6654967569

Committed 26 Oct 2023 01:39PM UTC coverage: 70.273% (-0.04%) from 70.317%

Build # 6654967569

Build Type

Pull #993

github

Committed by

web-flow

Commit Message

change version to 3.5.0 (#994)

Pull Request Pull Request #993: [WIP] Develop -> 3.5.0 (#980)

Run Details

69 of 69 new or added lines in 5 files covered. (100.0%)

9340 of 13291 relevant lines covered (70.27%)

0.7 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

0.0

/pod/video_encode_transcript/transcript_model.py

import numpy as np
import shlex
import subprocess
import json

import sys
import os
from timeit import default_timer as timer
import datetime as dt
from datetime import timedelta

from webvtt import WebVTT, Caption

try:
    from shhlex import quote
except ImportError:
    from pipes import quote

import logging

try:
    from ..custom import settings_local
except ImportError:
    from .. import settings as settings_local

DEBUG = getattr(settings_local, "DEBUG", False)

TRANSCRIPTION_MODEL_PARAM = getattr(settings_local, "TRANSCRIPTION_MODEL_PARAM", False)
USE_TRANSCRIPTION = getattr(settings_local, "USE_TRANSCRIPTION", False)
if USE_TRANSCRIPTION:
    TRANSCRIPTION_TYPE = getattr(settings_local, "TRANSCRIPTION_TYPE", "VOSK")
    if TRANSCRIPTION_TYPE == "VOSK":
        from vosk import Model, KaldiRecognizer
    elif TRANSCRIPTION_TYPE == "STT":
        from stt import Model
    elif TRANSCRIPTION_TYPE == "WHISPER":
        import whisper

TRANSCRIPTION_NORMALIZE = getattr(settings_local, "TRANSCRIPTION_NORMALIZE", False)
TRANSCRIPTION_NORMALIZE_TARGET_LEVEL = getattr(
    settings_local, "TRANSCRIPTION_NORMALIZE_TARGET_LEVEL", -16.0
)

TRANSCRIPTION_AUDIO_SPLIT_TIME = getattr(
    settings_local, "TRANSCRIPTION_AUDIO_SPLIT_TIME", 600
)  # 10min
# time in sec for phrase length
TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH = getattr(
    settings_local, "TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH", 2
)
TRANSCRIPTION_STT_SENTENCE_BLANK_SPLIT_TIME = getattr(
    settings_local, "TRANSCRIPTION_STT_SENTENCE_BLANK_SPLIT_TIME", 0.5
)
log = logging.getLogger(__name__)


def get_model(lang):
    """Get model for STT or Vosk software to transcript audio."""
    transript_model = Model(TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["model"])
    if TRANSCRIPTION_TYPE == "STT":
        if TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get("beam_width"):
            transript_model.setBeamWidth(
                TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["beam_width"]
            )
        if TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get("scorer"):
            print(
                "Loading scorer from files {}".format(
                    TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["scorer"]
                ),
                file=sys.stderr,
            )
            scorer_load_start = timer()
            transript_model.enableExternalScorer(
                TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["scorer"]
            )
            scorer_load_end = timer() - scorer_load_start
            print("Loaded scorer in {:.3}s.".format(scorer_load_end), file=sys.stderr)
            if TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get(
                "lm_alpha"
            ) and TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get("lm_beta"):
                transript_model.setScorerAlphaBeta(
                    TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["lm_alpha"],
                    TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["lm_beta"],
                )
    return transript_model


def start_transcripting(mp3filepath, duration, lang):
    """
    Start direct transcription.

    Normalize the audio if set, get the model according to the lang and start transcript.
    """
    if TRANSCRIPTION_NORMALIZE:
        mp3filepath = normalize_mp3(mp3filepath)
    if TRANSCRIPTION_TYPE == "WHISPER":
        msg, webvtt, all_text = main_whisper_transcript(mp3filepath, lang)
    else:
        transript_model = get_model(lang)
        msg, webvtt, all_text = start_main_transcript(
            mp3filepath, duration, transript_model
        )
    if DEBUG:
        print(msg)
        print(webvtt)
        print("\n%s\n" % all_text)

    return msg, webvtt


def start_main_transcript(mp3filepath, duration, transript_model):
    """Call transcription depending software type."""
    if TRANSCRIPTION_TYPE == "STT":
        msg, webvtt, all_text = main_stt_transcript(
            mp3filepath, duration, transript_model
        )
    elif TRANSCRIPTION_TYPE == "VOSK":
        msg, webvtt, all_text = main_vosk_transcript(
            mp3filepath, duration, transript_model
        )
    return msg, webvtt, all_text


