Script to parse changed Robot Framework tests from git diff

differ.py

'''
  $ git diff --unified=0 path/to/robot | python differ.py
'''
import re
import sys

from collections import defaultdict
from pathlib import Path

from robot.api import SuiteVisitor, TestSuiteBuilder

CURDIR = Path.cwd()

def get_absolute_filepath(fp):
    return CURDIR / fp

def parse(gitdiff):
    changed_files = defaultdict(set)
    current_suite_file = ''
    test_case_name = ''
    for line in gitdiff:
        if line.startswith('+++'):
            _, current_suite_file = line.split('b/', 1)
            current_suite_file = get_absolute_filepath(current_suite_file.strip())
            continue
        if line.startswith('@@'):
            _, test_case_name = line.rsplit('@@', 1)
        if re.match('^\+\w', line) != None:
            test_case_name = line[1:]
        changed_files[current_suite_file].add(test_case_name.strip())
    # if there is an entry where they key is empty string, remove it
    changed_files.pop('')
    return changed_files

def get_actual_tcs(filepaths):
    class TestCaseExtractor(SuiteVisitor):
        def __init__(self):
            self.data = defaultdict(list)

        def start_test(self, test):
            self.data[Path(test.parent.source)].append(test.name)

    extractor = TestCaseExtractor()
    suites = [TestSuiteBuilder().build(fp) for fp in filepaths]
    for s in suites:
        s.visit(extractor)
    return extractor.data

def filter_out_non_testcases(changed_files, actual_test_cases):
    retval = {}
    for suitepath, tests in changed_files.items():
      retval[suitepath] = [t for t in tests if t in actual_test_cases[suitepath]]
    return retval

def format_(changed_files):
    retval = []
    for changed_file, changed_tests in changed_files.items():
        suite_name = str(changed_file).replace(f'{CURDIR}/', '')
        # needs to end in slash, otherwise suite_name is '/foo/bar' instead
        # of 'foo/bar'
        suite_name = suite_name.replace('/', '.')
        suite_name, _ = suite_name.rsplit('.robot', 1)
        for test in changed_tests:
            t = test.replace(' ', '')
            retval.append(f'--suite {suite_name} --test {t}')
    return retval

def main(gitdiff):
    changed_files = parse(gitdiff)
    actual_test_cases = get_actual_tcs(changed_files.keys())
    changed_files = filter_out_non_testcases(changed_files, actual_test_cases)
    robot_args = format_(changed_files)
    print('\n'.join(robot_args))

if __name__ == '__main__':
   main(sys.stdin.readlines())