Skip to content

Latest commit

 

History

History
670 lines (497 loc) · 21.9 KB

README.md

File metadata and controls

670 lines (497 loc) · 21.9 KB
Raw

wpt.fyi API

This package defines and implements HTTP API endpoints for wpt.fyi, and this document covers usage and parameters of those endpoints.

Resource endpoints

Here's a list of endpoints to query various resources. An exhaustive list of the endpoints can be found in routes.go.

Also see results creation for endpoints to add new data.

TestRun entities

TestRun entities represent metadata about an execution of the wpt test suite, on a particular product. Tests are run on a regular basis, and each entry in /api/runs annotates when the tests were executed, which product they were executed on, and the where the results are stored.

/api/runs

Gets the TestRun metadata for all runs for a given SHA[0:10], sorted by time_start descending.

Parameters

sha : SHA[0:10] of the runs to get, or the keyword latest. Defaults to latest.

product : Product(s) to include (repeated param), e.g. chrome or firefox-60.

aligned : boolean for whether to get only SHAs which were executed across all of the requested products.

labels: A comma-separated list of labels, e.g. firefox,stable; only runs with all the given labels will be returned. There are currently two kinds of labels supported, browser names (chrome, edge, firefox, safari) and release channels (experimental or stable).

from : RFC3339 timestamp, for which to include runs that occured after the given time. NOTE: Runs are sorted by time_start descending, so be wary when combining this parameter with the max-count parameter below.

to : RFC3339 timestamp, for which to include runs that occured before the given time.

max-count : Maximum number of runs to get (for each browser). Maximum of 500.

staging.wpt.fyi only (Beta params)

pr (Beta): GitHub PR number. Shows runs for commits that belong to the PR.

Examples

Example JSON 
[
  {
    "browser_name": "chrome",
    "browser_version": "67.0.3396.62",
    "os_name": "linux",
    "os_version": "4.4",
    "revision": "2bd11b91d4",
    "full_revision_hash": "2bd11b91d490ddd5237bcb6d8149a7f25faaa101",
    "results_url": "https://storage.googleapis.com/wptd/2bd11b91d4/chrome-stable-linux-summary_v2.json.gz",
    "created_at": "2018-06-05T08:27:30.627865Z",
    "raw_results_url": "https://storage.googleapis.com/wptd-results/2bd11b91d490ddd5237bcb6d8149a7f25faaa101/chrome_67.0.3396.62_linux_4.4/report.json"
  }
]

/api/runs/{id}

Gets a specific (single) TestRun metadata by its datastore ID.

Example

https://wpt.fyi/api/runs/5184362994728960

Example JSON 
{
  "id": "5164888561287168",
  "browser_name": "chrome",
  "browser_version": "67.0.3396.62",
  "os_name": "linux",
  "os_version": "4.4",
  "revision": "2bd11b91d4",
  "full_revision_hash": "2bd11b91d490ddd5237bcb6d8149a7f25faaa101",
  "results_url": "https://storage.googleapis.com/wptd/2bd11b91d4/chrome-stable-linux-summary_v2.json.gz",
  "created_at": "2018-06-05T08:27:30.627865Z",
  "raw_results_url": "https://storage.googleapis.com/wptd-results/2bd11b91d490ddd5237bcb6d8149a7f25faaa101/chrome_67.0.3396.62_linux_4.4/report.json"
}

/api/run

Gets a specific (single) TestRun metadata by product and sha.