def convert_samplerate(audio_path, desired_sample_rate, trim_start, duration):
    """Convert audio to subaudio and add good sample rate."""
    sox_cmd = "sox {} --type raw --bits 16 --channels 1 --rate {} ".format(
        quote(audio_path), desired_sample_rate
    )
    sox_cmd += "--encoding signed-integer --endian little --compression 0.0 "
    sox_cmd += "--no-dither - trim {} {}".format(trim_start, duration)

    try:
        output = subprocess.check_output(shlex.split(sox_cmd), stderr=subprocess.PIPE)

    except subprocess.CalledProcessError as e:
        raise RuntimeError("SoX returned non-zero status: {}".format(e.stderr))
    except OSError as e:
        raise OSError(
            e.errno,
            "SoX not found, use {}hz files or install it: {}".format(
                desired_sample_rate, e.strerror
            ),
        )

    return np.frombuffer(output, np.int16)


def normalize_mp3(mp3filepath):
    """Normalize the audio to good format and sound level."""
    filename, file_extension = os.path.splitext(mp3filepath)
    mp3normfile = "{}{}{}".format(filename, "_norm", file_extension)
    normalize_cmd = "ffmpeg-normalize {} ".format(quote(mp3filepath))
    normalize_cmd += "-c:a libmp3lame -b:a 192k --normalization-type ebu "
    # normalize_cmd += \
    # '--loudness-range-target 7.0 --true-peak 0.0 --offset 0.0 '
    normalize_cmd += "--target-level {} -f -o {}".format(
        TRANSCRIPTION_NORMALIZE_TARGET_LEVEL, quote(mp3normfile)
    )
    if DEBUG:
        print(normalize_cmd)
    try:
        subprocess.check_output(shlex.split(normalize_cmd), stderr=subprocess.PIPE)
        return mp3normfile
    except subprocess.CalledProcessError as e:
        log.error("ffmpeg-normalize returned non-zero status: {}".format(e.stderr))
        return mp3filepath
    except OSError as e:
        log.error("ffmpeg-normalize not found {}".format(e.strerror))
        return mp3filepath


# #################################
# TRANSCRIPT VIDEO: MAIN FUNCTION
# #################################


def convert_vosk_samplerate(audio_path, desired_sample_rate, trim_start, duration):
    """Convert audio to the good sample rate."""
    sox_cmd = "sox {} --type raw --bits 16 --channels 1 --rate {} ".format(
        quote(audio_path), desired_sample_rate
    )
    sox_cmd += "--encoding signed-integer --endian little --compression 0.0 "
    sox_cmd += "--no-dither - trim {} {}".format(trim_start, duration)

    try:
        output = subprocess.Popen(shlex.split(sox_cmd), stdout=subprocess.PIPE)

    except subprocess.CalledProcessError as e:
        raise RuntimeError("SoX returned non-zero status: {}".format(e.stderr))
    except OSError as e:
        raise OSError(
            e.errno,
            "SoX not found, use {}hz files or install it: {}".format(
                desired_sample_rate, e.strerror
            ),
        )
    return output


def get_word_result_from_data(results, audio, rec):
    """Get subsound from audio and add transcription to result parameter."""
    while True:
        data = audio.stdout.read(4000)
        if len(data) == 0:
            break
        if rec.AcceptWaveform(data):
            results.append(rec.Result())
    results.append(rec.Result())


def words_to_vtt(
    words,
    start_trim,
    duration,
    is_first_caption,
    text_caption,
    start_caption,
    last_word_added,
    all_text,
    webvtt,
):
    """Convert word and time to webvtt captions."""
    for index, word in enumerate(words):
        start_key = "start_time"
        word_duration = word.get("duration", 0)
        last_word = words[-1]
        last_word_duration = last_word.get("duration", 0)
        if TRANSCRIPTION_TYPE == "VOSK":
            start_key = "start"
            word_duration = word["end"] - word["start"]
            last_word_duration = words[-1]["end"] - words[-1]["start"]
        next_word = None
        blank_duration = 0
        if word != words[-1] and (index + 1) < len(words):
            next_word = words[index + 1]
            blank_duration = ((next_word[start_key]) - start_caption) - (
                ((word[start_key]) - start_caption) + word_duration
            )
        all_text += word["word"] + " "
        # word: <class 'dict'> {'word': 'bonjour', 'start ':
        # 0.58, 'duration': 7.34}
        text_caption.append(word["word"])
        if not (
            (((word[start_key]) - start_caption) < TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH)
            and (
                next_word is not None
                and (blank_duration < TRANSCRIPTION_STT_SENTENCE_BLANK_SPLIT_TIME)
            )
        ):
            # on créé le caption
            if is_first_caption:
                # A revoir, fusion de la nouvelle ligne avec
                # l'ancienne...
                is_first_caption = False
                text_caption = get_text_caption(text_caption, last_word_added)

