web-features-mappings

Mapping external data to web-features IDs

The web-features project provides the minimum amount of data that's needed to support Baseline and otherwise acts as a repository of unique feature IDs, which other projects can point to.

This was done for maintainability reasons, to avoid adding a lot of third-party data to the web-features project to support other use cases than Baseline. This means that third-party data sources can map their own data to web-features IDs, and are responsible for maintaining that mapping.

Examples of data sources which map to web-features include:

• The web-platform-tests project, which maps certain tests to web-features via search keywords, e.g. the feature:grid keyword.
• The browser-compat-data project, which maps BCD keys to web-features via tags, e.g. the web-features:selection-api tag.
• Chrome Platform Status' Web features usage metrics, which maps Chrome page loads to web-features.

Adding new mappings

If you maintain data about features of the web platform, consider creating a resource which provides a mapping between your data and IDs from the web-features project and let us know about it.

About the files in this repository

There are two types of mapping files in this repository, under the /mappings/ folder:

• Files for data sources which do not yet map to web-features IDs.
• Files for data sources which already map to web-features IDs.

Data sources which don't yet map to web-features IDs

Currently, not all the data sources which are helpful to web developers and/or browser engineers are mapped to web-features IDs. Because these data sources are used on the Web platform features explorer website but a mapping did not exist, we maintain files in this repository to do the mapping instead.

Examples of data sources which are mapped in this repository:

• Origin trials
• MDN documentation
• Standard positions

Data sources which already map to web-features IDs

Data sources which already maintain a mapping to web-features on their own still have mapping files in this repository. These files are typically updated automatically, on a schedule, for convenience. This way, this repository can be used to retrieve all currently known web-features-mapped data, whether the mapping is maintained here or elsewhere.

Examples of data sources which already map to web-features IDs and for which we automatically update files in this repository:

• Interop focus areas.
• Chrome use counters.
• web-platform-tests.

The scripts folder

The /scripts/ folder contains the JavaScript files that are responsible for updating the mapping files.

To run these scripts:

1. cd scripts
2. npm install
3. node <the-script-you-want-to-run>.js

Or, to run all scripts:

1. cd scripts
2. npm install
3. npm run update:all

Mapping format

The mappings are JSON files that are formatted as follows:

{
  "a web-features id": <data that's specific to this mapping file>
}

Combined data

The combine script generates a combined-data.json file in the mappings directory. This file contains all the mapping data from the mappings folder, combined into a single file.

The format of the combined file is as follows:

{
  "<feature-id>": {
    "chrome-use-counters": { ... },
    "mdn-docs": [ ... ],
    "wpt": { ... }
    ...
  },
  ...
}

Updating the data

The data is updated automatically, once a day, by the GitHub Actions workflow in .github/workflows/update.yml.

To update the data locally:

1. cd scripts
2. npm run bump to make sure you have the latest versions of the dependencies locally.
3. npm run build to update all individual data files, check their format, and re-generate the combined data file.

TODO

• Add mapping to origin trials.
• Add mapping to TAG reviews.
• Add mapping to chromestatus entries (via spec URLs?)
• Try to find a better way than matching on spec URLs.
• Find a way to detect new standards positions automatically.
• Sometimes MDN doc pages get removed (:target-within was removed recently). Find a way to remove the mapping.
• Publish consolidated data to NPM.
• Also create releases on GitHub so consumers can download JSON from there too.
• Migrate the explorer to use this data instead of its own.