Adding support for diarization
Browse files- app.py +57 -4
- cli.py +18 -0
- config.json5 +11 -0
- src/config.py +11 -1
- src/diarization/diarization.py +188 -0
- src/diarization/requirements.txt +5 -0
- src/diarization/transcriptLoader.py +80 -0
- src/utils.py +21 -4
app.py
CHANGED
|
@@ -14,6 +14,7 @@ import numpy as np
|
|
| 14 |
import torch
|
| 15 |
|
| 16 |
from src.config import VAD_INITIAL_PROMPT_MODE_VALUES, ApplicationConfig, VadInitialPromptMode
|
|
|
|
| 17 |
from src.hooks.progressListener import ProgressListener
|
| 18 |
from src.hooks.subTaskProgressListener import SubTaskProgressListener
|
| 19 |
from src.hooks.whisperProgressHook import create_progress_listener_handle
|
|
@@ -73,6 +74,7 @@ class WhisperTranscriber:
|
|
| 73 |
self.deleteUploadedFiles = delete_uploaded_files
|
| 74 |
self.output_dir = output_dir
|
| 75 |
|
|
|
|
| 76 |
self.app_config = app_config
|
| 77 |
|
| 78 |
def set_parallel_devices(self, vad_parallel_devices: str):
|
|
@@ -89,19 +91,27 @@ class WhisperTranscriber:
|
|
| 89 |
# Entry function for the simple tab
|
| 90 |
def transcribe_webui_simple(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 91 |
vad, vadMergeWindow, vadMaxMergeSize,
|
| 92 |
-
word_timestamps: bool = False, highlight_words: bool = False
|
|
|
|
| 93 |
return self.transcribe_webui_simple_progress(modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 94 |
vad, vadMergeWindow, vadMaxMergeSize,
|
| 95 |
-
word_timestamps, highlight_words
|
|
|
|
| 96 |
|
| 97 |
# Entry function for the simple tab progress
|
| 98 |
def transcribe_webui_simple_progress(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 99 |
vad, vadMergeWindow, vadMaxMergeSize,
|
| 100 |
word_timestamps: bool = False, highlight_words: bool = False,
|
|
|
|
| 101 |
progress=gr.Progress()):
|
| 102 |
|
| 103 |
vadOptions = VadOptions(vad, vadMergeWindow, vadMaxMergeSize, self.app_config.vad_padding, self.app_config.vad_prompt_window, self.app_config.vad_initial_prompt_mode)
|
| 104 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
| 106 |
word_timestamps=word_timestamps, highlight_words=highlight_words, progress=progress)
|
| 107 |
|
|
@@ -112,14 +122,18 @@ class WhisperTranscriber:
|
|
| 112 |
word_timestamps: bool, highlight_words: bool, prepend_punctuations: str, append_punctuations: str,
|
| 113 |
initial_prompt: str, temperature: float, best_of: int, beam_size: int, patience: float, length_penalty: float, suppress_tokens: str,
|
| 114 |
condition_on_previous_text: bool, fp16: bool, temperature_increment_on_fallback: float,
|
| 115 |
-
compression_ratio_threshold: float, logprob_threshold: float, no_speech_threshold: float
|
|
|
|
|
|
|
| 116 |
|
| 117 |
return self.transcribe_webui_full_progress(modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 118 |
vad, vadMergeWindow, vadMaxMergeSize, vadPadding, vadPromptWindow, vadInitialPromptMode,
|
| 119 |
word_timestamps, highlight_words, prepend_punctuations, append_punctuations,
|
| 120 |
initial_prompt, temperature, best_of, beam_size, patience, length_penalty, suppress_tokens,
|
| 121 |
condition_on_previous_text, fp16, temperature_increment_on_fallback,
|
| 122 |
-
compression_ratio_threshold, logprob_threshold, no_speech_threshold
|
|
|
|
|
|
|
| 123 |
|
| 124 |
# Entry function for the full tab with progress
|
| 125 |
def transcribe_webui_full_progress(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
|
@@ -129,6 +143,8 @@ class WhisperTranscriber:
|
|
| 129 |
