User Interface Reference

Test farm

Click on the device name at the top left corner of the screen to access a menu where you can quickly switch between your devices:

_images/devices-dropdown.png

By default this displays the ID of each Stb-tester node (which based on its MAC address). You can configure a “friendly name” to display: See node.friendly_name in the Configuration Reference.

From this menu click “monitor” to access a “video wall” that allows you to monitor all the devices in your test farm:

_images/video-wall.png

Live video formats

The Stb-tester Portal shows live video from the device-under-test, right in your web browser. Click the settings icon underneath the video to choose the format used for streaming the video:

_images/video-format-selector.png

WebRTC video is high quality and low latency, but it may not work if you are on a different network than the Stb-tester Node. JPEG thumbnails are low quality and low framerate, but they will work in any network configuration.

WebRTC

JPEG thumbnails

High quality

Low quality

Video & audio

No audio

Low latency

1-2 seconds latency

Full frame-rate

1 frame per second

Delivered peer-to-peer (may not work if there is a firewall between you and the Stb-tester Node)

Delivered via the Stb-tester Portal (so it works in any network configuration)

40-220 kB/s outbound bandwidth per viewer

20-35 kB/s (total) if 1 or more viewers

Tags

When running tests, you can specify a set of tags (name and value, or just name). This allows you to change the test-script behaviour at run-time. For example, if you have a test that searches for a programme in your video-on-demand search screen, you might want to specify the name of the programme in a tag.

Tags are passed to the test script as an automatic configuration key and are recorded in the test-results. Tags set here can be retrived in Python during the test-run with stbt.get_config("result.tags", "tagname"). Tags can be retrieved after the test run with the REST API endpoint GET /api/v2/results/(result_id)/ and viewed in the result details pane of the Results View.

You can search for test-results (in the Portal or with the REST API) that match specific tags; see Filter.

Valid tag names:

  • Must not be blank (you can pass a tag without a value, but not a value without a tag).

  • Must not be the same as another tag name.

  • Must be shorter than 500 characters.

  • Must not contain :, =, $, <, >, newline, null, or other unprintable characters.

  • Must not begin with /, -, ., or _.

Valid tag values:

  • Must be shorter than 500 characters.

Job category

_images/run-tests.png

This is simply a way of categorising your test results into different “folders”. When viewing test results, you can filter by this category so that you only see the relevant results.

For example, you might use the system-under-test’s version as the job category. Or if you are testing changes you’ve made to the test scripts, you might use your own name or the name of the git branch, so that your results don’t pollute the “real” results.

Soak-testing options

_images/Soak-testing_options.png

Run once

Each test selected in the Testcases list will run once in the order in which they are selected in the list. Once all the testcases have been run the test job will finish.

Run forever (until stopped manually)

The testcases selected in the Testcases list will be run in the order in which they are selected in the list. Once all the testcases have been run the tests will start again from the beginning of the testcase list.

The testcases will continue to be run in a loop until the Stop button is clicked, or the testcase calls stbt.stop_job.

Run forever in random order (until stopped manually)

This option runs the given testcases in a random order, until the Stop button is clicked.

This can be useful if you have structured your test-pack as a large number of short targeted tests. You can then select many different tests to attempt a random walk of different journeys though your set-top-box UI. This can be particularly effective at finding hard to trigger bugs and get more value out of the testcases you have written.

Some tests may take much longer than other tests, which will then use up a disproportionate amount of your soaking time. To work around this we measure how long each test takes the first time it runs, and then we use that as a weighting when choosing the next test to run. Overall, we try to spend the same amount of time in each testcase, so we run the shorter testcases more frequently.

This makes it reasonable to include tests that take 10 seconds, and tests that take a few minutes or hours, in the same random soak.

These options correspond to the soak and shuffle parameters of the run_tests REST API.

Test results

_images/results-view.png

Filter

The search box at the top of the page allows you to filter the results that are shown.

  • Usually you can just type the words that you want to match. This will show results that contain all of the words you typed (in any order). It will search in the testcase name, the result / failure reason, and the job category (a label that you specify when you run a test job). All searches are case insensitive.

  • Use quotes to search for a specific phrase with spaces: “app crashed” or ‘app crashed’

  • Limit the search to specific fields: testcase:settings result:fail will show failing testruns from any testcases with the word “settings” in the testcase name. Note that these field names match the column headings of the results table.

