ScrapeMeAgain

ScrapeMeAgain is a Python 3 powered web scraper. It uses multithreading (ThreadPoolExecutor) and multiprocessing to get the work done quicker and stores collected data in an SQLite database.

Installation

pip install scrapemeagain

System requirements

Tor in combination with Privoxy are used for anonymity (i.e. regular IP address changes).

Using Docker and Docker Compose is the preferred (and easier) way to use ScrapeMeAgain.

You have to manually install and setup Tor and Privoxy on your system if not using Docker. For further information about installation and configuration refer to:

• A step-by-step guide how to use Python with Tor and Privoxy
• Crawling anonymously with Tor in Python (alternative link (Gist))

Usage

You have to provide your own database table description and an actual scraper class which must follow the BaseScraper interface. See examples/examplescraper for more details.

Dockerized

With Docker it is possible to use multiple Tor IPs at the same time and, unless you abuse it, scrape data faster.

The easiest way to start is to duplicate examples/examplescraper and then update, rename, expand, etc. your scraper and related classes.

Your scraper must define config.py and main_dockerized.py. These two names are hardcoded throughout the codebase.

scrapemeagain-compose.py dynamically creates a docker-compose.yml which orchestrates scraper instances. The idea is that the first scraper (scp1) is a master scraper and hence is the host for a couple of helper services which communicate over HTTP (see dockerized/apps).

1. Get Docker host Ip

ip addr show docker0

NOTE Your Docker interface name may be different from docker0.

2. Run examplesite on Docker host IP

python3 examples/examplescraper/examplesite/app.py 172.17.0.1

NOTE Your Docker host IP may be different from 172.17.0.1.

3. Start docker-compose

scrapemeagain-compose.py -s $(pwd)/examples/examplescraper -c tests.integration.fake_config | docker-compose -f - up

NOTE A special config file path is provided: -c tests.integration.fake_config. This is required only for test/demo purposes. You don't have to provide specific config for a real/production scraper.

Local

1. Run examplesite

python3 examples/examplescraper/examplesite/app.py

2. Start examplescraper

python3 examples/examplescraper/main.py

NOTE You may need to update your PYTHONPATH, e.g. export PYTHONPATH=$PYTHONPATH:$(pwd)/examples.

Development

To simplify running integration tests with latest changes:

• replace image: dusanmadar/scrapemeagain:x.y.z with image: scp:latest in the scrapemeagain/dockerized/docker-compose.yml template

• and make sure to rebuild the image locally before running tests, e.g.

docker build . -t scp:latest; python -m unittest discover -p test_integration.py

Legacy

The Python 2.7 version of ScrapeMeAgain, which also provides geocoding capabilities, is available under the legacy branch and is no longer maintained.