initial_prompt: str, temperature: float, best_of: int, beam_size: int, patience: float, length_penalty: float, suppress_tokens: str,
|
| 130 |
condition_on_previous_text: bool, fp16: bool, temperature_increment_on_fallback: float,
|
| 131 |
compression_ratio_threshold: float, logprob_threshold: float, no_speech_threshold: float,
|
|
|
|
|
|
|
| 132 |
progress=gr.Progress()):
|
| 133 |
|
| 134 |
# Handle temperature_increment_on_fallback
|
|
@@ -139,6 +155,13 @@ class WhisperTranscriber:
|
|
| 139 |
|
| 140 |
vadOptions = VadOptions(vad, vadMergeWindow, vadMaxMergeSize, vadPadding, vadPromptWindow, vadInitialPromptMode)
|
| 141 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 142 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
| 143 |
initial_prompt=initial_prompt, temperature=temperature, best_of=best_of, beam_size=beam_size, patience=patience, length_penalty=length_penalty, suppress_tokens=suppress_tokens,
|
| 144 |
condition_on_previous_text=condition_on_previous_text, fp16=fp16,
|
|
@@ -202,6 +225,19 @@ class WhisperTranscriber:
|
|
| 202 |
# Update progress
|
| 203 |
current_progress += source_audio_duration
|
| 204 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 205 |
source_download, source_text, source_vtt = self.write_result(result, filePrefix, outputDirectory, highlight_words)
|
| 206 |
|
| 207 |
if len(sources) > 1:
|
|
@@ -515,6 +551,17 @@ def create_ui(app_config: ApplicationConfig):
|
|
| 515 |
gr.Checkbox(label="Word Timestamps - Highlight Words", value=app_config.highlight_words),
|
| 516 |
]
|
| 517 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 518 |
is_queue_mode = app_config.queue_concurrency_count is not None and app_config.queue_concurrency_count > 0
|
| 519 |
|
| 520 |
simple_transcribe = gr.Interface(fn=ui.transcribe_webui_simple_progress if is_queue_mode else ui.transcribe_webui_simple,
|
|
@@ -522,6 +569,7 @@ def create_ui(app_config: ApplicationConfig):
|
|
| 522 |
*common_inputs(),
|
| 523 |
*common_vad_inputs(),
|
| 524 |
*common_word_timestamps_inputs(),
|
|
|
|
| 525 |
], outputs=[
|
| 526 |
gr.File(label="Download"),
|
| 527 |
gr.Text(label="Transcription"),
|
|
@@ -556,6 +604,11 @@ def create_ui(app_config: ApplicationConfig):
|
|
| 556 |
gr.Number(label="Compression ratio threshold", value=app_config.compression_ratio_threshold),
|
| 557 |
gr.Number(label="Logprob threshold", value=app_config.logprob_threshold),
|
| 558 |
gr.Number(label="No speech threshold", value=app_config.no_speech_threshold),
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 559 |
], outputs=[
|
| 560 |
gr.File(label="Download"),
|
| 561 |
gr.Text(label="Transcription"),
|
|
|
|
| 14 |
import torch
|
| 15 |
|
| 16 |
from src.config import VAD_INITIAL_PROMPT_MODE_VALUES, ApplicationConfig, VadInitialPromptMode
|
| 17 |
+
from src.diarization.diarization import Diarization
|
| 18 |
from src.hooks.progressListener import ProgressListener
|
| 19 |
from src.hooks.subTaskProgressListener import SubTaskProgressListener
|
| 20 |
from src.hooks.whisperProgressHook import create_progress_listener_handle
|
|
|
|
| 74 |
self.deleteUploadedFiles = delete_uploaded_files
|
| 75 |
self.output_dir = output_dir
|
| 76 |
|
| 77 |
+
self.diarization: Diarization = None
|
| 78 |
self.app_config = app_config
|
| 79 |
|
| 80 |
def set_parallel_devices(self, vad_parallel_devices: str):
|
|
|
|
| 91 |
# Entry function for the simple tab
|
| 92 |
def transcribe_webui_simple(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 93 |