Parameters

sha : SHA[0:10] of the runs to get, or the keyword latest. Defaults to latest.

product : browser[version[os[version]]]. e.g. chrome-63.0-linux

Example

https://wpt.fyi/api/run?sha=latest&product=chrome

Example JSON 
{
  "id": "5164888561287168",
  "browser_name": "chrome",
  "browser_version": "67.0.3396.62",
  "os_name": "linux",
  "os_version": "4.4",
  "revision": "2bd11b91d4",
  "full_revision_hash": "2bd11b91d490ddd5237bcb6d8149a7f25faaa101",
  "results_url": "https://storage.googleapis.com/wptd/2bd11b91d4/chrome-stable-linux-summary_v2.json.gz",
  "created_at": "2018-06-05T08:27:30.627865Z",
  "raw_results_url": "https://storage.googleapis.com/wptd-results/2bd11b91d490ddd5237bcb6d8149a7f25faaa101/chrome_67.0.3396.62_linux_4.4/report.json"
}

/api/shas

Gets an array of revisions (SHA[0:10]), in reverse chronological order. This method behaves similarly to /api/runs above, but projects the revision field's value.

Parameters

aligned : boolean for whether to get only SHAs which were executed across all of the requested products.

product : Product(s) to include (repeated param), e.g. chrome or firefox-60

from : RFC3339 timestamp, for which to include runs that occured after the given time. NOTE: Runs are sorted by time_start descending, so be wary when combining this parameter with the max-count parameter below.

to : RFC3339 timestamp, for which to include runs that occured before the given time.

max-count : Maximum number of runs to get (for each browser). Maximum of 500.

Example

https://wpt.fyi/api/shas?product=chrome

Example JSON 
[
  "98530fb944",
  "2bd11b91d4"//, ...
]

Results summaries

The following methods apply to the results summaries JSON blobs, which are linked to from TestRun entities.

/api/results

Performs an HTTP redirect for the results summary JSON blob of the given TestRun.

Response format

The summary JSON is in the format

{
  "/path/to/test.html": [1, 1],
}

Where the array contains [number of passes, total tests].

Parameters

product : Product to fetch the results for, e.g. chrome-66

sha : SHA[0:10] of the TestRun to fetch, or the keyword latest. Defaults to latest.

Example

https://wpt.fyi/api/results?product=chrome

Example JSON (from the summary_v2.json.gz output) 
{
  "/css/css-text/i18n/css3-text-line-break-opclns-213.html": [1, 1],
  "/css/css-writing-modes/table-progression-vrl-001.html": [1, 1],
  // ...
}

/api/diff

Computes a summary JSON blob of the differences between two TestRun summary blobs, in the format of an array of [improved, regressed, total-delta].

Parameters

run_ids : Exactly two numerical IDs for the "before" and "after" runs (in that order), separted by a comma. IDs associated with runs can be obtained by querying the /api/runs API. This overrides the before and after params.

before : [product]@[sha] spec for the TestRun to use as the before state.

after : [product]@[sha] spec for the TestRun to use as the after state.

path : Test path to filter by. path is a repeatable query parameter.

filter : Differences to include in the summary.

  • A : Added - tests which are present after, but not before.
  • D : Deleted - tests which are present before, but not after.
  • C : Changed - tests which are present before and after, but the results summary is different.
  • U : Unchanged - tests which are present before and after, and the results summary count is not different.

Test Manifest

The following methods apply to the retrieval and filtering of the Test Manifest in WPT, which contains metadata about each test.

/api/manifest

Gets the JSON of the WPT manifest GitHub release asset, for a given sha (defaults to latest).

Parameters

sha : SHA of the WPT repo PR for which to fetch, the manifest, or the keyword latest. (Defaults to latest.)

NOTE: The full SHA of the fetched manifest is returned in the HTTP response header X-WPT-SHA, e.g.

X-WPT-SHA: abcdef0123456789abcdef0123456789abcdef01

Response format

The high-level structure of the v4 manifest is as follows:

{
  "items": {
    "manual": {
        "file/path": [
          manifest_item,
          ...
        ],
        ...
    },
    "reftest": {...},
    "testharness": {...},
    "visual", {...},
    "wdspec": {...},
  },
}

manifest_item is an array (nested in the map's "file/path" value's array) with varying contents. Loosely,

  • For testharness entries: [url, extras]
    • extras example: {"timeout": "long", "testdriver": True}
  • For reftest entries: [url, references, extras]
    • references example: [[reference_url1, "=="], [reference_url2, "!="], ...]
    • extras example: {"timeout": "long", "viewport_size": ..., "dpi": ...}

Results creation

/api/results/upload

Uploads a wptreport to the dashboard to create the test run.

