Skip to content

Latest commit

 

History

History
26 lines (17 loc) · 2.8 KB

CONTRIBUTING.md

File metadata and controls

26 lines (17 loc) · 2.8 KB
Raw

Contributing

Hi there! We welcome contributions to Libraries to improve Sourcerer and help it learn more about the engineering world.

Adding a library

We try only to add libraries once they have some usage. We prefer libraries to be in use in hundreds of repositories before supporting them in Sourcerer. You can also refer to coverage chart to see what languages need more support.

To add support for a new library:

Check out libs directory of this repository to find out whether your favorite language is supported. The folder contains json files, one for each language, for example, cpp.json defines libraries for C++. If the folder doesn't contain a file for your language, then please add one. Then, for each library you will need to provide the following information:

  • Library name: as used to commonly refer to a library, e.g. "TensorFlow" or "Laravel Lumen".
  • Library id: this should be a unique ID prefixed with a language code, and generally repeating the library name, e.g. "cpp.tensor-flow" or "php.laravel-lumen".
  • Library technology group (tech): look at technologies.json for existing categories ("iot", "faster-code", etc). If none is fitting feel free to add yours. Technology groups are used to group libraries together on profiles.
  • A compact and descriptive tag for a library: see technologies.json again for examples of tags. Tags are more or less free form, and are used next to library name in Sourcerer to give reader an idea of what the library is about.
  • Library repository URL (repo): If it's GitHub, you can just use a relative path. If no open source code for the library is available then keep repo empty and add examples field with list of github repositories that extensively use the library.
  • Library imports: a list of common prefixes for files you import when using a library. If the library is used without imports, leave the list empty. For example, when importing OpenCV in python, you can use import cv2 or import cv, so you would list ["cv2", "cv"] in imports.
  • A path to examples of library use (optional): This is optional, but if present it will make it much faster for library to appear in Sourcerer. We use examples to train our classifiers for library detection.
  • status of library in Sourcerer. Fill it with awaiting-model. When the library is in production, we will change it to model-ready.

Run node verify.js to validate and sort libs, and also to update technologies.json file.

Open a pull request. Note, it will take us a couple of days after the pull requrest is merged to update classifiers for your libraries.

Every library you list will make the product better for the entire community. Thank you for doing this!