vad, vadMergeWindow, vadMaxMergeSize,
|
| 94 |
+
word_timestamps: bool = False, highlight_words: bool = False,
|
| 95 |
+
diarization: bool = False, diarization_speakers: int = 2):
|
| 96 |
return self.transcribe_webui_simple_progress(modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 97 |
vad, vadMergeWindow, vadMaxMergeSize,
|
| 98 |
+
word_timestamps, highlight_words,
|
| 99 |
+
diarization, diarization_speakers)
|
| 100 |
|
| 101 |
# Entry function for the simple tab progress
|
| 102 |
def transcribe_webui_simple_progress(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 103 |
vad, vadMergeWindow, vadMaxMergeSize,
|
| 104 |
word_timestamps: bool = False, highlight_words: bool = False,
|
| 105 |
+
diarization: bool = False, diarization_speakers: int = 2,
|
| 106 |
progress=gr.Progress()):
|
| 107 |
|
| 108 |
vadOptions = VadOptions(vad, vadMergeWindow, vadMaxMergeSize, self.app_config.vad_padding, self.app_config.vad_prompt_window, self.app_config.vad_initial_prompt_mode)
|
| 109 |
|
| 110 |
+
if diarization:
|
| 111 |
+
self.diarization = Diarization(auth_token=self.app_config.auth_token, num_speakers=diarization_speakers)
|
| 112 |
+
else:
|
| 113 |
+
self.diarization = None
|
| 114 |
+
|
| 115 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
| 116 |
word_timestamps=word_timestamps, highlight_words=highlight_words, progress=progress)
|
| 117 |
|
|
|
|
| 122 |
word_timestamps: bool, highlight_words: bool, prepend_punctuations: str, append_punctuations: str,
|
| 123 |
initial_prompt: str, temperature: float, best_of: int, beam_size: int, patience: float, length_penalty: float, suppress_tokens: str,
|
| 124 |
condition_on_previous_text: bool, fp16: bool, temperature_increment_on_fallback: float,
|
| 125 |
+
compression_ratio_threshold: float, logprob_threshold: float, no_speech_threshold: float,
|
| 126 |
+
diarization: bool = False, diarization_speakers: int = 2,
|
| 127 |
+
diarization_min_speakers = 1, diarization_max_speakers = 5):
|
| 128 |
|
| 129 |
return self.transcribe_webui_full_progress(modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
| 130 |
vad, vadMergeWindow, vadMaxMergeSize, vadPadding, vadPromptWindow, vadInitialPromptMode,
|
| 131 |
word_timestamps, highlight_words, prepend_punctuations, append_punctuations,
|
| 132 |
initial_prompt, temperature, best_of, beam_size, patience, length_penalty, suppress_tokens,
|
| 133 |
condition_on_previous_text, fp16, temperature_increment_on_fallback,
|
| 134 |
+
compression_ratio_threshold, logprob_threshold, no_speech_threshold,
|
| 135 |
+
diarization, diarization_speakers,
|
| 136 |
+
diarization_min_speakers, diarization_max_speakers)
|
| 137 |
|
| 138 |
# Entry function for the full tab with progress
|
| 139 |
def transcribe_webui_full_progress(self, modelName, languageName, urlData, multipleFiles, microphoneData, task,
|
|
|
|
| 143 |
initial_prompt: str, temperature: float, best_of: int, beam_size: int, patience: float, length_penalty: float, suppress_tokens: str,
|
| 144 |
condition_on_previous_text: bool, fp16: bool, temperature_increment_on_fallback: float,
|
| 145 |
compression_ratio_threshold: float, logprob_threshold: float, no_speech_threshold: float,
|
| 146 |
+
diarization: bool = False, diarization_speakers: int = 2,
|
| 147 |
+
diarization_min_speakers = 1, diarization_max_speakers = 5,
|
| 148 |
progress=gr.Progress()):
|
| 149 |
|
| 150 |
# Handle temperature_increment_on_fallback
|
|
|
|
| 155 |
|
| 156 |
