def test_properly_converts_timing(self):
caption_set = DFXPReader().read(DFXP_WITH_ALTERNATIVE_TIMING_FORMATS)
caps = caption_set.get_captions('en-US')
self.assertEqual(caps[0].start, 1900000)
self.assertEqual(caps[0].end, 3050000)
self.assertEqual(caps[1].start, 4000000)
self.assertEqual(caps[1].end, 5200000)
def test_proper_xml_entity_escaping(self):
caption_set = DFXPReader().read(DFXP_WITH_ESCAPED_APOSTROPHE)
cue_text = caption_set.get_captions('en-US')[0].nodes[0].content
self.assertEqual(cue_text,
"<< \"Andy's Caf\xe9 & Restaurant\" this way")
result = DFXPWriter().write(caption_set)
self.assertIn("<< \"Andy's Café & Restaurant\" this way",
result)
def test_individual_timings_of_captions_with_matching_timespec_are_kept(
self): # noqa
captionset = DFXPReader().read(
SAMPLE_DFXP_MULTIPLE_CAPTIONS_WITH_THE_SAME_TIMING)
expected_timings = [(9209000, 12312000)] * 3
actual_timings = [(c_.start, c_.end)
for c_ in captionset.get_captions('en-US')]
self.assertEqual(expected_timings, actual_timings)
def test_merge_concurrent_captions(self):
caption_set = DFXPReader().read(DFXP_WITH_CONCURRENT_CAPTIONS)
captions = caption_set.get_captions('en-US')
self.assertEqual(len(captions), 5)
caption_set = merge_concurrent_captions(caption_set)
captions = caption_set.get_captions('en-US')
self.assertEqual(len(captions), 3)
def test_individual_texts_of_captions_with_matching_timespec_are_kept(
self): # noqa
captionset = DFXPReader().read(
SAMPLE_DFXP_MULTIPLE_CAPTIONS_WITH_THE_SAME_TIMING)
expected_texts = [
'Some text here', 'Some text there',
'Caption texts are everywhere!'
]
actual_texts = [
c_.nodes[0].content for c_ in captionset.get_captions("en-US")
]
self.assertEqual(expected_texts, actual_texts)
def test_default_region_p_tags(self):
caption_set = DFXPReader().read(SAMPLE_DFXP)
result = DFXPWriter().write(caption_set)
soup = BeautifulSoup(result, 'lxml')
for p in soup.find_all('p'):
self.assertEqual(p.attrs.get('region'), DFXP_DEFAULT_REGION_ID)
def test_fit_to_screen(self):
# Check if caption width and height are is explicitly set and
# recalculate it if necessary. This prevents long captions from being
# cut out of the screen.
caption_set = DFXPReader().read(SAMPLE_DFXP_LONG_CUE)
result = DFXPWriter().write(caption_set)
self.assertEqual(result, SAMPLE_DFXP_LONG_CUE_FIT_TO_SCREEN)
def test_dfxp_with_positioning_to_webvtt_conversion(self):
caption_set = DFXPReader().read(SAMPLE_DFXP_WITH_POSITIONING)
results = WebVTTWriter(video_width=VIDEO_WIDTH,
video_height=VIDEO_HEIGHT).write(caption_set)
self.assertTrue(isinstance(results, text_type))
self.assertWebVTTEquals(
SAMPLE_WEBVTT_FROM_DFXP_WITH_POSITIONING_AND_STYLE, results)
def test_dfxp_to_webvtt_preserves_proper_alignment(self):