This endpoint only accepts POST requests. Requests need to be authenticated via HTTP basic auth. Please file an issue if you want to register as a "test runner", to upload results.

File payload

Content type: multipart/form-data

Parameters

labels: (Optional) A comma-separated string of labels for this test run. Currently recognized labels are "experimental" and "stable" (the release channel of the tested browser).

callback_url: (Optional) A URL that the processor should POST when successful, which will create the TestRun. Defaults to /api/results/create in the current project's environment (e.g. wpt.fyi for wptdashboard, staging.wpt.fyi for wptdashboard-staging).

result_file: A gzipped JSON file, with the filename ending with .gz extension, produced by wpt run --log-wptreport. This field can be repeated to include multiple files (for chunked reports).

screenshot_file: A gzipped screenshot database produced by wpt run --log-screenshot. This field can be repeated to include multiple links (for chunked reports).

The JSON file roughly looks like this:

{
  "results": [...],
  "time_start": MILLISECONDS_SINCE_EPOCH,
  "time_end": MILLISECONDS_SINCE_EPOCH,
  "run_info": {
    "revision": "WPT revision of the test run",
    "product": "your browser",
    "browser_version": "version of the browser",
    "os": "your os",
    "os_version": "OPTIONAL OS version",
    ...
  }
}

Notes

The time_start and time_end fields are numerical timestamps (in milliseconds since the UNIX epoch) when the whole test run starts and finishes. They are optional, but encouraged. wpt run produces them in the report by default.

run_info.{revision,product,browser_version,os} are required, and should be automatically generated by wpt run. If for some reason the report does not contain these fields (e.g. old WPT version, Sauce Labs, or custom runners), they can be overridden with the following optional parameters in the POST payload (this is NOT recommended; please include metadata in the reports whenever possible):

  • revision (note this should be the full revision hash, not a 10-char truncation)
  • browser_name (note that it is not called product here)
  • browser_version
  • os_name (note that it is not called os here)
  • os_version

URL payload

Content type: application/x-www-form-urlencoded

Parameters

result_url: A URL to a gzipped JSON file produced by wpt run --log-wptreport (see above for its format). This field can be repeated to include multiple links (for chunked reports).

screenshot_url: A URL to a gzipped screenshot database produced by wpt run --log-screenshot. This field can be repeated to include multiple links (for chunked reports).

callback_url: (Optional) A URL that the processor should POST when successful, which will create the TestRun. Defaults to /api/results/create in the current project's environment (e.g. wpt.fyi for wptdashboard, staging.wpt.fyi for wptdashboard-staging).

labels: (Optional) A comma-separated string of labels for this test run. Currently recognized labels are "experimental" and "stable" (the release channel of the tested browser).

/api/results/create

This is an internal endpoint used by the results processor.

Querying test results

/api/search

Search for test results over some set of test runs. This endpoint accepts POST and GET requests.

  • POST requests are forwarded to the searchcache for structured queries, with run_ids and query fields in the JSON payload; see search query documentaton for more information.

  • GET requests are unstructured queries with the following parameters:

Parameters

run_ids : (Optional) A comma-separated list of numerical ids associated with the runs over which to search. IDs associated with runs can be obtained by querying the /api/runs API. Defaults to the default runs returned by /api/runs. NOTE: This is not the same set of runs as is shown on wpt.fyi by default.

q: (Optional) A query string for search. Only results data for tests that contain the q value as a substring of the test name will be returned. Defaults to the empty string, which will yield all test results for the selected runs. NOTE: structured search queries are not supported.

Examples

Example JSON 
{
  "runs": [
    {
      "id": 6.311104602964e+15,
      "browser_name": "chrome",
      "browser_version": "68.0.3440.106",
      "os_name": "linux",
      "os_version": "16.04",
      "revision": "2dda7b8c10",
      "full_revision_hash": "2dda7b8c10c7566fa6167a32b09c85d51baf2a85",
      "results_url": "https:\/\/storage.googleapis.com\/wptd-staging\/2dda7b8c10c7566fa6167a32b09c85d51baf2a85\/chrome-68.0.3440.106-linux-16.04-edf200244e-summary_v2.json.gz",
      "created_at": "2018-08-17T08:12:29.219847Z",
      "time_start": "2018-08-17T06:26:52.33Z",
      "time_end": "2018-08-17T07:50:09.155Z",
      "raw_results_url": "https:\/\/storage.googleapis.com\/wptd-results-staging\/2dda7b8c10c7566fa6167a32b09c85d51baf2a85\/chrome-68.0.3440.106-linux-16.04-edf200244e\/report.json",
      "labels": [
        "buildbot",
        "chrome",
        "stable"
      ]
    },
    {
      "id": 5.132783244542e+15,
      "browser_name": "firefox",
      "browser_version": "61.0.2",
      "os_name": "linux",
      "os_version": "16.04",
      "revision": "2dda7b8c10",
      "full_revision_hash": "2dda7b8c10c7566fa6167a32b09c85d51baf2a85",
      "results_url": "https:\/\/storage.googleapis.com\/wptd-staging\/2dda7b8c10c7566fa6167a32b09c85d51baf2a85\/firefox-61.0.2-linux-16.04-75ff911c43-summary_v2.json.gz",
      "created_at": "2018-08-17T08:31:38.580221Z",
      "time_start": "2018-08-17T06:47:29.643Z",
      "time_end": "2018-08-17T08:15:18.612Z",
      "raw_results_url": "https:\/\/storage.googleapis.com\/wptd-results-staging\/2dda7b8c10c7566fa6167a32b09c85d51baf2a85\/firefox-61.0.2-linux-16.04-75ff911c43\/report.json",
      "labels": [
        "buildbot",
        "firefox",
        "stable"
      ]
    }
  ],
  "results": [
    {
      "test": "\/html\/dom\/elements\/global-attributes\/lang-xyzzy.html",
      "legacy_status": [
        {
          "passes": 1,
          "total": 1
        },
        {
          "passes": 1,
          "total": 1
        }
      ]
    }
  ]
}

Metadata results

/api/metadata

API endpoint for fetching all of the link metadata stored in the wpt-metadata repository, with the (normally file-sharded) data all flattened into a JSON object which is keyed by test name.

This endpoint accepts POST and GET requests.

  • GET request returns Metadata Link Information by product, and requires product parameters;

  • POST request searches Metadata Link by link url, and requires product parameters and payload.

URL Parameters

product : browser[version[os[version]]]. e.g. chrome-63.0-linux

JSON Request Payload

[
  {
    "link": "[pattern]"
  }
]

Where [pattern] is any substring of the url field of a wpt-metadata link node.

Get Examples

  • /api/metadata?product=chrome&product=safari
Example JSON 
{
  "/FileAPI/blob/Blob-constructor.html": [
    {
      "url": "https://github.com/web-platform-tests/results-collection/issues/661",
      "product": "chrome",
      "results:" [
        {
          "subtest": "Blob with type \"image/gif;\"",
          "status": "UNKNOWN"
        },
        {
          "subtest": "Invalid contentType (\"text/plain\")",
          "status": "UNKNOWN"
        }
      ]
    }
  ],
  "/service-workers/service-worker/fetch-request-css-base-url.https.html": [
    {
      "url": "https://bugzilla.mozilla.org/show_bug.cgi?id=1201160",
      "product": "firefox",
    }
  ],
  "/service-workers/service-worker/fetch-request-css-images.https.html": [
    {
      "url": "https://bugzilla.mozilla.org/show_bug.cgi?id=1532331",
      "product": "firefox"
    }
  ]
}