vadOptions = VadOptions(vad, vadMergeWindow, vadMaxMergeSize, vadPadding, vadPromptWindow, vadInitialPromptMode)
|
| 157 |
|
| 158 |
+
# Set diarization
|
| 159 |
+
if diarization:
|
| 160 |
+
self.diarization = Diarization(auth_token=self.app_config.auth_token, num_speakers=diarization_speakers,
|
| 161 |
+
min_speakers=diarization_min_speakers, max_speakers=diarization_max_speakers)
|
| 162 |
+
else:
|
| 163 |
+
self.diarization = None
|
| 164 |
+
|
| 165 |
return self.transcribe_webui(modelName, languageName, urlData, multipleFiles, microphoneData, task, vadOptions,
|
| 166 |
initial_prompt=initial_prompt, temperature=temperature, best_of=best_of, beam_size=beam_size, patience=patience, length_penalty=length_penalty, suppress_tokens=suppress_tokens,
|
| 167 |
condition_on_previous_text=condition_on_previous_text, fp16=fp16,
|
|
|
|
| 225 |
# Update progress
|
| 226 |
current_progress += source_audio_duration
|
| 227 |
|
| 228 |
+
# Diarization
|
| 229 |
+
if self.diarization:
|
| 230 |
+
print("Diarizing ", source.source_path)
|
| 231 |
+
diarization_result = list(self.diarization.run(source.source_path))
|
| 232 |
+
|
| 233 |
+
# Print result
|
| 234 |
+
print("Diarization result: ")
|
| 235 |
+
for entry in diarization_result:
|
| 236 |
+
print(f" start={entry.start:.1f}s stop={entry.end:.1f}s speaker_{entry.speaker}")
|
| 237 |
+
|
| 238 |
+
# Add speakers to result
|
| 239 |
+
result = self.diarization.mark_speakers(diarization_result, result)
|
| 240 |
+
|
| 241 |
source_download, source_text, source_vtt = self.write_result(result, filePrefix, outputDirectory, highlight_words)
|
| 242 |
|
| 243 |
if len(sources) > 1:
|
|
|
|
| 551 |
gr.Checkbox(label="Word Timestamps - Highlight Words", value=app_config.highlight_words),
|
| 552 |
]
|
| 553 |
|
| 554 |
+
has_diarization_libs = Diarization.has_libraries()
|
| 555 |
+
|
| 556 |
+
if not has_diarization_libs:
|
| 557 |
+
print("Diarization libraries not found - disabling diarization")
|
| 558 |
+
app_config.diarization = False
|
| 559 |
+
|
| 560 |
+
common_diarization_inputs = lambda : [
|
| 561 |
+
gr.Checkbox(label="Diarization", value=app_config.diarization, interactive=has_diarization_libs),
|
| 562 |
+
gr.Number(label="Diarization - Speakers", precision=0, value=app_config.diarization_speakers, interactive=has_diarization_libs)
|
| 563 |
+
]
|
| 564 |
+
|
| 565 |
is_queue_mode = app_config.queue_concurrency_count is not None and app_config.queue_concurrency_count > 0
|
| 566 |
|
| 567 |
simple_transcribe = gr.Interface(fn=ui.transcribe_webui_simple_progress if is_queue_mode else ui.transcribe_webui_simple,
|
|
|
|
| 569 |
*common_inputs(),
|
| 570 |
*common_vad_inputs(),
|
| 571 |
*common_word_timestamps_inputs(),
|
| 572 |
+
*common_diarization_inputs(),
|
| 573 |
], outputs=[
|
| 574 |
gr.File(label="Download"),
|
| 575 |
gr.Text(label="Transcription"),
|
|
|
|
| 604 |
gr.Number(label="Compression ratio threshold", value=app_config.compression_ratio_threshold),
|
| 605 |
gr.Number(label="Logprob threshold", value=app_config.logprob_threshold),
|
| 606 |
gr.Number(label="No speech threshold", value=app_config.no_speech_threshold),
|
| 607 |
+
|
| 608 |
+
*common_diarization_inputs(),
|
| 609 |
+
gr.Number(label="Diarization - Min Speakers", precision=0, value=app_config.diarization_min_speakers, interactive=has_diarization_libs),
|
| 610 |
+
gr.Number(label="Diarization - Max Speakers", precision=0, value=app_config.diarization_max_speakers, interactive=has_diarization_libs),
|
| 611 |
+
|
