stefanw / scrapa

Python 3 AsyncIO powered scraping framework with batteries included

scrapa

A scraping library on top of Python 3 asyncio and aiohttp.

Example

# example.py
from itertools import chain

import scrapa
from scrapa.storage.database import DatabaseStorage


class MyScraper(scrapa.Scraper):
    BASE_URL = 'http://example.org/'

    # All methods that do IO need to be marked as coroutines
    async def start(self):
        # Get links on a page
        page_links = yield from self.get_page_links()
        # Get more links on all these pages in parallel.
        # This returns a set of link generators (one result for each task)
        page_link_sets = yield from self.run_many(self.get_page_links, page_links)
        # Chain the generators in these sets to make one generator
        page_link_sets = chain.from_iterable(page_link_sets)
        # Schedule tasks to get all images from these pages
        # Scheduled tasks are put in a queue without blocking
        yield from self.schedule_many(self.get_images, page_link_sets)
        # This returns immediately and the queue is processed in the event loop
        print('DONE')

    async def get_page_links(self, url=''):
        # Get lxml DOM from url
        doc = yield from self.get_dom(url)
        # Return a generator of urls (but could return anything)
        return (link.attrib['href'] for link in doc.xpath('.//a'))

    # Store these tasks
    @scrapa.store
    async def get_images(self, url):
        doc = yield from self.get_dom(url)
        for i in doc.xpath('.//img'):
            # Store result with id, type and some data
            yield from self.store_result(i.attrib['src'], 'image', i.attrib['alt'])


def main():
    path = os.path.join(os.getcwd(), 'tasks.db'),
    db_url = 'sqlite:///%s' % path
    scraper = MyScraper(storage=DatabaseStorage(db_url=db_url))
    # Run with command line arguments
    scraper.run_from_cli()

if __name__ == '__main__':
    main()

Save this to a file and Run it like this with python >= 3.4:

python example.py scrape

Dump pending tasks and split to 3 files:

python example.py dump_tasks -s 3

Load tasks from dump file:

python example.py load_tasks -f scrapa-001.json