            stop_caption = word[start_key] + word_duration

            # on evite le chevauchement
            change_previous_end_caption(webvtt, start_caption)

            caption = Caption(
                format_time_caption(start_caption),
                format_time_caption(stop_caption),
                " ".join(text_caption),
            )

            webvtt.captions.append(caption)
            # on remet tout à zero pour la prochaine phrase
            start_caption = word[start_key]
            text_caption = []
            last_word_added = word["word"]
    if start_trim + TRANSCRIPTION_AUDIO_SPLIT_TIME > duration:
        # on ajoute ici la dernière phrase de la vidéo
        stop_caption = start_trim + words[-1][start_key] + last_word_duration
        caption = Caption(
            format_time_caption(start_caption),
            format_time_caption(stop_caption),
            " ".join(text_caption),
        )
        webvtt.captions.append(caption)
    return all_text, webvtt


def main_vosk_transcript(norm_mp3_file, duration, transript_model):
    """Vosk transcription."""
    msg = ""
    inference_start = timer()
    msg += "\nInference start %0.3fs." % inference_start
    desired_sample_rate = 16000

    rec = KaldiRecognizer(transript_model, desired_sample_rate)
    rec.SetWords(True)

    webvtt = WebVTT()
    all_text = ""
    for start_trim in range(0, duration, TRANSCRIPTION_AUDIO_SPLIT_TIME):
        audio = convert_vosk_samplerate(
            norm_mp3_file,
            desired_sample_rate,
            start_trim,
            TRANSCRIPTION_AUDIO_SPLIT_TIME,  # dur
        )
        msg += "\nRunning inference."
        results = []
        get_word_result_from_data(results, audio, rec)
        for res in results:
            words = json.loads(res).get("result")
            text = json.loads(res).get("text")
            if not words:
                continue
            start_caption = words[0]["start"]
            stop_caption = words[-1]["end"]
            caption = Caption(
                format_time_caption(start_caption),
                format_time_caption(stop_caption),
                text,
            )
            webvtt.captions.append(caption)
            """
            text_caption = []
            is_first_caption = True
            all_text, webvtt = words_to_vtt(
                words,
                start_trim,
                duration,
                is_first_caption,
                text_caption,
                start_caption,
                last_word_added,
                all_text,
                webvtt,
            )
            """
    inference_end = timer() - inference_start

    msg += "\nInference took %0.3fs." % inference_end
    return msg, webvtt, all_text


def main_stt_transcript(norm_mp3_file, duration, transript_model):
    """STT transcription."""
    msg = ""
    inference_start = timer()
    msg += "\nInference start %0.3fs." % inference_start
    desired_sample_rate = transript_model.sampleRate()
    webvtt = WebVTT()
    last_word_added = ""
    metadata = None
    all_text = ""
    for start_trim in range(0, duration, TRANSCRIPTION_AUDIO_SPLIT_TIME):
        end_trim = (
            duration
            if start_trim + TRANSCRIPTION_AUDIO_SPLIT_TIME > duration
            else (
                start_trim
                + TRANSCRIPTION_AUDIO_SPLIT_TIME
                + TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH
            )
        )

        dur = (
            (TRANSCRIPTION_AUDIO_SPLIT_TIME + TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH)
            if (
                (
                    start_trim
                    + TRANSCRIPTION_AUDIO_SPLIT_TIME
                    + TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH
                )
                < duration
            )
            else (duration - start_trim)
        )

        msg += "\ntake audio from %s to %s - %s" % (start_trim, end_trim, dur)

        audio = convert_samplerate(norm_mp3_file, desired_sample_rate, start_trim, dur)
        msg += "\nRunning inference."

