Introduction

Saffron | News & announcements aggregation framework

Simple Abstract Framework For the Retrieval Of News

The project is maintained by po/iw (ποιώ), a Greek university team that supports Free/Libre and Open Source Software and hardware. Saffron was built to help teams collect and compile huge lists of aggregated news and announcement content intuitively and efficiently.

What is this?

Saffron is an abstraction engine that helps you collect news and announcements from websites in a uniform way. All you have to do is write a file for each website you want to be scraped and run it in your NodeJS app. Saffron manages the scraping interval, data parsing and - if you want to - the data offloading to your database of choice.

Architecture

Saffron's architecture is based on a main node that issues scraping instructions and several worker nodes that do the scraping & upload the data to the database.

The communication between the nodes is happening through the Grid. The grid will generate events to communicate with other classes. Saffron supports remote nodes by using socket.io server and clients as a middleware to connect to the main node.

Versions

3.0.1

• Fixes
  • Issue new workers for imported jobs.

3.0.0

• New features
  • Type support for configuration.
  • Changed embedded database to user-end implementation.
  • Finished Grid.
  • More than one category per url.
  • Added debug tools.
  • Allow scheduler to continue from previous session.
  • Grid authorization using tokens.
  • Saffron will abort the job if a middleware is rejected with an error.
  • One extension can be used multiple times.
  • Support for re-reading the source from file system without interrupting the scheduler.
• Changes
  • Code remodelling.
  • Removed hash field from article.
  • Merged serialization functions.
  • Grid supports HTTPS instead of in-app encryption.
  • Removed redundant code.
  • SourceException error messages.
  • Saffron is remodeled in a class.
  • Listeners workers.job.finished, workers.job.failed and workers.job.failed emit the job's id and not the job class anymore.
  • Removed listeners database.connection.okay, database.connection.failed and workers.articles.errorOffloading.
  • Added listener database.set.okay, database.get.error, database.set.error and middleware.error.
  • Job handling is done by the Scheduler (previously was Grid).
  • Unified parsers' arguments in the util class.
  • Dynamic parser will receive both initialized utils and uninitialized Article class.
  • Axios was abstracted inside the utils class.
  • parse function will return an array with articles per url instead of a merged array.
• Fixes
  • Listener middleware.after not called on database none.
  • Grid will not emit to remote clients unnecessary events.

2.2.x

• Changes
  • Remodeled WordPress parser.
• Fixes
  • WordPress parser did not put url-category at the categories list.

2.1.x

• New features
  • Allow custom encoding using TextDecoder at axios response.
• Fixes
  • Dynamic parser missing article id.
  • HTML parser fixed attachment list pushed as one attachment.

2.1.0

• New features
  • Added event listeners middleware.before & middleware.after.
  • Added log levels info & errors.

2.0.0

• Changes
  • Reformatted configuration fields.
  • Reformatted source fields.
• New features
  • Added more event listeners
  • Logs can be switched off.

1.x.x

Version 1.x.x was created for test purposes, and it is quite unstable. It is recommended that you're using any version above 2.0.0.