def validate_sample(ext, expected_transcripts, out_segments):
base_output = f"{test_dir}/good"
convert(f"{sample_dir}/invalid.stm", base_output + ext)
validated_transcript = time_aligned_text(base_output + ext)
assert len(validated_transcript.segments) == out_segments
for seg, expected_text in zip(validated_transcript.segments, expected_transcripts):
assert seg.text == expected_text
def validate(input_file):
"""
Overwrites an STM file, leaving only valid input lines
"""
# after reading in, only valid lines will remain
# so write it back in place
convert(input_file, input_file)
def validate_sample(ext, expected_transcripts, out_segments):
base_output = 'tests/good'
convert('samples/invalid.stm', base_output + ext)
validated_transcript = time_aligned_text(base_output + ext)
assert len(validated_transcript.segments) == out_segments
for seg, expected_text in zip(validated_transcript.segments,
expected_transcripts):
assert seg.text == expected_text
def validate():
parser = argparse.ArgumentParser(
description='convert between text file formats')
parser.add_argument('input_file',
metavar='input_file',
type=str,
help='input file')
args = parser.parse_args()
# after reading in, only valid lines will remain
# so write it back in place
convert(args.input_file, args.input_file)