    • branch: matches the git branch of the test-pack.

      • Note: For test jobs initiated via the REST API, the git branch is only recorded if it was specified by name, not by sha.

    • category: matches the job category (a label that you specify when you run a test job).

    • date: matches the start timestamp of a testrun.

      • date:>”2015-10-21 13:27:00” shows testruns that started after the specified time.

      • date:>2015-10-21 – if you leave out the time, it defaults to midnight (the start of the day). Similarly, you can leave out the seconds, or the minutes.

      • date:<”2015-10-21 13:27:00” shows testruns that started before the specified time.

      • date:[2015-10-21 13:00 to 2015-10-21 06:30] shows testruns that started within the specified range – from 13:00:00 to 06:30:59.

      • date:>now-1d shows testruns that started within the last day (the current time minus 24 hours). The following units are understood:

        • seconds, second, sec, s

        • minutes, minute, min, m

        • hours, hour, hr, h

        • days, day, d

        • weeks, week, w

        • months, month, M

        The time units can be combined, for example “2hr30min”.

    • duration: matches how long a testrun took.

      • This duration can be specified with units, for example duration:>2hr30min. If no units are specified it is assumed to be in seconds. See the date syntax above for the supported units.

      • You can use ranges with square brackets, similar to the date syntax above.

    • job: matches the specified job, for example job:/stb-tester-eca86bfd3b76/1jkw/4. You can find the job uid in the result details pane on the right, or via the REST API.

    • node: matches tests than ran on a specific stb-tester node, for example node:stb-tester-eca86bfd3b76.

    • result: matches the result or failure reason.

      • result:pass shows passing results.

      • result:fail shows failing results.

      • result:error shows errored results (for example, the test did not complete because of a bug in the test script).

      • result:”some failure reason” shows failing or errored results with the specified failure reason (that is, the text of the exception raised by your test script).

    • sha: matches the git sha of the test-pack.

    • tag: matches a tag name, and optionally a value. For example tag:sanity matches any result that had the tag “sanity” specified when the test was run, and tag:model=T-1000 matches the tag “model” only if it has the value “T-1000”. See Tags.

    • testcase: matches the testcase name.

    • user: matches the username of the person who started the test job.

  • Combine search terms with boolean operators (and, or, not). A space is the same as and (unless inside quotes).

    • testcase:settings and result:pass shows results from testcases with the word “settings” in their name, that passed.

    • testcase:settings result:pass is the same as the previous example.

    • testcase:settings or result:pass shows all results that passed, as well as all results for testcases with the word “settings” in their name.

    • (testcase:settings or testcase:menu) result:pass – use parentheses for precedence in complex boolean queries. Without parentheses, and has higher precedence than or.

Save screenshot to Object Repository

_images/save-screenshot-to-object-repository.png
  • App: The app that this screenshot is from (for example “youtube”). If you are testing the operating system itself, use the platform’s name (for example “androidtv”).

  • Page: Name of the Python class that will represent this page of the app. Use CamelCase capitalization. For example “Home”, “Settings”, or “SearchResults”.

  • Device type: Optional platform name if you need to save examples of the same page across different platforms. For example “androidtv” or “roku”. See also device_type configuration.

  • Save to: This previews the filename that the screenshot will be saved as. It’s determined by what you enter in App, Page, Device type and Filename. Learn more about the test-pack file layout here.

  • Branch: The screenshot is saved to your test-pack by creating a new commit in your test-pack git repository. This controls which branch the commit will be created on. See also Development Workflow.

What is the difference between “app” and “device type”?

App is the name of the app you are testing. Device type is the device the app is running on. A single app may run on multiple device types and a single device type may run multiple different apps. A specific device under test has a single unchanging device type.

Examples:

  • If you are testing the YouTube app on Roku 4K device the app is “youtube” and the device type is “roku_4k”

  • If you are testing the same YouTube app on Amazon Fire Stick the app is “youtube” and the device_type is “fire_stick”

  • If you are testing the Sony UI on an Sony TV (i.e. not in an app) then the app is “sony” and the device type would be “sony”.

App and device type are always lower case with no spaces, and may not begin with a number.

Learn more about apps and device types in How to structure your test scripts for re-use across multiple platforms

Learn how to configure what device a given node is attached to in Node-specific configuration files.

REST API

The Stb-tester portal provides a REST API for accessing test results programmatically.