Pytest HTML Reporter for VS Code

View pytest-html-reporter results in the VS Code sidebar — failures grouped by suite, error previews, one-click navigation to the test, and flaky-test detection drawn from the build history the plugin already keeps.

Features

Failures at a glance. Every failing test grouped under its suite, with a truncated error message on each card.
Jump to the test. Click a failure to open the file at the test's def.
Copy the real error. The card shows readable text; the clipboard gets the original, terminal colouring intact.
Flaky vs broken. Past builds in archive/ are read to tell a test that sometimes fails from one that has never passed. Stable tests get no badge, so the ones that do mean something.
Charts. A pass-rate ring, a proportional status bar, and failures per build across the recent archive — the trend no single report can show.
Per-test trend. The last dozen outcomes for a flaky test, oldest to newest.
Full report, in the editor. Open the generated HTML report — Steps, Attachments, Analytics, Coverage, screenshots — in a tab instead of a browser.
Live refresh. The sidebar reloads when a run finishes.
Several reports. Switch between them from a dropdown.

Getting started

pip install pytest-html-reporter
pytest --html-report=./report

Open the Pytest HTML Reporter view in the activity bar. The report is found automatically; if you keep it somewhere unusual, run Pytest HTML Reporter: Configure Report Path.

How it finds your report

The extension reads output.json, never the generated HTML — the HTML is only ever displayed. It looks for output.json in the workspace root (the plugin's default, since it writes relative to pytest's working directory), then in report/, reports/, test-reports/, .reports/ and test-results/, then falls back to a bounded workspace search.

Every candidate is checked for the plugin's own content.suites shape before it is offered. output.json is a common filename, and rendering an unrelated one as test results would be worse than finding nothing.

Alongside it the extension expects the layout the plugin writes:

<report dir>/output.json                       the current run
<report dir>/archive/output_<timestamp>.json   past runs — the history
<report dir>/pytest_html_report.html           the full report
<report dir>/pytest_screenshots/               screenshots

Settings

Setting	Default	What it does
`pytestHtmlReporter.reportJsonPaths`	`[]`	Explicit `output.json` paths. Empty means auto-detect.
`pytestHtmlReporter.autoRefresh`	`true`	Reload when the report changes on disk.
`pytestHtmlReporter.showErrorSnippets`	`true`	Show a truncated error on each failure.
`pytestHtmlReporter.maxErrorLength`	`150`	Snippet length, 50–500.
`pytestHtmlReporter.history.enabled`	`true`	Read `archive/*.json` for flaky/broken verdicts.
`pytestHtmlReporter.history.maxBuilds`	`25`	Recent builds to consider, 2–200.

Launching the Extension Development Host to check a colour is a slow loop, so the sidebar can be rendered straight to HTML:

npm run preview                # the bundled fixture, with failures
npm run preview -- all-passed  # a green run
npm run preview -- no-config   # nothing found yet
npm run preview -- error       # a corrupt report

It writes a dark and a light file into .preview/; open either in a browser. The real VS Code theme variables are stubbed in, so what you see is what the webview renders. Point it at one of your own runs with PREVIEW_REPORT=/path/to/output.json npm run preview — the history charts then read that report's own archive/ too.

For the full thing, press F5 for an Extension Development Host and open a folder containing a report.

Known limitation: locating a test

output.json records no line numbers, and no nodeid. Each result carries only its suite path and pytest's item.name — which means the class segment of the nodeid is not written down anywhere.

So "jump to test" opens the suite file and scans it for the matching def. That is exact for the overwhelming majority of tests. Where it cannot be exact is when one file defines the same test name twice — two classes with a method of the same name — because the record has nothing to tell them apart. In that case you get a pick list rather than a silent guess at the wrong one.

If a test has been renamed or deleted since the run, the file still opens, at line 1, with a note explaining why.

What it does not do

It does not run your tests — the built-in Python extension's Test Explorer does that well already. This is a viewer for reports the plugin has produced, and it never shells out to pytest.

Steps, attachments, logs, BDD scenarios and per-file coverage are the HTML report's job; the sidebar links out to it rather than reproducing it.

License

MIT

Pytest HTML Reporter

Prashanth Sams