|
import re |
|
|
|
def getSpeechBlocks(whispered, silence_time=2): |
|
text_blocks, (st, et, txt) = [], (0,0,"") |
|
for i, seg in enumerate(whispered['segments']): |
|
if seg['start'] - et > silence_time: |
|
if txt: text_blocks.append([[st, et], txt]) |
|
(st, et, txt) = (seg['start'], seg['end'], seg['text']) |
|
else: |
|
et, txt = seg['end'], txt + seg['text'] |
|
|
|
if txt: text_blocks.append([[st, et], txt]) |
|
|
|
return text_blocks |
|
|
|
def cleanWord(word): |
|
return re.sub(r'[^\w\s\-_"\'\']', '', word) |
|
|
|
def interpolateTimeFromDict(word_position, d): |
|
for key, value in d.items(): |
|
if key[0] <= word_position <= key[1]: |
|
return value |
|
return None |
|
|
|
def getTimestampMapping(whisper_analysis): |
|
index = 0 |
|
locationToTimestamp = {} |
|
for segment in whisper_analysis['segments']: |
|
for word in segment['words']: |
|
newIndex = index + len(word['text'])+1 |
|
locationToTimestamp[(index, newIndex)] = word['end'] |
|
index = newIndex |
|
return locationToTimestamp |
|
|
|
|
|
def splitWordsBySize(words, maxCaptionSize): |
|
halfCaptionSize = maxCaptionSize / 2 |
|
captions = [] |
|
while words: |
|
caption = words[0] |
|
words = words[1:] |
|
while words and len(caption + ' ' + words[0]) <= maxCaptionSize: |
|
caption += ' ' + words[0] |
|
words = words[1:] |
|
if len(caption) >= halfCaptionSize and words: |
|
break |
|
captions.append(caption) |
|
return captions |
|
|
|
def getCaptionsWithTime(whisper_analysis, maxCaptionSize=15, considerPunctuation=False): |
|
wordLocationToTime = getTimestampMapping(whisper_analysis) |
|
position = 0 |
|
start_time = 0 |
|
CaptionsPairs = [] |
|
text = whisper_analysis['text'] |
|
|
|
if considerPunctuation: |
|
sentences = re.split(r'(?<=[.!?]) +', text) |
|
words = [word for sentence in sentences for word in splitWordsBySize(sentence.split(), maxCaptionSize)] |
|
else: |
|
words = text.split() |
|
words = [cleanWord(word) for word in splitWordsBySize(words, maxCaptionSize)] |
|
|
|
for word in words: |
|
position += len(word) + 1 |
|
end_time = interpolateTimeFromDict(position, wordLocationToTime) |
|
if end_time and word: |
|
CaptionsPairs.append(((start_time, end_time), word)) |
|
start_time = end_time |
|
|
|
return CaptionsPairs |
