Tool to support Desktop Test Engineering triage test cases for future automation

If you find an issue with this tool, please create a bug on Mozilla's BugZilla instance, that blocks this bug.

The tool requires a .env file to exist on the root of the directory where you cloned the repository. Please do not name this file anything else, and DO NOT SAVE SECRETS to any other file, as you will likely push those secrets to our upstream, where anyone can see them. The following values need to be set:

• TESTRAIL_BASE_URL=https://mozilla.testrail.io
• TESTRAIL_USERNAME=<valid testrail username>
• TESTRAIL_API_KEY=<valid testrail api key>
• BUGZILLA_BASE_URL=https://bugzilla.mozilla.org
• BUGZILLA_API_KEY=<valid bugzilla api key>

To obtain usernames and API keys for DTE service accounts, please talk to Ben C to be added to the vault or receive a secure link.

The project uses semantic versioning. The current version can be found in the pyproject file. We use bump-my-version to bump the version. To bump the version, run the following command:

uv run bump-my-version bump [major|minor|patch]

If you just want to start the application the fastest way possible, simply run ./quickstart.sh from a compatible terminal's command prompt. (Linux including WSL, Unix including MacOS terminals). If you don't want to do that or it does not work, try the following:

You will need to build the UI before starting up streamlit. Inside src/UI/:

• npm install

If you wish to run the custom component in a separate server, then:

• npm run start

Otherwise, to build the UI component:

• npm run build

After either starting the UI server or building the component, run the streamlit app from the root of the project:

• uv run streamlit run main.py

The UI should be relatively intuitive. In the beta version, we default to TestRail project 17 (Fx Desktop). If you want to use the tool to triage test cases in other projects, change the Project ID to the appropriate number.

Note that the Commit Changes button might be below the break on your browser and you will have to scroll the left hand panel to find it.

When you hit the Commit Changes button, TestRail changes are immediately performed. Bugzilla changes are not, but JSON files containing the relevant info are produced. They start with session_ and contain timestamps. It is expected that occasionally a single session will produce multiple files.

To commit all of these to Bugzilla, there is a script. Take note: only run the script once per suite modified. Don't just produce a bunch of session files for multiple suites and run them all together. Given any number of sessions for the same suite:

uv run python update_bugzilla.py session_2025-09-26T16_17_18Z.json session_2025-09-26T16_19_20Z.json

Where, of course, the session JSON files are replaced by the one(s) that live on your filesystem. The new suite and case bugs will block the functional root bug by default. Change the global variable to assign bugs to a new root metabug.

We use uv to manage our environment, as a way to test-pilot this tool to replace pipenv in other DTE-managed Python projects. See Astral's docs for more information, including how to add, remove, and update dependencies.

To initialize env:

• uv sync
• source .venv/bin/activate on Mac / Linux / WSL
• .venv\Scripts\activate on PowerShell

We use hurl to test our API requests. See documentation at Hurl's page.

To add a test:

• Create the hurl description file at tests/hurl/hurl_files/[name].hurl
• If you need variables defined, add those additional key-value pairs to a file in that same folder, with file extension .vars
• Update /tests/hurl/manifest.yaml like this example:

  get_sample_payload: result: pass variables-file: suite_tests

• If you need to test that something fails, set result to fail
• If you need to skip a test, set result to skip

We use playwright to test our UI. See documentation at the Playwright project page for more.

We use pytest to bring it all together.

• Pytest documentation
• Pytest Playwright

We store secrets securely. Please note the following:

• Do not use the -u flag in hurl when writing test harness logic, this exposes secrets to the public in the GHA readouts in event of a failure.
• We save the entire value of the .env file as a secret in GHA, which contains the keys mentioned in the "Pre-Setup" section.