Report for Assignment 1

Project chosen

Name: SimpleParsing

URL: https://github.com/lebrice/SimpleParsing

Number of lines of code and the tool used to count it: lizard, 18161 nloc

Programming language: Python

Coverage measurement

Existing tool

Coverage.py via the command: coverage run -m pytest

Your own coverage tool

Karina Sudnicina

Function 1 name: get_bound(t) from simple_parsing/utils.py

Function 2 name: _description_from_docstring(docstring: dp.Docstring) from simple_parsing/decorators.py

Nora Cai

Function 1 name: def post_init from examples/ugly_example_after.py

Function 2 name: def contains_dataclass_type_arg from simple_parsing/utils.py

Che Lai

Function 1 name: def setattr_recursive from simple_parsing/utils.py

Function 2 name: def getattr_recursive from simple_parsing/utils.py

Sanne Aerts

Function 1: get_item_type from simple_parsing/utils.py

The instrumented function

Coverage results output from the instrumentation

Function 2: get_argparse_type_for_container from simple_parsing/utils.py

The instrumented function

Coverage results output from the instrumentation

Coverage improvement

Individual tests

Karina Sudnicina

Test 1: get_bound(t)

Coverage before:

Coverage after:

As seen from the screenshot, the function was initially not covered completely by test cases. I have added the test cases to make sure all of the brances are covered. As a result, the function is covered by 100% and the number of 'missing' statements in utils.py got decreased by 1 - from 77 to 76.

Test 2: def _description_from_docstring(docstring: dp.Docstring)

Coverage before:
Coverage after:

Similarly, the function was not covered completely. Adding new test cases for this function raised the coverage of decorators.py file for 2% - from 91 to 93. The function is now covered by 100% and the 'missing' statements number has also decreased by 1 - from 6 to 5.

Nora Cai

Test 1: test/test_ugly_example_after.py

added new test:

old coverage:

new coverage:

percentage of improvement: 60%%

The coverage of the function def post_init went from 40% to 100%, this is because the coverage for the function has been improved by writing test case to cover the second and third branch, which were not previously covered.

Test 2: test/test_base.py

enhanced existing test:

old coverage:

new coverage:

percentage of improvement: 14%

The coverage of the function def contains_dataclass_type_arg went from 86% to 100%, the increase is caused by the enhancement of test cases that that covers the second branch of the function, which was not previously covered.

Che Lai

Test 1: setattr_recursive

old coverage results: (coverage: 0%)

new coverage results: (coverage: 100%)

In the original test file, there was no test case for this function, so the coverage of this function was 0%, then I added the test code for this function, then its coverage increased to 100%.

Test 2: getattr_recursive

old coverage results: (coverage: 0%)

new coverage results: (coverage: 100%)

percentage of improvement: 100%

Same as the function in Test 1, there was no test case for this function, so the old covearage was 0%, after I added test case for it, the coverage of this function increased to 100%.

Sanne Aerts

Test 1: New test for Function 1 (get_item_type from simple_parsing/utils.py) to hit the third branch.

Old coverage result

New coverage result

Adding a new test case to cover the third branch of the function resulted in a coverage improvement of 17%, bringing the total coverage for the function to 100%.

Test 2: New test for Function 2 (get_argparse_type_for_container from simple_parsing/utils.py) to hit the second branch.

Old coverage result

New coverage result

Adding a new test case to cover the second branch of the function resulted in a coverage improvement of 11%, bringing the total coverage for the function to 100%.

Overall

Old result:

New result:

Statement of individual contributions

Karina Sudnicina: found the project, implemented coverage tool for two functions mentioned above, created a test cases for them.

Nora Cai: implemeted coverage tool for two functions mentioned above, created test cases for them.

Sanne Aerts: Created a function-based coverage tool for two chosen functions and developed new test cases for them to enhance their coverage.

Che Lai: imlemented coverage tool for two functions and enhance their coverage by creating new test cases.