| 612 |
], outputs=[
|
| 613 |
gr.File(label="Download"),
|
| 614 |
gr.Text(label="Transcription"),
|
cli.py
CHANGED
|
@@ -8,6 +8,7 @@ import numpy as np
|
|
| 8 |
import torch
|
| 9 |
from app import VadOptions, WhisperTranscriber
|
| 10 |
from src.config import VAD_INITIAL_PROMPT_MODE_VALUES, ApplicationConfig, VadInitialPromptMode
|
|
|
|
| 11 |
from src.download import download_url
|
| 12 |
from src.languages import get_language_names
|
| 13 |
|
|
@@ -106,6 +107,14 @@ def cli():
|
|
| 106 |
parser.add_argument("--threads", type=optional_int, default=0,
|
| 107 |
help="number of threads used by torch for CPU inference; supercedes MKL_NUM_THREADS/OMP_NUM_THREADS")
|
| 108 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 109 |
args = parser.parse_args().__dict__
|
| 110 |
model_name: str = args.pop("model")
|
| 111 |
model_dir: str = args.pop("model_dir")
|
|
@@ -142,10 +151,19 @@ def cli():
|
|
| 142 |
compute_type = args.pop("compute_type")
|
| 143 |
highlight_words = args.pop("highlight_words")
|
| 144 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 145 |
transcriber = WhisperTranscriber(delete_uploaded_files=False, vad_cpu_cores=vad_cpu_cores, app_config=app_config)
|
| 146 |
transcriber.set_parallel_devices(args.pop("vad_parallel_devices"))
|
| 147 |
transcriber.set_auto_parallel(auto_parallel)
|
| 148 |
|
|
|
|
|
|
|
|
|
|
| 149 |
model = create_whisper_container(whisper_implementation=whisper_implementation, model_name=model_name,
|
| 150 |
device=device, compute_type=compute_type, download_root=model_dir, models=app_config.models)
|
| 151 |
|
|
|
|
| 8 |
import torch
|
| 9 |
from app import VadOptions, WhisperTranscriber
|
| 10 |
from src.config import VAD_INITIAL_PROMPT_MODE_VALUES, ApplicationConfig, VadInitialPromptMode
|
| 11 |
+
from src.diarization.diarization import Diarization
|
| 12 |
from src.download import download_url
|
| 13 |
from src.languages import get_language_names
|
| 14 |
|
|
|
|
| 107 |
parser.add_argument("--threads", type=optional_int, default=0,
|
| 108 |
help="number of threads used by torch for CPU inference; supercedes MKL_NUM_THREADS/OMP_NUM_THREADS")
|
| 109 |
|
| 110 |
+
# Diarization
|
| 111 |
+
parser.add_argument('--auth_token', type=str, default=None, help='HuggingFace API Token (optional)')
|
| 112 |
+
parser.add_argument("--diarization", type=str2bool, default=app_config.diarization, \
|
| 113 |
+
help="whether to perform speaker diarization")
|
| 114 |
+
parser.add_argument("--num_speakers", type=int, default=None, help="Number of speakers")
|
| 115 |
+
parser.add_argument("--min_speakers", type=int, default=None, help="Minimum number of speakers")
|
| 116 |
+
parser.add_argument("--max_speakers", type=int, default=None, help="Maximum number of speakers")
|
| 117 |
+
|
| 118 |
args = parser.parse_args().__dict__
|
| 119 |
model_name: str = args.pop("model")
|
| 120 |
model_dir: str = args.pop("model_dir")
|
|
|
|
| 151 |
compute_type = args.pop("compute_type")
|
| 152 |
highlight_words = args.pop("highlight_words")
|
| 153 |
|
| 154 |
+
diarization = args.pop("diarization")
|
| 155 |
+
auth_token = args.pop("auth_token")
|
| 156 |
+
num_speakers = args.pop("num_speakers")
|
| 157 |
+
min_speakers = args.pop("min_speakers")
|
| 158 |
+
max_speakers = args.pop("max_speakers")
|
| 159 |
+
|
| 160 |
transcriber = WhisperTranscriber(delete_uploaded_files=False, vad_cpu_cores=vad_cpu_cores, app_config=app_config)
|
| 161 |
transcriber.set_parallel_devices(args.pop("vad_parallel_devices"))
