Ink 1.0 is here!

Jul 17, 2017

The Collaborative Knowledge Foundation is pleased to announce the release of INK 1.0.

It’s an exciting milestone for us: After almost a year of development, INK has grown from a scribble on a bit of paper to a fully fledged, production-ready framework ready to automate some of the tricky and time-consuming aspects of document and content processing in publishing.

INK 1.0 allows users to assemble reusable custom-built document pipelines. It’s an extensible, modular, customizable open-source framework.

By facilitating the building of custom automated processes, INK enables smaller publishers and authors to leverage functionality generally only available to specialist vendors or larger publishers.

The next phase of INK development is where we really show off what it can do — file conversion is really only the beginning of highlighting INK’s capabilities. In the future, INK create some steps to directly address some of the challenges seen by our partners and collaborators in the publishing industry to directly automate some of their most time-consuming production processes. You can check out INK 1.0 here.

For the technically minded readers, here are some details on the release.

The full feature set of INK 1.0

These were all added since the last update:

  • INK automatically picks up step classes and makes them available to users of the system to use in recipes
  • A JSON file manifest is included with output files, which includes a checksum and semantically tagged as “new”, “modified”, or “existing”
  • A user can create a recipe with recipe steps with execution parameters as part of the step
  • A user can execute a recipe with one-off execution parameters (these parameters will take priority over parameters with the same name specified by the recipe step)

Recipe step parameters are a really exciting feature. They allow the same step to be customized many different ways depending on the needs of the user without having to write many step classes that all do something similar. For example, a user can specify options when doing a Pandoc file conversion. That Pandoc step could therefore be re-used for many different uses, rather than requireing a different step class for every possible permutation of Pandoc options (and are a lot of them!).

  • A user can create recipes with any steps they like, in any order.
  • A recipe can be public or private – private recipes can only be viewed and executed by the account who created it.
  • A user can change a recipe
  • A user can execute a recipe with any number and size of files (limited only by system resources and the web server settings)
  • A user can see a recipe being executed step by step (the client they use would subscribe to events)
  • A user can download the logs from any step of an execution
  • A Service (associated with an account) can perform executions

An INK administrator can:

  • See the status of essential INK dependencies (e.g. Sidekiq) to quickly determine something that isn’t working or “up” that INK needs to execute recipes
  • See the list of accounts on that INK instance
  • See which step classes are available to execute on that INK instance (sorted by the step gem they appear in)
  • See which step gems are outdated (have an available version upgrade)
  • Monitor the asynchronous processes and check for failed processes
  • Add any combination of step gems to their INK instance

For everyone who’s been following INK since it was merely a scribble, thanks for your continued support and enthusiasm for the project. We’ve been able to create a flexible, extensible, customizable, modular framework for processing documents. INK is useful on its own as well as a workhorse for projects of all types.

Getting started with INK is now made easier thanks to the Docker image provided by Vasilis Kefallinos.