        metadata = transript_model.sttWithMetadata(audio)

        for transcript in metadata.transcripts:
            msg += "\nConfidence: %s" % transcript.confidence
            words = words_from_candidate_transcript(transcript)
            start_caption = start_trim + words[0]["start_time"]
            text_caption = []
            is_first_caption = True
            all_text, webvtt = words_to_vtt(
                words,
                start_trim,
                duration,
                is_first_caption,
                text_caption,
                start_caption,
                last_word_added,
                all_text,
                webvtt,
            )
    inference_end = timer() - inference_start

    msg += "\nInference took %0.3fs." % inference_end
    return msg, webvtt, all_text


def main_whisper_transcript(norm_mp3_file, lang):
    """Whisper transcription."""
    msg = ""
    all_text = ""
    webvtt = WebVTT()
    inference_start = timer()
    msg += "\nInference start %0.3fs." % inference_start

    model = whisper.load_model(
        TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["model"],
        download_root=TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang][
            "download_root"
        ],
    )

    transcription = model.transcribe(norm_mp3_file, language=lang)
    msg += "\nRunning inference."

    for segment in transcription["segments"]:
        caption = Caption(
            format_time_caption(segment["start"]),
            format_time_caption(segment["end"]),
            segment["text"],
        )
        webvtt.captions.append(caption)

    inference_end = timer() - inference_start
    msg += "\nInference took %0.3fs." % inference_end
    return msg, webvtt, all_text


def change_previous_end_caption(webvtt, start_caption):
    """Change the end time for caption."""
    if len(webvtt.captions) > 0:
        prev_end = dt.datetime.strptime(webvtt.captions[-1].end, "%H:%M:%S.%f")
        td_prev_end = timedelta(
            hours=prev_end.hour,
            minutes=prev_end.minute,
            seconds=prev_end.second,
            microseconds=prev_end.microsecond,
        ).total_seconds()
        if td_prev_end > start_caption:
            webvtt.captions[-1].end = format_time_caption(start_caption)


def format_time_caption(time_caption):
    """Format time for webvtt caption."""
    return (
        dt.datetime.utcfromtimestamp(0) + timedelta(seconds=float(time_caption))
    ).strftime("%H:%M:%S.%f")[:-3]


def get_text_caption(text_caption, last_word_added):
    """Get the text for a caption."""
    try:
        first_index = text_caption.index(last_word_added)
        return text_caption[first_index + 1 :]
    except ValueError:
        return text_caption


def words_from_candidate_transcript(metadata):
    """Get words list from transcription."""
    word = ""
    word_list = []
    word_start_time = 0
    # Loop through each character
    for i, token in enumerate(metadata.tokens):
        # Append character to word if it's not a space
        if token.text != " ":
            if len(word) == 0:
                # Log the start time of the new word
                word_start_time = token.start_time

            word = word + token.text
        # Word boundary is either a space or the last character in the array
        if token.text == " " or i == len(metadata.tokens) - 1:
            word_duration = token.start_time - word_start_time

            if word_duration < 0:
                word_duration = 0

            each_word = dict()
            each_word["word"] = word
            each_word["start_time"] = round(word_start_time, 4)
            each_word["duration"] = round(word_duration, 4)

            word_list.append(each_word)
            # Reset
            word = ""
            word_start_time = 0

