class IncRecoTest(unittest.TestCase):
def setUp(self):
self.reco = IncReco('data/test.inc_reco')
def test_methods(self):
self.failUnlessEqual(self.reco[0], {
'Chunk': [['0.08', '0.13', '\xc3\xa4h']],
'Time': 0.13
})
self.failUnlessEqual(str(self.reco)[143:155], "'Time': 0.71")
self.failUnlessEqual(len(self.reco), 106)
self.failUnlessEqual(
list(self.reco)[55], {
'Chunk': [['6.95', '7.10', 'ph']],
'Time': 7.1
})
self.failUnlessEqual(self.reco.get_latest_chunk(13.66)['Time'], 13.58)
self.failUnlessEqual(self.reco.get_latest_chunk(13.67)['Time'], 13.67)
self.failUnlessEqual(self.reco.get_latest_chunk(13.68)['Time'], 13.67)
self.failUnlessEqual(self.reco.get_last_chunk()['Time'], 38.4)
self.failUnlessEqual(self.reco.get_times()[33], 3.56)
self.failUnlessEqual([x['Time'] for x in self.reco[5:10]],
[0.82, 0.93, 0.98, 1.02, 1.08])
class IncRecoTest(unittest.TestCase):
def setUp(self):
self.reco = IncReco('data/test.inc_reco')
def test_methods(self):
self.failUnlessEqual(self.reco[0], {'Chunk': [['0.08',
'0.13',
'\xc3\xa4h']],
'Time': 0.13})
self.failUnlessEqual(str(self.reco)[143:155], "'Time': 0.71")
self.failUnlessEqual(len(self.reco), 106)
self.failUnlessEqual(list(self.reco)[55], {'Chunk': [['6.95',
'7.10',
'ph']],
'Time': 7.1})
self.failUnlessEqual(self.reco.get_latest_chunk(13.66)['Time'], 13.58)
self.failUnlessEqual(self.reco.get_latest_chunk(13.67)['Time'], 13.67)
self.failUnlessEqual(self.reco.get_latest_chunk(13.68)['Time'], 13.67)
self.failUnlessEqual(self.reco.get_last_chunk()['Time'], 38.4)
self.failUnlessEqual(self.reco.get_times()[33], 3.56)
self.failUnlessEqual([x['Time'] for x in self.reco[5:10]],
[0.82, 0.93, 0.98, 1.02, 1.08])
def open_intervalframe_from_increco(filepath,
encoding='utf-8',
lastonly=False):
""" Create an interval frame from an inc_reco file
Creates a dictionary with an intervalframe per chunk. The dictionary
keys are the times of the chunks
Arguments:
filepath - the path to the inc_reco file
Kwargs:
encoding -- the encoding of the file
lastonly -- Read only the last chunk in the inc_reco file
rather than all chunks
"""
reco = IncReco(filepath)
start_chunk = -1 if lastonly else 0
frame_dict = dict()
for chunk in reco[start_chunk:]:
cur_frame = None
for line in chunk['Chunk']:
reco_frame = pd.DataFrame([{"start_time": \
float(line[0].decode(encoding)),
"end_time": \
float(line[1].decode(encoding)),
"text": line[2].decode(encoding)}])
if cur_frame is None:
cur_frame = reco_frame
else:
cur_frame = cur_frame.append(reco_frame, ignore_index=True)
#check for minor errors in the times
#Sometimes the end of the previous word is 1 ms later than
#the start of the next word
for i in cur_frame.index[1:]:
cur_frame.ix[i,
'start_time'] = max(cur_frame['start_time'].ix[i],
cur_frame['end_time'].ix[i - 1])
frame_dict[str(
chunk['Time'])] = cur_frame.ix[:,
['start_time', 'end_time', 'text']]
return frame_dict
def setUp(self):
self.reco = IncReco('data/test.inc_reco')
def setUp(self):
self.reco = IncReco('data/test.inc_reco')