Skip to content
Dynamic definitions and types provider for ruby static analysis
Kotlin Java Ruby C
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
arg_scanner Update plugin version to 0.3.3 Dec 18, 2019
contract-creator Automatically create DB schema on connect Dec 11, 2019
gradle/wrapper Code refactoring and other doesn't matter stuff Dec 13, 2018
ide-plugin Update plugin version to 0.3.3 Dec 18, 2019
screenshots RTI-CR-49 Update Feb 27, 2019
signature-viewer Automatically create DB schema on connect Dec 11, 2019
state-tracker Update Kotlin version Nov 1, 2018
storage-server-api Create indecies for speedup H2 database Dec 12, 2019
.gitignore Remove .idea folder and add it to .gitignore Dec 6, 2018
.travis.yml allow ruby-head to fail for now Nov 9, 2017 RTI-CR-49 Update Feb 27, 2019
LICENSE add license Sep 11, 2017
build.gradle Update gradle and org.jetbrains.intellij Dec 6, 2018 Update plugin dependencies Dec 4, 2019
gradlew Update gradle and org.jetbrains.intellij Dec 6, 2018
settings.gradle Add common module Dec 6, 2019

Automated Type Contracts Generation JetBrains incubator project Build Status

ruby-type-inference project is a completely new approach to tackle the problems of Ruby dynamic nature and provide more reliable symbol resolution and type inference. It collects some run time data to build type contracts for the methods.

Every time a method is being called, some arguments of particular types are being passed to it. Type Tracker collects all such argument combinations and then builds a special contract which satisfies all encountered argument type tuples.

The approach has its own pros and cons:

  • The obtained contracts utilize real-world usages of code of any complexity so it provides true results even if a method utilizes dynamic Ruby features heavily.
  • The completeness of the contracts obtained for a method highly depends on the coverage of that method, including its callees. That implies the need to merge the data obtained from the different sources (e.g. different projects using the same gem).

This implementation addresses the stated coverage problem by providing the possibility to merge any type contracts at any time.


For simple usage you need to install the Ruby Dynamic Code Insight plugin for RubyMine. Then this plugin will require the arg_scanner gem to be installed. See arg_scanner installation instruction if you have problems while installation.

After that, you will have the possibility to run your programs under type tracker:

Run with type tracker

Or you can run your programs in terminal via the rubymine-type-tracker binary (But you have to keep your project opened in RubyMine). E.g.:

rubymine-type-tracker bin/rails server

The rubymine-type-tracker binary is included into the arg_scanner gem.

See for understanding what benefits you will have after running your program under type tracker.


  • arg_scanner is a gem with a native extension to attach to ruby processes and trace and intercept all method calls to log type-wise data flow in runtime.

    See arg_scanner documentation for details on usage.

  • The type contract processor server listens for incoming type data (from arg_scanner) and processes it to a compact format.

    The data stored may be used later for better code analysis and also can be shared with other users.

  • Code analysis clients (a RubyMine/IJ+Ruby plugin) use the contract data to provide features for the users such as code completion, better resolution, etc.

  • (todo) Signature server receives contracts anonymously from the users and provides a compiled contract collections for popular gems.

Running project from sources


The arg_scanner gem is used for collecting type information. It can be installed manually to the target SDK and requires MRI Ruby at least 2.3.

Running type tracker

There are two possibilities to use the type tracker: (I) using IJ/RubyMine plugin or (II) requiring it from Ruby code.

Using RubyMine plugin

The easiest way to run the plugin (and the most convenient for its development) is running it with special gradle task against IJ Ultimate snapshot:

./gradlew ide-plugin:runIde

The task will compile the plugin, run IJ Ultimate with plugin "installed" in it. There is no need in running anything manually in that case.

If you want to try it with existing RubyMine instance, you should:

  1. Build it via ./gradlew ide-plugin:buildPlugin
  2. Install plugin in the IDE
    • Navigate to File | Settings | Plugins | Install plugin from disk...
    • Locate plugin in ide-plugin/build/distributions and select.
    • Restart IDE.

Note that due to API changes the plugin may be incompatible with older RM instances.

Using command line
  1. In order to collect the data for the script needs a contract server to be up and running; it could be run by running

    ./gradlew contract-creator:runServer --args

    where is path where type contracts will be stored (H2 database file).

  2. Run the ruby script to be processed via arg-scanner binary.

  3. Use the data collected by the contract server.


Any kind of ideas, use cases, contributions and questions are very welcome as the project is just incubating. Please feel free to create issues for any sensible request.

You can’t perform that action at this time.