    return word_list

1	import numpy as np	×
2	import shlex	×
3	import subprocess	×
4	import json	×
5
6	import sys	×
7	import os	×
8	from timeit import default_timer as timer	×
9	import datetime as dt	×
10	from datetime import timedelta	×
11
12	from webvtt import WebVTT, Caption	×
13
14	try:	×
15	from shhlex import quote	×
16	except ImportError:	×
17	from pipes import quote	×
18
19	import logging	×
20
21	try:	×
22	from ..custom import settings_local	×
23	except ImportError:	×
24	from .. import settings as settings_local	×
25
26	DEBUG = getattr(settings_local, "DEBUG", False)	×
27
28	TRANSCRIPTION_MODEL_PARAM = getattr(settings_local, "TRANSCRIPTION_MODEL_PARAM", False)	×
29	USE_TRANSCRIPTION = getattr(settings_local, "USE_TRANSCRIPTION", False)	×
30	if USE_TRANSCRIPTION:	×
31	TRANSCRIPTION_TYPE = getattr(settings_local, "TRANSCRIPTION_TYPE", "VOSK")	×
32	if TRANSCRIPTION_TYPE == "VOSK":	×
33	from vosk import Model, KaldiRecognizer	×
34	elif TRANSCRIPTION_TYPE == "STT":	×
35	from stt import Model	×
36	elif TRANSCRIPTION_TYPE == "WHISPER":	×
37	import whisper	×
38
39	TRANSCRIPTION_NORMALIZE = getattr(settings_local, "TRANSCRIPTION_NORMALIZE", False)	×
40	TRANSCRIPTION_NORMALIZE_TARGET_LEVEL = getattr(	×
41	settings_local, "TRANSCRIPTION_NORMALIZE_TARGET_LEVEL", -16.0
42	)
43
44	TRANSCRIPTION_AUDIO_SPLIT_TIME = getattr(	×
45	settings_local, "TRANSCRIPTION_AUDIO_SPLIT_TIME", 600
46	) # 10min
47	# time in sec for phrase length
48	TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH = getattr(	×
49	settings_local, "TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH", 2
50	)
51	TRANSCRIPTION_STT_SENTENCE_BLANK_SPLIT_TIME = getattr(	×
52	settings_local, "TRANSCRIPTION_STT_SENTENCE_BLANK_SPLIT_TIME", 0.5
53	)
54	log = logging.getLogger(__name__)	×
55
56
57	def get_model(lang):	×
58	"""Get model for STT or Vosk software to transcript audio."""
59	transript_model = Model(TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["model"])	×
60	if TRANSCRIPTION_TYPE == "STT":	×
61	if TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get("beam_width"):	×
62	transript_model.setBeamWidth(	×
63	TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["beam_width"]
64	)
65	if TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get("scorer"):	×
66	print(	×
67	"Loading scorer from files {}".format(
68	TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["scorer"]
69	),
70	file=sys.stderr,
71	)
72	scorer_load_start = timer()	×
73	transript_model.enableExternalScorer(	×
74	TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["scorer"]
75	)
76	scorer_load_end = timer() - scorer_load_start	×
77	print("Loaded scorer in {:.3}s.".format(scorer_load_end), file=sys.stderr)	×
78	if TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get(	×
79	"lm_alpha"
80	) and TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang].get("lm_beta"):
81	transript_model.setScorerAlphaBeta(	×
82	TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["lm_alpha"],
83	TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["lm_beta"],
84	)
85	return transript_model	×
86
87
88	def start_transcripting(mp3filepath, duration, lang):	×
89	"""
90	Start direct transcription.
91
92	Normalize the audio if set, get the model according to the lang and start transcript.
93	"""
94	if TRANSCRIPTION_NORMALIZE:	×
95	mp3filepath = normalize_mp3(mp3filepath)	×
96	if TRANSCRIPTION_TYPE == "WHISPER":	×
97	msg, webvtt, all_text = main_whisper_transcript(mp3filepath, lang)	×
98	else:
99	transript_model = get_model(lang)	×
100	msg, webvtt, all_text = start_main_transcript(	×
101	mp3filepath, duration, transript_model
102	)
103	if DEBUG:	×
104	print(msg)	×
105	print(webvtt)	×
106	print("\n%s\n" % all_text)	×
107
108	return msg, webvtt	×
109
110
111	def start_main_transcript(mp3filepath, duration, transript_model):	×
112	"""Call transcription depending software type."""
113	if TRANSCRIPTION_TYPE == "STT":	×
114	msg, webvtt, all_text = main_stt_transcript(	×
115	mp3filepath, duration, transript_model
116	)
117	elif TRANSCRIPTION_TYPE == "VOSK":	×