|
| 162 |
transcriber.set_auto_parallel(auto_parallel)
|
| 163 |
|
| 164 |
+
if diarization:
|
| 165 |
+
transcriber.set_diarization(Diarization(auth_token=auth_token, num_speakers=num_speakers, min_speakers=min_speakers, max_speakers=max_speakers))
|
| 166 |
+
|
| 167 |
model = create_whisper_container(whisper_implementation=whisper_implementation, model_name=model_name,
|
| 168 |
device=device, compute_type=compute_type, download_root=model_dir, models=app_config.models)
|
| 169 |
|
config.json5
CHANGED
|
@@ -138,4 +138,15 @@
|
|
| 138 |
"append_punctuations": "\"\'.。,,!!??::”)]}、",
|
| 139 |
// (requires --word_timestamps True) underline each word as it is spoken in srt and vtt
|
| 140 |
"highlight_words": false,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 141 |
}
|
|
|
|
| 138 |
"append_punctuations": "\"\'.。,,!!??::”)]}、",
|
| 139 |
// (requires --word_timestamps True) underline each word as it is spoken in srt and vtt
|
| 140 |
"highlight_words": false,
|
| 141 |
+
|
| 142 |
+
// Diarization settings
|
| 143 |
+
"auth_token": null,
|
| 144 |
+
// Whether to perform speaker diarization
|
| 145 |
+
"diarization": false,
|
| 146 |
+
// The number of speakers to detect
|
| 147 |
+
"diarization_speakers": 2,
|
| 148 |
+
// The minimum number of speakers to detect
|
| 149 |
+
"diarization_min_speakers": 1,
|
| 150 |
+
// The maximum number of speakers to detect
|
| 151 |
+
"diarization_max_speakers": 5,
|
| 152 |
}
|
src/config.py
CHANGED
|
@@ -69,7 +69,10 @@ class ApplicationConfig:
|
|
| 69 |
# Word timestamp settings
|
| 70 |
word_timestamps: bool = False, prepend_punctuations: str = "\"\'“¿([{-",
|
| 71 |
append_punctuations: str = "\"\'.。,,!!??::”)]}、",
|
| 72 |
-
highlight_words: bool = False
|
|
|
|
|
|
|
|
|
|
| 73 |
|
| 74 |
self.models = models
|
| 75 |
|
|
@@ -121,6 +124,13 @@ class ApplicationConfig:
|
|
| 121 |
self.append_punctuations = append_punctuations
|
| 122 |
self.highlight_words = highlight_words
|
| 123 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 124 |
def get_model_names(self):
|
| 125 |
return [ x.name for x in self.models ]
|
| 126 |
|
|
|
|
| 69 |
# Word timestamp settings
|
| 70 |
word_timestamps: bool = False, prepend_punctuations: str = "\"\'“¿([{-",
|
| 71 |
append_punctuations: str = "\"\'.。,,!!??::”)]}、",
|
| 72 |
+
highlight_words: bool = False,
|
| 73 |
+
# Diarization
|
| 74 |
+
auth_token: str = None, diarization: bool = False, diarization_speakers: int = 2,
|
| 75 |
+
diarization_min_speakers: int = 1, diarization_max_speakers: int = 5):
|
| 76 |
|
| 77 |
self.models = models
|
| 78 |
|
|
|
|
| 124 |
self.append_punctuations = append_punctuations
|
| 125 |
self.highlight_words = highlight_words
|
| 126 |
|
| 127 |
+
# Diarization settings
|
| 128 |
+
self.auth_token = auth_token
|
| 129 |
+
self.diarization = diarization
|
| 130 |
+
self.diarization_speakers = diarization_speakers
|
| 131 |
+
self.diarization_min_speakers = diarization_min_speakers
|
| 132 |
+
self.diarization_max_speakers = diarization_max_speakers
|
| 133 |
+
|
| 134 |
def get_model_names(self):
|
| 135 |
return [ x.name for x in self.models ]
|
| 136 |
|
src/diarization/diarization.py
ADDED
|
@@ -0,0 +1,188 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import argparse
|
| 2 |
+
import json
|
| 3 |
+
import os
|
| 4 |
+
from pathlib import Path
|
| 5 |
+
import tempfile
|
| 6 |
+
from typing import TYPE_CHECKING, List
|
| 7 |
+
import torch
|
| 8 |
+
|
| 9 |
+
import ffmpeg
|
| 10 |
+