# This failed at one point when the CaptionSet had node breaks with
# different positioning. It was fixed both at the DFXPReader AND the
# WebVTTWriter.
caption_set = DFXPReader().read(DFXP_STYLE_REGION_ALIGN_CONFLICT)
results = WebVTTWriter().write(caption_set)
self.assertEqual(WEBVTT_FROM_DFXP_WITH_CONFLICTING_ALIGN, results)
def test_legacy_convert(self):
caption_set = DFXPReader(read_invalid_positioning=True).read(
SAMPLE_DFXP_FOR_LEGACY_WRITER_INPUT)
result = LegacyDFXPWriter().write(caption_set)
self.assertEqual(result, SAMPLE_DFXP_FOR_LEGACY_WRITER_OUTPUT)
def test_default_styling_p_tags(self):
caption_set = DFXPReader().read(SAMPLE_DFXP)
result = DFXPWriter().write(caption_set)
soup = BeautifulSoup(result, 'lxml')
for p in soup.find_all('p'):
self.assertEqual(p.attrs.get('style'), 'p')
def test_offset_time(self):
reader = DFXPReader()
self.assertEqual(1, reader._translate_time("0.001ms"))
self.assertEqual(2000, reader._translate_time("2ms"))
self.assertEqual(1000000, reader._translate_time("1s"))
self.assertEqual(1234567, reader._translate_time("1.234567s"))
self.assertEqual(180000000, reader._translate_time("3m"))
self.assertEqual(14400000000, reader._translate_time("4h"))
# Tick values are not supported
self.assertRaises(InvalidInputError, reader._translate_time, "2.3t")
def test_individual_layouts_of_captions_with_matching_timespec_are_kept(
self): # noqa
captionset = DFXPReader().read(
SAMPLE_DFXP_MULTIPLE_CAPTIONS_WITH_THE_SAME_TIMING)
expected_layouts = [
(((10, UnitEnum.PERCENT), (10, UnitEnum.PERCENT)), None, None,
(HorizontalAlignmentEnum.CENTER, VerticalAlignmentEnum.BOTTOM)),
(((40, UnitEnum.PERCENT), (40, UnitEnum.PERCENT)), None, None,
(HorizontalAlignmentEnum.CENTER, VerticalAlignmentEnum.BOTTOM)),
(((10, UnitEnum.PERCENT), (70, UnitEnum.PERCENT)), None, None,
(HorizontalAlignmentEnum.CENTER, VerticalAlignmentEnum.BOTTOM))
]
actual_layouts = [
c_.layout_info.serialized()
for c_ in captionset.get_captions('en-US')
]
self.assertEqual(expected_layouts, actual_layouts)
def test_dfxp_to_sami_with_margins(self):
caption_set = DFXPReader().read(SAMPLE_DFXP_FROM_SAMI_WITH_MARGINS)
results = SAMIWriter(video_width=VIDEO_WIDTH,
video_height=VIDEO_HEIGHT).write(caption_set)
margins = [
"margin-right: 6.04%;", "margin-bottom: 0%;", "margin-top: 0%;",
"margin-left: 6.04%;"
]
for margin in margins:
self.assertIn(margin, results)
def test_dont_create_style_tags_with_no_id(self):
# The <style> tags can have no 'xml:id' attribute. Previously, in this
# case, the style was copied to the output file, with the 'xml:id'
# property declared, but no value assigned to it. Since such a style
# can not be referred anyway, and <style> elements, children of
# <region> tags shouldn't be referred to anyway, we don't include
# these styles in the output file
caption_set = DFXPReader().read(
SAMPLE_DFXP_STYLE_TAG_WITH_NO_XML_ID_INPUT)
result = DFXPWriter().write(caption_set)
self.assertEqual(result, SAMPLE_DFXP_STYLE_TAG_WITH_NO_XML_ID_OUTPUT)
def test_default_styling_tag(self):
caption_set = DFXPReader().read(SAMPLE_DFXP_WITHOUT_REGION_AND_STYLE)
result = DFXPWriter().write(caption_set)
default_style = _recreate_style(DFXP_DEFAULT_STYLE, None)
default_style['xml:id'] = DFXP_DEFAULT_STYLE_ID
soup = BeautifulSoup(result, 'lxml-xml')
style = soup.find('style', {'xml:id': DFXP_DEFAULT_STYLE_ID})
self.assertTrue(style)
self.assertEqual(style.attrs, default_style)
def test_default_region_tag(self):
caption_set = DFXPReader().read(SAMPLE_DFXP)
result = DFXPWriter().write(caption_set)
soup = BeautifulSoup(result, 'lxml-xml')
region = soup.find('region', {'xml:id': DFXP_DEFAULT_REGION_ID})
default_region = _convert_layout_to_attributes(DFXP_DEFAULT_REGION)
default_region['xml:id'] = DFXP_DEFAULT_REGION_ID
self.assertTrue(region)
self.assertEqual(region.attrs['xml:id'], DFXP_DEFAULT_REGION_ID)
self.assertEqual(region.attrs, default_region)
def test_incorrectly_specified_positioning_is_explicitly_accepted(self):