118	msg, webvtt, all_text = main_vosk_transcript(	×
119	mp3filepath, duration, transript_model
120	)
121	return msg, webvtt, all_text	×
122
123
124	def convert_samplerate(audio_path, desired_sample_rate, trim_start, duration):	×
125	"""Convert audio to subaudio and add good sample rate."""
126	sox_cmd = "sox {} --type raw --bits 16 --channels 1 --rate {} ".format(	×
127	quote(audio_path), desired_sample_rate
128	)
129	sox_cmd += "--encoding signed-integer --endian little --compression 0.0 "	×
130	sox_cmd += "--no-dither - trim {} {}".format(trim_start, duration)	×
131
132	try:	×
133	output = subprocess.check_output(shlex.split(sox_cmd), stderr=subprocess.PIPE)	×
134
135	except subprocess.CalledProcessError as e:	×
136	raise RuntimeError("SoX returned non-zero status: {}".format(e.stderr))	×
137	except OSError as e:	×
138	raise OSError(	×
139	e.errno,
140	"SoX not found, use {}hz files or install it: {}".format(
141	desired_sample_rate, e.strerror
142	),
143	)
144
145	return np.frombuffer(output, np.int16)	×
146
147
148	def normalize_mp3(mp3filepath):	×
149	"""Normalize the audio to good format and sound level."""
150	filename, file_extension = os.path.splitext(mp3filepath)	×
151	mp3normfile = "{}{}{}".format(filename, "_norm", file_extension)	×
152	normalize_cmd = "ffmpeg-normalize {} ".format(quote(mp3filepath))	×
153	normalize_cmd += "-c:a libmp3lame -b:a 192k --normalization-type ebu "	×
154	# normalize_cmd += \
155	# '--loudness-range-target 7.0 --true-peak 0.0 --offset 0.0 '
156	normalize_cmd += "--target-level {} -f -o {}".format(	×
157	TRANSCRIPTION_NORMALIZE_TARGET_LEVEL, quote(mp3normfile)
158	)
159	if DEBUG:	×
160	print(normalize_cmd)	×
161	try:	×
162	subprocess.check_output(shlex.split(normalize_cmd), stderr=subprocess.PIPE)	×
163	return mp3normfile	×
164	except subprocess.CalledProcessError as e:	×
165	log.error("ffmpeg-normalize returned non-zero status: {}".format(e.stderr))	×
166	return mp3filepath	×
167	except OSError as e:	×
168	log.error("ffmpeg-normalize not found {}".format(e.strerror))	×
169	return mp3filepath	×
170
171
172	# #################################
173	# TRANSCRIPT VIDEO: MAIN FUNCTION
174	# #################################
175
176
177	def convert_vosk_samplerate(audio_path, desired_sample_rate, trim_start, duration):	×
178	"""Convert audio to the good sample rate."""
179	sox_cmd = "sox {} --type raw --bits 16 --channels 1 --rate {} ".format(	×
180	quote(audio_path), desired_sample_rate
181	)
182	sox_cmd += "--encoding signed-integer --endian little --compression 0.0 "	×
183	sox_cmd += "--no-dither - trim {} {}".format(trim_start, duration)	×
184
185	try:	×
186	output = subprocess.Popen(shlex.split(sox_cmd), stdout=subprocess.PIPE)	×
187
188	except subprocess.CalledProcessError as e:	×
189	raise RuntimeError("SoX returned non-zero status: {}".format(e.stderr))	×
190	except OSError as e:	×
191	raise OSError(	×
192	e.errno,
193	"SoX not found, use {}hz files or install it: {}".format(
194	desired_sample_rate, e.strerror
195	),
196	)
197	return output	×
198
199
200	def get_word_result_from_data(results, audio, rec):	×
201	"""Get subsound from audio and add transcription to result parameter."""
202	while True:
203	data = audio.stdout.read(4000)	×
204	if len(data) == 0:	×
205	break	×
206	if rec.AcceptWaveform(data):	×
207	results.append(rec.Result())	×
208	results.append(rec.Result())	×
209
210
211	def words_to_vtt(	×
212	words,
213	start_trim,
214	duration,
215	is_first_caption,
216	text_caption,
217	start_caption,
218	last_word_added,
219	all_text,
220	webvtt,
221	):
222	"""Convert word and time to webvtt captions."""
223	for index, word in enumerate(words):	×
224	start_key = "start_time"	×
225	word_duration = word.get("duration", 0)	×
226	last_word = words[-1]	×
227	last_word_duration = last_word.get("duration", 0)	×
228	if TRANSCRIPTION_TYPE == "VOSK":	×
229	start_key = "start"	×
230	word_duration = word["end"] - word["start"]	×
231	last_word_duration = words[-1]["end"] - words[-1]["start"]	×
232	next_word = None	×
233	blank_duration = 0	×
234	if word != words[-1] and (index + 1) < len(words):	×
235	next_word = words[index + 1]	×
236	blank_duration = ((next_word[start_key]) - start_caption) - (	×
237	((word[start_key]) - start_caption) + word_duration
238	)
239	all_text += word["word"] + " "	×
240	# word: <class 'dict'> {'word': 'bonjour', 'start ':
241	# 0.58, 'duration': 7.34}
242	text_caption.append(word["word"])	×
243	if not (	×
244	(((word[start_key]) - start_caption) < TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH)
245	and (
246	next_word is not None
247	and (blank_duration < TRANSCRIPTION_STT_SENTENCE_BLANK_SPLIT_TIME)
248	)
249	):
250	# on créé le caption
251	if is_first_caption:	×
252	# A revoir, fusion de la nouvelle ligne avec
253	# l'ancienne...
254	is_first_caption = False	×
255	text_caption = get_text_caption(text_caption, last_word_added)	×
256
257	stop_caption = word[start_key] + word_duration	×
258
259	# on evite le chevauchement
260	change_previous_end_caption(webvtt, start_caption)	×
261
262	caption = Caption(	×
263	format_time_caption(start_caption),
264	format_time_caption(stop_caption),
265	" ".join(text_caption),
266	)
267
268	webvtt.captions.append(caption)	×
269	# on remet tout à zero pour la prochaine phrase
270	start_caption = word[start_key]	×
271	text_caption = []	×
272	last_word_added = word["word"]	×
273	if start_trim + TRANSCRIPTION_AUDIO_SPLIT_TIME > duration:	×
274	# on ajoute ici la dernière phrase de la vidéo
275	stop_caption = start_trim + words[-1][start_key] + last_word_duration	×
276	caption = Caption(	×
277	format_time_caption(start_caption),
278	format_time_caption(stop_caption),
279	" ".join(text_caption),
280	)
281	webvtt.captions.append(caption)	×
282	return all_text, webvtt	×
283
284
285	def main_vosk_transcript(norm_mp3_file, duration, transript_model):	×
286	"""Vosk transcription."""
287	msg = ""	×
288	inference_start = timer()	×
289	msg += "\nInference start %0.3fs." % inference_start	×
290	desired_sample_rate = 16000	×
291
292	rec = KaldiRecognizer(transript_model, desired_sample_rate)	×
293	rec.SetWords(True)	×
294
295	webvtt = WebVTT()	×
296	all_text = ""	×
297	for start_trim in range(0, duration, TRANSCRIPTION_AUDIO_SPLIT_TIME):	×
298	audio = convert_vosk_samplerate(	×
299	norm_mp3_file,
300	desired_sample_rate,
301	start_trim,
302	TRANSCRIPTION_AUDIO_SPLIT_TIME, # dur
303	)
304	msg += "\nRunning inference."	×
305	results = []	×
306	get_word_result_from_data(results, audio, rec)	×
307	for res in results:	×
308	words = json.loads(res).get("result")	×
309	text = json.loads(res).get("text")	×
310	if not words:	×
311	continue	×
312	start_caption = words[0]["start"]	×
313	stop_caption = words[-1]["end"]	×
314	caption = Caption(	×
315	format_time_caption(start_caption),
316	format_time_caption(stop_caption),
317	text,
318	)
319	webvtt.captions.append(caption)	×
320	"""	×
321	text_caption = []
322	is_first_caption = True
323	all_text, webvtt = words_to_vtt(
324	words,
325	start_trim,
326	duration,
327	is_first_caption,
328	text_caption,
329	start_caption,
330	last_word_added,
331	all_text,
332	webvtt,
333	)
334	"""
335	inference_end = timer() - inference_start	×
336
337	msg += "\nInference took %0.3fs." % inference_end	×
338	return msg, webvtt, all_text	×
339
340
341	def main_stt_transcript(norm_mp3_file, duration, transript_model):	×
342	"""STT transcription."""
343	msg = ""	×
344	inference_start = timer()	×
345	msg += "\nInference start %0.3fs." % inference_start	×
346	desired_sample_rate = transript_model.sampleRate()	×
347	webvtt = WebVTT()	×
348	last_word_added = ""	×
349	metadata = None	×
350	all_text = ""	×
351	for start_trim in range(0, duration, TRANSCRIPTION_AUDIO_SPLIT_TIME):	×
352	end_trim = (	×
353	duration
354	if start_trim + TRANSCRIPTION_AUDIO_SPLIT_TIME > duration
355	else (
356	start_trim
357	+ TRANSCRIPTION_AUDIO_SPLIT_TIME
358	+ TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH
359	)
360	)
361
362	dur = (	×
363	(TRANSCRIPTION_AUDIO_SPLIT_TIME + TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH)
364	if (
365	(
366	start_trim
367	+ TRANSCRIPTION_AUDIO_SPLIT_TIME
368	+ TRANSCRIPTION_STT_SENTENCE_MAX_LENGTH
369	)
370	< duration
371	)
372	else (duration - start_trim)
373	)
374
375	msg += "\ntake audio from %s to %s - %s" % (start_trim, end_trim, dur)	×
376
377	audio = convert_samplerate(norm_mp3_file, desired_sample_rate, start_trim, dur)	×
378	msg += "\nRunning inference."	×
379
380	metadata = transript_model.sttWithMetadata(audio)	×
381
382	for transcript in metadata.transcripts:	×
383	msg += "\nConfidence: %s" % transcript.confidence	×
384	words = words_from_candidate_transcript(transcript)	×
385	start_caption = start_trim + words[0]["start_time"]	×
386	text_caption = []	×
387	is_first_caption = True	×
388	all_text, webvtt = words_to_vtt(	×
389	words,
390	start_trim,
391	duration,
392	is_first_caption,
393	text_caption,
394	start_caption,
395	last_word_added,
396	all_text,
397	webvtt,
398	)
399	inference_end = timer() - inference_start	×
400
401	msg += "\nInference took %0.3fs." % inference_end	×
402	return msg, webvtt, all_text	×
403
404
405	def main_whisper_transcript(norm_mp3_file, lang):	×
406	"""Whisper transcription."""
407	msg = ""	×
408	all_text = ""	×
409	webvtt = WebVTT()	×
410	inference_start = timer()	×
411	msg += "\nInference start %0.3fs." % inference_start	×
412
413	model = whisper.load_model(	×
414	TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang]["model"],
415	download_root=TRANSCRIPTION_MODEL_PARAM[TRANSCRIPTION_TYPE][lang][
416	"download_root"
417	],
418	)
419
420	transcription = model.transcribe(norm_mp3_file, language=lang)	×
421	msg += "\nRunning inference."	×
422
423	for segment in transcription["segments"]:	×
424	caption = Caption(	×
425	format_time_caption(segment["start"]),
426	format_time_caption(segment["end"]),
427	segment["text"],
428	)
429	webvtt.captions.append(caption)	×
430
431	inference_end = timer() - inference_start	×
432	msg += "\nInference took %0.3fs." % inference_end	×
433	return msg, webvtt, all_text	×
434
435
436	def change_previous_end_caption(webvtt, start_caption):	×
437	"""Change the end time for caption."""
438	if len(webvtt.captions) > 0:	×
439	prev_end = dt.datetime.strptime(webvtt.captions[-1].end, "%H:%M:%S.%f")	×
440	td_prev_end = timedelta(	×
441	hours=prev_end.hour,
442	minutes=prev_end.minute,
443	seconds=prev_end.second,
444	microseconds=prev_end.microsecond,
445	).total_seconds()
446	if td_prev_end > start_caption:	×
447	webvtt.captions[-1].end = format_time_caption(start_caption)	×
448
449
450	def format_time_caption(time_caption):	×
451	"""Format time for webvtt caption."""
452	return (	×
453	dt.datetime.utcfromtimestamp(0) + timedelta(seconds=float(time_caption))
454	).strftime("%H:%M:%S.%f")[:-3]
455
456
457	def get_text_caption(text_caption, last_word_added):	×
458	"""Get the text for a caption."""
459	try:	×
460	first_index = text_caption.index(last_word_added)	×
461	return text_caption[first_index + 1 :]	×
462	except ValueError:	×
463	return text_caption	×
464
465
466	def words_from_candidate_transcript(metadata):	×
467	"""Get words list from transcription."""
468	word = ""	×
469	word_list = []	×
470	word_start_time = 0	×
471	# Loop through each character
472	for i, token in enumerate(metadata.tokens):	×
473	# Append character to word if it's not a space
474	if token.text != " ":	×
475	if len(word) == 0:	×
476	# Log the start time of the new word
477	word_start_time = token.start_time	×
478
479	word = word + token.text	×
480	# Word boundary is either a space or the last character in the array
481	if token.text == " " or i == len(metadata.tokens) - 1:	×
482	word_duration = token.start_time - word_start_time	×
483
484	if word_duration < 0:	×
485	word_duration = 0	×
486
487	each_word = dict()	×
488	each_word["word"] = word	×
489	each_word["start_time"] = round(word_start_time, 4)	×
490	each_word["duration"] = round(word_duration, 4)	×
491
492	word_list.append(each_word)	×
493	# Reset
494	word = ""	×
495	word_start_time = 0	×
496
497	return word_list	×

EsupPortail / Esup-Pod / 6654967569

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous