KeyArgo
/
youtube-transcript-api
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

formatters.py python 2 compliant and misc. 

Remove use of ABC, not part of Python 2 and implementing ABCMeta changes between Python 2 and 3 so left it out entirely. Base class now raises NotImplementedError manually.

Fix parse_timecode issue with start and end times being identical

Replaced uses of F-strings with .format() also for compatibility.
This commit is contained in:
Chris Howell 2020-07-09 20:13:58 -07:00
parent 74d36a821e
commit c4b8b5b18d
1 changed files with 22 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

43
youtube_transcript_api/formatters.py
View File

@ -1,5 +1,3 @@
from abc import ABC
from abc import abstractclassmethod
from collections import defaultdict from collections import defaultdict
import json import json
import re import re
@ -17,17 +15,18 @@ def parse_timecode(time):
>>> parse_timecode(6.93) >>> parse_timecode(6.93)
'00:00:06,930' '00:00:06,930'
""" """
time = float(time) time = float(time)
hours, mins, secs = ( times = {
str(int(time)//3600).rjust(2, '0'), 'hours': str(int(time) // 3600).rjust(2, '0'),
str(int(time)//60).rjust(2, '0'), 'mins': str(int(time) // 60).rjust(2, '0'),
str(int(time)%60).rjust(2, '0'), 'secs': str(int(time) % 60).rjust(2, '0'),
) 'ms': str(int(round((time - int(time))*1000, 2))).rjust(3, '0')
ms = str(int(round((time - int(time))*1000, 2))).rjust(3, '0') }
return f"{hours}:{mins}:{secs},{ms}" return "{hours}:{mins}:{secs},{ms}".format(**times)
class TranscriptFormatter(ABC): class TranscriptFormatter(object):
"""Abstract Base TranscriptFormatter class """Abstract Base TranscriptFormatter class
This class should be inherited from to create additional This class should be inherited from to create additional
@ -51,7 +50,7 @@ class TranscriptFormatter(ABC):
return cls.DELIMITER.join( return cls.DELIMITER.join(
str(transcript) for transcript in transcripts) str(transcript) for transcript in transcripts)
@abstractclassmethod @classmethod
def format(cls, transcript_data): def format(cls, transcript_data):
"""Any subclass must implement this format class method. """Any subclass must implement this format class method.
@ -61,7 +60,9 @@ class TranscriptFormatter(ABC):
as a string. as a string.
:rtype: list[str] :rtype: list[str]
""" """
pass raise NotImplementedError(
cls.__name__ + '.format'
)
class JSONTranscriptFormatter(TranscriptFormatter): class JSONTranscriptFormatter(TranscriptFormatter):
@ -103,22 +104,20 @@ class SRTTranscriptFormatter(TranscriptFormatter):
output = [] output = []
for frame, item in enumerate(transcript_data, start=1): for frame, item in enumerate(transcript_data, start=1):
start_time = float(item.get('start')) start_time = float(item.get('start'))
duration = float(item.get('dur', '0.0')) duration = float(item.get('duration', '0.0'))
end_time = parse_timecode(start_time + duration)
start_time = parse_timecode(start_time)
output.append("{frame}\n".format(frame=frame)) output.append("{frame}\n".format(frame=frame))
output.append("{start_time} --> {end_time}\n".format( output.append("{start_time} --> {end_time}\n".format(
start_time=start_time, end_time=end_time)) start_time=parse_timecode(start_time),
end_time=parse_timecode(start_time + duration)
))
output.append("{text}".format(text=item.get('text'))) output.append("{text}".format(text=item.get('text')))
if frame < len(transcript_data): if frame < len(transcript_data):
output.append('\n\n') output.append('\n\n')
return '{}\n'.format(''.join(output)) return '{}\n'.format(''.join(output))
class TranscriptFormatterFactory: class TranscriptFormatterFactory(object):
"""A Transcript Class Factory """A Transcript Class Factory
Allows for adding additional custom Transcript classes for the API Allows for adding additional custom Transcript classes for the API
@ -139,8 +138,10 @@ class TranscriptFormatterFactory:
:rtype None :rtype None
""" """
if not issubclass(formatter_class, TranscriptFormatter): if not issubclass(formatter_class, TranscriptFormatter):
raise TypeError( raise TypeError((
f'{formatter_class} must be a subclass of TranscriptFormatter') '{0} must be a subclass of TranscriptFormatter'
).format(formatter_class)
)
self._formatters.update({name: formatter_class}) self._formatters.update({name: formatter_class})
def add_formatters(self, formatters_dict): def add_formatters(self, formatters_dict):