There are sometimes that we want to be first person that get notice about new posts in a website. For example when you're looking for a house, you want to be first person that apply to its advertisement and by this way you want to increase your chance in getting it. Besides, sometime you may want to get new articles from your favorite websites on a daily. This framework is designed specifically for these goals.
You can setup the project using docker-compose command.
git clone [email protected]:ghorbani-mohammad/crawler-framework.git
cd crawler-framework
docker-compose up
- For production you can use
docker-compose -f docker-compose-prod.yml up
. - I've used gunicorn to served the requests.
- For serving static files, I've used Nginx. Checkout crawler_api_nginx.conf configuration.
- I've used selenium hub(grid) to provide multiple browser sessions at same time.
- You can configure smtp server credentials and set your email to get error logs in your email inbox
- You can also check all logs (all levels) in Django admin (DBLogEntry model)
- You can use provided shell commands to easily manage the project (checkout mng-api.sh file)
This is a framework for crawling data from websites. You define a website, then the pages of that website and in the last step you define the structure of those pages so the crawler engine can retrieve data from them. In this framework we have three main entities:
- Agency
- Page
- Structure
Agencies are the websites, like CNN and BBC. First step is defining agencies.
Pages, are different pages of an agency or a website. For instance, the CNN website has political, entertainment and etc pages in it. After defining your agencies you can specify pages of that website which you want crawl data from them.
Structures define how crawler engine should gather data from a page. When you defining a page, you should specify its structure. So you need define structure of a page, before creating the page. This model has three important fields that probably you need fill those.
First one is news_links_structure. This field specifies how we should get links of news or articles or anything that we want. At the below picture you can see an example. As you can see, we gather elements with tag a that has class attribute with value c-jobListView__titleLink
Guest user have read-only access to some models, So using that, you can login into Django admin and see what potentials can have this project. By log in into the Django admin, you can see I have defined a bunch of websites that I get their new posts periodically. These are some examples that help you to create your own crawler.
- Guest User Credential:
- Username:
guest
- Password:
RPxzsoen4O
- Username:
- Django Admin Panel:
There are some TG channels that I use this crawler framework to keep them updated. Example: