This package defines and implements HTTP API endpoints for wpt.fyi, and this document covers usage and parameters of those endpoints.
Here's a list of endpoints to query various resources. An exhaustive list of
the endpoints can be found in routes.go
.
- /api/runs
- /api/runs/{id}
- /api/run
- /api/shas
- /api/diff
- /api/results
- /api/manifest
- /api/search
- /api/metadata
- /api/metadata/pending
- /api/metadata/triage
- /api/bsf
Also see results creation for endpoints to add new data.
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.
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 product
s.
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.
pr
(Beta): GitHub PR number. Shows runs for commits that belong to the PR.
- https://wpt.fyi/api/runs?product=chrome&product=safari
- https://wpt.fyi/api/runs?product=chrome&from=2018-01-01T00:00:00Z&max-count=10
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"
}
]
Gets a specific (single) TestRun metadata by its datastore ID.
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"
}
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
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"
}
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 product
s.
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.
https://wpt.fyi/api/shas?product=chrome
Example JSON
[
"98530fb944",
"2bd11b91d4"//, ...
]
The following methods apply to the results summaries JSON blobs, which are linked to from TestRun entities.
Performs an HTTP redirect for the results summary JSON blob of the given TestRun.
Response format
The summary JSON format has been updated as of July 2022, and all requisite
summary files should now follow this newformat. Summary files with the new format
are denoted with the _v2
file name suffix. This change was made to
differentiate a test's overall status value from the subtest passes and
totals.
The v2 summary JSON is in the format
{
"/path/to/test.html": {
"s": "O",
"c": [1, 1]
},
}
Each test path has two properties.
s
, or status, which is an abbreviated value to the test's overall status.
c
, or counts, which is an array containing
[number of subtest passes
, total subtests
].
Status abbrevations
Status | Abbreviation |
---|---|
OK | O |
PASS | P |
FAIL | F |
SKIP | S |
ERROR | E |
NOTRUN | N |
CRASH | C |
TIMEOUT | T |
PRECONDITION_FAILED | PF |
Any summary files before this update follow the old JSON format (v1). The v1 summary format has no additional name suffix, unlike v2.
The v1 JSON is in the format
{
"/path/to/test.html": [2, 2],
}
Where the array contains [number of subtest passes
, total subtests
].
The test's overall status is added with these subtest values. A passing status
value (OK
or PASS
) will increment the number of subtest passes.
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
.
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],
// ...
}
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.
The following methods apply to the retrieval and filtering of the Test Manifest in WPT, which contains metadata about each test.
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": ...}
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.
Content type: multipart/form-data
Parameters
labels
: (Optional) A comma-separated string of labels for this test run. Currently recognized
labels are "experimental", "stable" (the release channel of the tested browser) and "master" (test run
from the master branch).
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 calledproduct
here)browser_version
os_name
(note that it is not calledos
here)os_version
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).
This is an internal endpoint used by the results processor.
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
andquery
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.
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
}
]
}
]
}
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
[
{
"link": "[pattern]"
}
]
Where [pattern]
is any substring of the url field of a wpt-metadata link
node.
- /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 /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 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.
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"
}
]
}
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"
]
]
}