# The arguments used here illustrate how we will try to read
# and write incorrectly specified positioning information.
# By incorrect, I mean the specs say that those attributes should be
# ignored, not that the attributes themselves are outside of the specs
caption_set = DFXPReader(read_invalid_positioning=True).read(
SAMPLE_DFXP_INVALID_BUT_SUPPORTED_POSITIONING_INPUT)
result = DFXPWriter(relativize=False,
fit_to_screen=False,
write_inline_positioning=True).write(caption_set)
if six.PY2:
self.assertDFXPEquals(
result, SAMPLE_DFXP_INVALID_BUT_SUPPORTED_POSITIONING_OUTPUT)
else:
# attributes are sorted differently I guess testing for same
# length is close enough
self.assertDFXPEquals(
result, SAMPLE_DFXP_INVALID_BUT_SUPPORTED_POSITIONING_OUTPUT)
def test_dfxp_to_sami_conversion(self):
caption_set = DFXPReader().read(SAMPLE_DFXP)
results = SAMIWriter().write(caption_set)
self.assertTrue(isinstance(results, text_type))
self.assertSAMIEquals(SAMPLE_SAMI, results)
def test_caption_error_for_invalid_positioning_values(self):
invalid_value_dfxp = (
SAMPLE_DFXP_INVALID_POSITIONING_VALUE_TEMPLATE.format(
origin="px 5px"))
self.assertRaises(CaptionReadSyntaxError,
DFXPReader().read, invalid_value_dfxp)
def test_caption_error_for_invalid_or_unsupported_positioning_units(self):
invalid_dfxp = (SAMPLE_DFXP_INVALID_POSITIONING_VALUE_TEMPLATE.format(
origin="6foo 7bar"))
self.assertRaises(CaptionReadSyntaxError,
DFXPReader().read, invalid_dfxp)
def test_dfxp_to_dfxp_conversion(self):
caption_set = DFXPReader().read(SAMPLE_DFXP)
results = DFXPWriter().write(caption_set)
self.assertTrue(isinstance(results, text_type))
self.assertDFXPEquals(SAMPLE_DFXP_OUTPUT, results)
def test_dfxp_reader_only_supports_unicode_input(self):
with self.assertRaises(InvalidInputError):
DFXPReader().read(b'')
def test_dfxp_to_webvtt_adds_explicit_size(self):
caption_set = DFXPReader().read(SAMPLE_DFXP_LONG_CUE)
results = WebVTTWriter().write(caption_set)
self.assertTrue(isinstance(results, text_type))
self.assertEqual(SAMPLE_WEBVTT_OUTPUT_LONG_CUE, results)
def test_is_relativized(self):
# Absolute positioning settings (e.g. px) are converted to percentages
caption_set = DFXPReader().read(SAMPLE_DFXP_WITH_POSITIONING)
result = DFXPWriter(video_width=VIDEO_WIDTH,
video_height=VIDEO_HEIGHT).write(caption_set)
self.assertEqual(result, SAMPLE_DFXP_WITH_RELATIVIZED_POSITIONING)
def test_dfxp_with_inherited_style_to_webvtt_conversion(self):
caption_set = DFXPReader().read(SAMPLE_DFXP_WITH_INHERITED_STYLE)
results = WebVTTWriter().write(caption_set)
self.assertTrue(isinstance(results, text_type))
self.assertWebVTTEquals(SAMPLE_WEBVTT_FROM_DFXP_WITH_STYLE, results)
def test_dfxp_to_webvtt_conversion(self):
caption_set = DFXPReader().read(SAMPLE_DFXP)
results = WebVTTWriter().write(caption_set)
self.assertTrue(isinstance(results, text_type))
self.assertWebVTTEquals(SAMPLE_WEBVTT_FROM_DFXP, results)
def test_invalid_markup_is_properly_handled(self):
captions = DFXPReader().read(SAMPLE_DFXP_SYNTAX_ERROR)
self.assertEqual(2, len(captions.get_captions("en-US")))
def test_correct_region_attributes_are_recreated(self):
caption_set = DFXPReader().read(SAMPLE_DFXP_MULTIPLE_REGIONS_INPUT)
result = DFXPWriter(relativize=False,
fit_to_screen=False).write(caption_set)
self.assertDFXPEquals(result, SAMPLE_DFXP_MULTIPLE_REGIONS_OUTPUT)
def test_empty_file(self):
self.assertRaises(CaptionReadNoCaptions,
DFXPReader().read, SAMPLE_DFXP_EMPTY)