Unit Testing

From WormBaseWiki
Revision as of 19:42, 19 July 2013 by Joachim.baran (talk | contribs) (Added explanation on how to create the GBrowse reference image set.)
Jump to navigationJump to search

Executing Unit Tests

  1. change directory into your Github Website clone.
  2. run tests of the REST API: perl t/rest.t
  3. run tests of the WormBase Perl API: API_TESTS=1 perl t/api.t
  4. run comparative GBrowse tests: perl --base http://your.url.org:1234/cgi-bin/gb2/gbrowse

Successful tests are preceded by an "ok". For example:

ok 1 - require 't/rest_tests/template.t';
    ok 1 - port open
    1..1
ok 2 - test_port_open

How to read the output: The first "ok 1" means that the first executed test require ... completed sucessfully. "ok 2" shows that the second test was also successful, where sub-tests (i.e., tests within the test_port_open Perl sub) are shown above as indented lines. The interval "1..1" denotes the tests were run within test_port_open.

Failed tests are preceded by a "not ok" and a short description is displayed that explains why the test failed (lines preceded by "#"). For example:

    not ok 1 - port open
    #   Failed test 'port open'
    #   at t/rest_tests/template.t line 53.
    #          got: undef
    #     expected: anything else
    1..1
    # Looks like you failed 1 test of 1.
not ok 2 - test_port_open

API Unit Testing Framework: How it works / Write your own

Unit tests are

  • based on Test::More.
  • run on a fully populated WormBase database backend.
  • API tests: start (and stop) a Catalyst web server (random port between 28,000 and 31,999)

Unit test templates are available as

  • t/rest_tests/template.t for writing REST API tests
  • t/api_tests/template.t for writing WormBase Perl API tests

New unit tests should be using the following workflow:

  1. create a copy of template.t within the same directory; choose a descriptive filename
  2. make appropriate changes to the newly created file

Comparative Testing Framework: GBrowse

Comparative GBrowse tests are

  • based on a reference image set (one image per species, per track, per example landmark)
  • comparing a live gbrowse_img rendering to the respective reference image

The reference image set can be created via