def test_to_file(self):
for codec in ['utf-8', 'latin_1', 'mac_roman']:
self.tg = TextGrid(xmax=20)
tier1 = self.tg.add_tier('tier')
tier1.add_interval(1, 2, 'i1')
tier1.add_interval(2, 3, 'i2')
tier1.add_interval(4, 5, 'i3')
tier4 = self.tg.add_tier('tier')
tier4.add_interval(1, 2, u'i1ü')
tier4.add_interval(2.0, 3, 'i2')
tier4.add_interval(4, 5.0, 'i3')
tier2 = self.tg.add_tier('tier2', tier_type='TextTier')
tier2.add_point(1, u'p1ü')
tier2.add_point(2, 'p1')
tier2.add_point(3, 'p1')
tempf = tempfile.mkstemp()[1]
# Normal mode
self.tg.to_file(tempf, codec=codec)
TextGrid(tempf, codec=codec)
# Short mode
self.tg.to_file(tempf, codec=codec, mode='s')
TextGrid(tempf, codec=codec)
# Binary mode
self.tg.to_file(tempf, mode='b')
TextGrid(tempf)
os.remove(tempf)
def test_to_file(codec, tmp_path):
tg = TextGrid(xmax=20)
tier1 = tg.add_tier('tier')
tier1.add_interval(1, 2, 'i1')
tier1.add_interval(2, 3, 'i2')
tier1.add_interval(4, 5, 'i3')
tier4 = tg.add_tier('tier')
tier4.add_interval(1, 2, u'i1ü')
tier4.add_interval(2.0, 3, 'i2')
tier4.add_interval(4, 5.0, 'i3')
tier2 = tg.add_tier('tier2', tier_type='TextTier')
tier2.add_point(1, u'p1ü')
tier2.add_point(2, 'p1')
tier2.add_point(3, 'p1')
tempf = str(tmp_path / 'test')
# Normal mode
tg.to_file(pathlib.Path(tempf), codec=codec)
TextGrid(tempf, codec=codec)
# Short mode
tg.to_file(tempf, codec=codec, mode='s')
TextGrid(tempf, codec=codec)
# Binary mode
tg.to_file(tempf, mode='b')
TextGrid(tempf)
def write_to_text_grid(words: List[dict], sentences: List[dict],
duration: float):
"""Write results to Praat TextGrid. Because we are using pympi, we can also export to Elan EAF.
Args:
words (List[dict]): List of word times containing start, end, and value keys
sentences (List[dict]): List of sentence times containing start, end, and value keys
duration (float): duration of entire audio
Returns:
TextGrid: Praat TextGrid with word and sentence alignments
"""
text_grid = TextGrid(xmax=duration)
sentence_tier = text_grid.add_tier(name="Sentence")
word_tier = text_grid.add_tier(name="Word")
for s in sentences:
sentence_tier.add_interval(
begin=s[0]["start"],
end=s[-1]["end"],
value=" ".join([w["text"] for w in s]),
)
for w in words:
word_tier.add_interval(begin=w["start"], end=w["end"], value=w["text"])
return text_grid
def toTextGrid(self, filePath, excludedTiers=[], includedTiers=[]):
"""
Convert the elan file to praat's TextGrid, returns 0 if succesfull<br />
<br />
filePath -- The output file path - for stdout<br />
excludedTiers -- Tiers to exclude<br />
includedTiers -- Tiers to include if empty all tiers are included"""
try:
from pympi.Praat import TextGrid
except ImportError:
warnings.warn(
'Please install the pympi.Praat module from the py' +
'mpi module found at https://github.com/dopefishh/pympi')
return 1
tgout = TextGrid()
for tier in [
a for a in self.tiers if a not in excludedTiers and (
not includedTiers or a in includedTiers)
]:
currentTier = tgout.addTier(tier)
for interval in self.getAnnotationDataForTier(tier):
if interval[0] == interval[1]:
continue
currentTier.addInterval(interval[0] / 1000.0,
interval[1] / 1000.0, interval[2])
tgout.tofile(filePath)
return 0
def setUp(self):
self.tg = TextGrid(xmax=20)
self.maxdiff = None
def test_to_file(self):
for codec in ['utf-8', 'utf-16', 'latin_1', 'mac_roman']:
self.tg = TextGrid(xmax=20)
tier1 = self.tg.add_tier('tier')
tier1.add_interval(1, 2, 'i1')
tier1.add_interval(2, 3, 'i2')
tier1.add_interval(4, 5, 'i3')
tier4 = self.tg.add_tier('tier')
tier4.add_interval(1, 2, u'i1ü')
tier4.add_interval(2.0, 3, 'i2')
tier4.add_interval(4, 5.0, 'i3')
tier2 = self.tg.add_tier('tier2', tier_type='TextTier')
tier2.add_point(1, u'p1ü')
tier2.add_point(2, 'p1')
tier2.add_point(3, 'p1')
# Normal mode
tgfile = io.StringIO()
self.tg.to_stream(tgfile, codec=codec)
tgfile.seek(0)
tg1 = tgfile.read()
tgfile.seek(0)
self.tg = TextGrid(tgfile, codec=codec, stream=True)
tgfile = io.StringIO()
self.tg.to_stream(tgfile, codec=codec)
tgfile.seek(0)
tg2 = tgfile.read()
tgfile.seek(0)
self.assertEqual(tg2, tg1)
# Short mode
tgfile = io.StringIO()
self.tg.to_stream(tgfile, codec=codec, mode='s')
tgfile.seek(0)
tg1 = tgfile.read()
tgfile.seek(0)
self.tg = TextGrid(tgfile, codec=codec, stream=True)
tgfile = io.StringIO()
self.tg.to_stream(tgfile, codec=codec, mode='s')
tgfile.seek(0)
tg2 = tgfile.read()
tgfile.seek(0)
self.assertEqual(tg2, tg1)
# Binary mode
tgfile = io.BytesIO()
self.tg.to_stream(tgfile, codec=codec, mode='b')
tgfile.seek(0)
tg1 = tgfile.read()
tgfile.seek(0)
self.tg = TextGrid(tgfile, codec=codec, stream=True)
tgfile = io.BytesIO()
self.tg.to_stream(tgfile, codec=codec, mode='b')
tgfile.seek(0)
tg2 = tgfile.read()
tgfile.seek(0)
self.assertEqual(tg2, tg1)