Post Examples

  • POST /api/metadata?product=chrome&product=firefox
    exists:='[{"link":"bugs.chromium.org"}]'
Example JSON 
{
  "/IndexedDB/bindings-inject-key.html": [
    {
      "url": "bugs.chromium.org/p/chromium/issues/detail?id=934844"
    }
  ],
  "/html/browsers/history/the-history-interface/007.html": [
    {
      "url": "bugs.chromium.org/p/chromium/issues/detail?id=592874"
    }
  ]
}

/api/metadata/pending

API endpoint for retrieving pending metadata whose PRs are not merged yet. This endpoint is used along with the /api/metadata endpoint to retrieve all metadata, pending or non-pending. It accepts GET requests without any parameters. It returns the same JSON response as /api/metadata.

This endpoint is a best-effort API, because in some rare cases, e.g. both the Redis server and its replica go down, pending metadata information can be lost temporarily.

/api/metadata/triage

This API is available for trusted third parties.

To use the Triage Metadata API, you first need to sign in to wpt.fyi (top-right corner; 'Sign in with GitHub'). For more information on wpt.fyi login, see here.

The logged-in user also needs to belong to the 'web-platform-tests' GitHub organization. To join, please file an issue, including the reason you need access to the Triage Metadata API.

Once logged in, you can send a request to /api/metadata/triage to triage metadata. This endpoint only accepts PATCH requests and appends a triage JSON object to the existing Metadata YML files. The JSON object is a flattened YAML Links structure that is keyed by test name Test path; see below for an example.

This endpoint returns the URL of a PR that is created in the wpt-metadata repo.

Example JSON Body 
{
  "/FileAPI/blob/Blob-constructor.html": [
    {
      "url": "https://github.com/web-platform-tests/results-collection/issues/661",
      "product": "chrome",
      "results:" [
        {
          "subtest": "Blob with type \"image/gif;\"",
          "status": 6
        },
        {
          "subtest": "Invalid contentType (\"text/plain\")",
          "status": 0
        }
      ]
    }
  ],
  "/service-workers/service-worker/fetch-request-css-base-url.https.html": [
    {
      "url": "https://bugzilla.mozilla.org/show_bug.cgi?id=1201160",
      "product": "firefox",
    }
  ],
  "/service-workers/service-worker/fetch-request-css-images.https.html": [
    {
      "url": "https://bugzilla.mozilla.org/show_bug.cgi?id=1532331",
      "product": "firefox"
    }
  ]
}

Browser Specific Failure

/api/bsf

Gets the BSF data of Chrome, Firefox, Safari for the home directory.

The endpoint accepts GET requests.

Parameters

from : (Optional) RFC3339 timestamp, for which to include BSF data that occured after the given time inclusively.

to : (Optional) RFC3339 timestamp, for which to include BSF data that occured before the given time exclusively.

experimental : A boolean to return BSF data for experimental or stable runs. Defaults to false.

JSON Response

The response has three top-level fields:

lastUpdateRevision indicates the latest WPT Revision updated in data.

fields corresponds to the fields (columns) in the data table and has the format of an array of:

  • sha, date, [product-version, product-score]+

data returns BSF data in chronological order.

Example JSON 
{
   "lastUpdateRevision":"eea0b54014e970a2f94f1c35ec6e18ece76beb76",
   "fields":[
      "sha",
      "date",
      "chrome-version",
      "chrome",
      "firefox-version",
      "firefox",
      "safari-version",
      "safari"
   ],
   "data":[
      [
         "eea0b54014e970a2f94f1c35ec6e18ece76beb76",
         "2018-08-07",
         "70.0.3510.0 dev",
         "602.0505256721168",
         "63.0a1",
         "1617.1788882804883",
         "12.1",
         "2900.3438625831423"
      ],
      [
         "203c34855f6871d6e55eaf7b55b50dad563f781f",
         "2018-08-18",
         "70.0.3521.2 dev",
         "605.3869030161061",
         "63.0a1",
         "1521.908686731921",
         "12.1",
         "2966.686195133767"
      ]
   ]
}