A browser for your agent, built on Chrome and Pyppeteer.
Project description
agentbrowser
A browser for your agent, built on Chrome and Pyppeteer.
Installation
pip install agentbrowser
Usage
Importing into your project
from agentbrowser import (
get_browser,
init_browser,
navigate_to,
get_body_html,
get_body_text,
get_body_text_raw,
get_document_html,
create_page,
close_page,
evaluate_javascript,
)
Quickstart
from agentbrowser import (
navigate_to,
get_body_text,
)
# Navigate to a URL
page = navigate_to("https://google.com")
# Get the text from the page
text = get_body_text(page)
print(text)
Basic:
Create a new page
Equivalent of ctrl+t in Chrome. Makes a new blank page.
page = create_page()
Close a page
Equivalent of ctrl+w in Chrome. Closes the current page.
close_page(page)
Navigate to a URL
Equivalent of typing a URL into the address bar and hitting enter. If you haven't created a page yet, it will create one for you.
page = navigate_to("https://google.com")
Get the HTML of the page
Get the entire document HTML
html = get_document_html(page)
Get the HTML of the body
Get just the HTML of the body and inner. Useful for parsing out the content of the page.
html = get_body_html(page)
Get the text of the body
Get just the text of the body. Unlike the raw function, tries to remove some useless tags and divs and things. Not perfect, though.
text = get_body_text(page)
Get the raw text of the body
Get the raw text of the body. This will include all the tags and divs and things.
text = get_body_text_raw(page)
Advanced Usage
Get browser
This will give you a reference to the browser object, which you can use for advanced stuff. The browser object comes from Pyppeteer, so anything you can do with Pyppeteer, you can do with this.
browser = get_browser()
Evaluate Javascript
Call some Javascript on the page. Equivalent of opening the console and typing in some Javascript.
result = evaluate_javascript(page, "document.title")
Initialize browser
This will initialize the browser object. You can pass headless
and executable_path
. Headless will control whether the actual window appears on screen. Executable path will control which browser is used. By default, it will try to find Chrome first, then fall back to Chromium if it can't find Chrome.
The browser will be auto-initialized by default so you don't need to call this. The only reason you would is because you want to use headful or swap the browser.
init_browser(headless=True, executable_path="/path/to/chrome")
Asynchronous Usage
The library also supports asyncio and offers asynchronous versions of the methods to facilitate non-blocking operations. Here's how to use them:
Importing into your project
from agentbrowser import (
async_get_browser,
async_init_browser,
async_navigate_to,
async_get_body_html,
async_get_body_text,
async_get_body_text_raw,
async_get_document_html,
async_create_page,
async_close_page,
async_evaluate_javascript,
)
Quickstart
import asyncio
from agentbrowser import (
async_navigate_to,
async_get_body_text,
)
async def main():
# Navigate to a URL
page = await async_navigate_to("https://google.com")
# Get the text from the page
text = await async_get_body_text(page)
print(text)
# Run the asyncio event loop
asyncio.run(main())
Basic:
Create a new page
Equivalent of ctrl+t in Chrome. Makes a new blank page.
page = await async_create_page()
Close a page
Equivalent of ctrl+w in Chrome. Closes the current page.
await async_close_page(page)
Navigate to a URL
Equivalent of typing a URL into the address bar and hitting enter. If you haven't created a page yet, it will create one for you.
page = await async_navigate_to("https://google.com")
Get the HTML of the page
Get the entire document HTML
html = await async_get_document_html(page)
Get the HTML of the body
Get just the HTML of the body and inner. Useful for parsing out the content of the page.
html = await async_get_body_html(page)
Get the text of the body
Get just the text of the body. Unlike the raw function, tries to remove some useless tags and divs and things. Not perfect, though.
text = await async_get_body_text(page)
Get the raw text of the body
Get the raw text of the body. This will include all the tags and divs and things.
text = await async_get_body_text_raw(page)
Advanced Usage
Get browser
This will give you a reference to the browser object, which you can use for advanced stuff. The browser object comes from Pyppeteer, so anything you can do with Pyppeteer, you can do with this.
browser = await async_get_browser()
Evaluate Javascript
Call some Javascript on the page. Equivalent of opening the console and typing in some Javascript.
result = await async_evaluate_javascript(page, "document.title")
Initialize browser
This will initialize the browser object. You can pass headless
and executable_path
. Headless will control whether the actual window appears on screen. Executable path will control which browser is used. By default, it will try to find Chrome first, then fall back to Chromium if it can't find Chrome.
The browser will be auto-initialized by default so you don't need to call this. The only reason you would is because you want to use headful or swap the browser.
await async_init_browser(headless=True, executable_path="/path/to/chrome")
Remember to use asyncio.run(main())
to start the asynchronous event loop when using these functions.
Contributions Welcome
If you like this library and want to contribute in any way, please feel free to submit a PR and I will review it. Please note that the goal here is simplicity and accesibility, using common language and few dependencies.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for agentbrowser-0.1.4-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2019c84dfdaa2935906f1c933a29891b4ed6b22ee1d9e814946d75f25a0af219 |
|
MD5 | 2c21147b323b70ab147bd461bc1bf54d |
|
BLAKE2b-256 | 5589223710898fea83a4fc518dbc293e400b88bd97cc0d84ea0a025d8750e453 |