||
|
docs: List[Document] = list()
|
||
|
|
||
|
with sync_playwright() as p:
|
||
|
browser = p.chromium.launch(headless=self.headless)
|
||
|
for url in self.urls:
|
||
|
try:
|
||
|
page = browser.new_page()
|
||
|
response = page.goto(url)
|
||
|
if response is None:
|
||
|
raise ValueError(f"page.goto() returned None for url {url}")
|
||
|
|
||
|
text = self.evaluator.evaluate(page, browser, response)
|
||
|
metadata = {"source": url}
|
||
|
docs.append(Document(page_content=text, metadata=metadata))
|
||
|
except Exception as e:
|
||
|
if self.continue_on_failure:
|
||
|
logger.error(
|
||
|
f"Error fetching or processing {url}, exception: {e}"
|
||
|
)
|
||
|
else:
|
||
|
raise e
|
||
|
browser.close()
|
||
|
return docs
|
||
|
|
||
|
async def aload(self) -> List[Document]:
|
||
|
"""Load the specified URLs with Playwright and create Documents asynchronously.
|
||
|
Use this function when in a jupyter notebook environment.
|
||
|
|
||
|
Returns:
|
||
|
List[Document]: A list of Document instances with loaded content.
|
||
|
"""
|
||
|
from playwright.async_api import async_playwright
|
||
|
|
||
|
docs: List[Document] = list()
|
||
|
|
||
|
async with async_playwright() as p:
|
||
|
browser = await p.chromium.launch(headless=self.headless)
|
||
|
for url in self.urls:
|
||
|
try:
|
||
|
page = await browser.new_page()
|
||
|
response = await page.goto(url)
|
||
|
if response is None:
|
||
|
raise ValueError(f"page.goto() returned None for url {url}")
|
||
|
|
||
|
text = await self.evaluator.evaluate_async(page, browser, response)
|
||
|
metadata = {"source": url}
|
||
|
docs.append(Document(page_content=text, metadata=metadata))
|
||
|
except Exception as e:
|
||
|
if self.continue_on_failure:
|
||
|
logger.error(
|
||
|
f"Error fetching or processing {url}, exception: {e}"
|
||
|
)
|
||
|
else:
|
||
|
raise e
|
||
|
await browser.close()
|
||
|
return docs
|