Skip to content

Latest commit

 

History

History
68 lines (57 loc) · 3.09 KB

README.md

File metadata and controls

68 lines (57 loc) · 3.09 KB
Raw

Readable-Web Proxy

Reading long-form content on the internet is a shitty experience.
This is a web-proxy that tries to make it better.

This is a rewriting proxy. In other words, it proxies arbitrary web content, while allowing the rewriting of the remote content as driven by a set of rule-files. The goal is to effectively allow the complete customization of any existing web-sites as driven by predefined rules.

Functionally, it's used for extracting just the actual content body of a site and reproducing it in a clean layout. It also modifies all links on the page to point to internal addresses, so following a link points to the proxied version of the file, rather then the original.

While the above was the original scope, the project has mutated heavily. At this point, it has a complete web spider and archives entire websites to local storage. Additionally, multiple versions of each page are kept, with a overall rolling refresh of the entire database at configurable intervals (configurable on a per-domain, or global basis).

There are also a lot of facilities responsible for feeding the releases/RSS views as part of wlnupdates.com.

Quick installation overview:

  • Install Redis
  • (optional) install InfluxDB
  • (optional) install Graphite
  • Install Postgresql >= 10.
  • Build the community extensions for Postgresql.
  • Create a database for the project.
  • In the project database, install the pg_trgm and citext extensions from the community extensions modules.
  • Copy settings.example.py to settings.py.
  • Fill in all settings in settings.py
  • Setup virtualhost by running build-venv.sh
  • Activate vhost: source flask/bin/activate
  • Bootstrap DB: alembic uprade head
  • (on another machine/session) Run local fetch RPC server run_local.sh from https://github.com/fake-name/AutoTriever
  • Run server: python3 run.py
  • If you want to run the spider, it has a LOT more complicated components:
    • Main scraper is started by python runScrape.py
    • Raw scraper is started by python runScrape.py raw
    • Scraper periodic scheduler is started by python runScrape.py scheduler
    • The scraper requires substantial RPC infrastructure. You will need:
      • A RabbitMQ instance with a public DNS address
      • A machine running saltstack + salt-master with a public DNS address On the salt machine, run https://github.com/fake-name/AutoTriever/tree/master/marshaller/salt_scheduler.py
      • A variable number of RPC workers to execute fetch tasks. The AutoTriever project can be used to manage these.
      • A machine to run the RPC local demultiplexing agent (run_agent.sh) The RPC agent allows multiple projects to use the RPC system simultaneously. Since the RPC system basically allows executing either predefined jobs, or arbitrary code on the worker swarm. This is fairly useful in general, so I've implemented it as a service that multiple of my projects then use.

Ubuntu dependencies

  • postgresql-common libpq-dev libenchant-dev
  • probably more I've forgotten