|
| 11 |
+
from src.diarization.transcriptLoader import load_transcript
|
| 12 |
+
from src.utils import write_srt
|
| 13 |
+
|
| 14 |
+
class DiarizationEntry:
|
| 15 |
+
def __init__(self, start, end, speaker):
|
| 16 |
+
self.start = start
|
| 17 |
+
self.end = end
|
| 18 |
+
self.speaker = speaker
|
| 19 |
+
|
| 20 |
+
def __repr__(self):
|
| 21 |
+
return f"<DiarizationEntry start={self.start} end={self.end} speaker={self.speaker}>"
|
| 22 |
+
|
| 23 |
+
def toJson(self):
|
| 24 |
+
return {
|
| 25 |
+
"start": self.start,
|
| 26 |
+
"end": self.end,
|
| 27 |
+
"speaker": self.speaker
|
| 28 |
+
}
|
| 29 |
+
|
| 30 |
+
class Diarization:
|
| 31 |
+
def __init__(self, auth_token=None, **kwargs):
|
| 32 |
+
if auth_token is None:
|
| 33 |
+
auth_token = os.environ.get("HK_ACCESS_TOKEN")
|
| 34 |
+
if auth_token is None:
|
| 35 |
+
raise ValueError("No HuggingFace API Token provided - please use the --auth_token argument or set the HK_ACCESS_TOKEN environment variable")
|
| 36 |
+
|
| 37 |
+
self.auth_token = auth_token
|
| 38 |
+
self.initialized = False
|
| 39 |
+
self.pipeline = None
|
| 40 |
+
self.pipeline_kwargs = kwargs
|
| 41 |
+
|
| 42 |
+
@staticmethod
|
| 43 |
+
def has_libraries():
|
| 44 |
+
try:
|
| 45 |
+
import pyannote.audio
|
| 46 |
+
import intervaltree
|
| 47 |
+
return True
|
| 48 |
+
except ImportError:
|
| 49 |
+
return False
|
| 50 |
+
|
| 51 |
+
def initialize(self):
|
| 52 |
+
if self.initialized:
|
| 53 |
+
return
|
| 54 |
+
from pyannote.audio import Pipeline
|
| 55 |
+
|
| 56 |
+
self.pipeline = Pipeline.from_pretrained("pyannote/[email protected]", use_auth_token=self.auth_token)
|
| 57 |
+
|
| 58 |
+
# Load GPU mode if available
|
| 59 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
| 60 |
+
if device == "cuda":
|
| 61 |
+
print("Diarization - using GPU")
|
| 62 |
+
self.pipeline = self.pipeline.to(torch.device(0))
|
| 63 |
+
else:
|
| 64 |
+
print("Diarization - using CPU")
|
| 65 |
+
|
| 66 |
+
def run(self, audio_file):
|
| 67 |
+
self.initialize()
|
| 68 |
+
audio_file_obj = Path(audio_file)
|
| 69 |
+
|
| 70 |
+
# Supported file types in soundfile is WAV, FLAC, OGG and MAT
|
| 71 |
+
if audio_file_obj.suffix in [".wav", ".flac", ".ogg", ".mat"]:
|
| 72 |
+
target_file = audio_file
|
| 73 |
+
else:
|
| 74 |
+
# Create temp WAV file
|
| 75 |
+
target_file = tempfile.mktemp(prefix="diarization_", suffix=".wav")
|
| 76 |
+
try:
|
| 77 |
+
ffmpeg.input(audio_file).output(target_file, ac=1).run()
|
| 78 |
+
except ffmpeg.Error as e:
|
| 79 |
+
print(f"Error occurred during audio conversion: {e.stderr}")
|
| 80 |
+
|
| 81 |
+
diarization = self.pipeline(target_file, **self.pipeline_kwargs)
|
| 82 |
+
|
| 83 |
+
if target_file != audio_file:
|
| 84 |
+
# Delete temp file
|
| 85 |
+
os.remove(target_file)
|
| 86 |
+
|
| 87 |
+
# Yield result
|
| 88 |
+
for turn, _, speaker in diarization.itertracks(yield_label=True):
|
| 89 |
+
yield DiarizationEntry(turn.start, turn.end, speaker)
|
| 90 |
+
|
| 91 |
+
def mark_speakers(self, diarization_result: List[DiarizationEntry], whisper_result: dict):
|
| 92 |
+
from intervaltree import IntervalTree
|
| 93 |
+
result = whisper_result.copy()
|
| 94 |
+
|
| 95 |
+
# Create an interval tree from the diarization results
|
| 96 |
+
tree = IntervalTree()
|
| 97 |
+
for entry in diarization_result:
|
| 98 |
+
tree[entry.start:entry.end] = entry
|
| 99 |
+
|
| 100 |
+
# Iterate through each segment in the Whisper JSON
|
| 101 |
+
for segment in result["segments"]:
|
| 102 |
+
segment_start = segment["start"]
|
| 103 |
+
segment_end = segment["end"]
|
| 104 |
+
|
| 105 |
+
# Find overlapping speakers using the interval tree
|
| 106 |
+
overlapping_speakers = tree[segment_start:segment_end]
|
| 107 |
+
|
| 108 |
+
# If no speakers overlap with this segment, skip it
|
| 109 |
+
if not overlapping_speakers:
|
| 110 |
+
continue
|
| 111 |
+
|
| 112 |
+
# If multiple speakers overlap with this segment, choose the one with the longest duration
|
| 113 |
+
longest_speaker = None
|
| 114 |
+
longest_duration = 0
|
| 115 |
+
|
| 116 |
+
for speaker_interval in overlapping_speakers:
|
| 117 |
+
overlap_start = max(speaker_interval.begin, segment_start)
|
| 118 |
+
overlap_end = min(speaker_interval.end, segment_end)
|
| 119 |
+
overlap_duration = overlap_end - overlap_start
|
| 120 |
+
|
| 121 |
+
if overlap_duration > longest_duration:
|
| 122 |
+
longest_speaker = speaker_interval.data.speaker
|
| 123 |
+
longest_duration = overlap_duration
|
| 124 |
+
|
| 125 |
+
# Add speakers
|
| 126 |
+
segment["longest_speaker"] = longest_speaker
|
| 127 |
+
segment["speakers"] = list([speaker_interval.data.toJson() for speaker_interval in overlapping_speakers])
|
| 128 |
+
|
| 129 |
+
# The write_srt will use the longest_speaker if it exist, and add it to the text field
|
| 130 |
+
|
| 131 |
+
return result
|
| 132 |
+
|
| 133 |
+
def _write_file(input_file: str, output_path: str, output_extension: str, file_writer: lambda f: None):
|
| 134 |
+
if input_file is None:
|
| 135 |
+
raise ValueError("input_file is required")
|
| 136 |
+
if file_writer is None:
|
| 137 |
+
raise ValueError("file_writer is required")
|
| 138 |
+
|
| 139 |
+
# Write file
|
| 140 |
+
if output_path is None:
|
| 141 |
+
effective_path = os.path.splitext(input_file)[0] + "_output" + output_extension
|
| 142 |
+
else:
|
| 143 |
+
effective_path = output_path
|
| 144 |
+
|
| 145 |
+
with open(effective_path, 'w+', encoding="utf-8") as f:
|
| 146 |
+
file_writer(f)
|
| 147 |
+
|
| 148 |
+
print(f"Output saved to {effective_path}")
|
| 149 |
+
|
| 150 |
+
def main():
|
| 151 |
+
parser = argparse.ArgumentParser(description='Add speakers to a SRT file or Whisper JSON file using pyannote/speaker-diarization.')
|
| 152 |
+
parser.add_argument('audio_file', type=str, help='Input audio file')
|
| 153 |
+
parser.add_argument('whisper_file', type=str, help='Input Whisper JSON/SRT file')
|
| 154 |
+
parser.add_argument('--output_json_file', type=str, default=None, help='Output JSON file (optional)')
|
| 155 |
+
parser.add_argument('--output_srt_file', type=str, default=None, help='Output SRT file (optional)')
|
| 156 |
+
parser.add_argument('--auth_token', type=str, default=None, help='HuggingFace API Token (optional)')
|
| 157 |
+
parser.add_argument("--max_line_width", type=int, default=40, help="Maximum line width for SRT file (default: 40)")
|
| 158 |
+
parser.add_argument("--num_speakers", type=int, default=None, help="Number of speakers")
|
| 159 |
+
parser.add_argument("--min_speakers", type=int, default=None, help="Minimum number of speakers")
|
| 160 |
+
parser.add_argument("--max_speakers", type=int, default=None, help="Maximum number of speakers")
|
| 161 |
+
|
| 162 |
+
args = parser.parse_args()
|
| 163 |
+
|
| 164 |
+
print("\nReading whisper JSON from " + args.whisper_file)
|
| 165 |
+
|
| 166 |
+
# Read whisper JSON or SRT file
|
| 167 |
+
whisper_result = load_transcript(args.whisper_file)
|
| 168 |
+
|
| 169 |
+
diarization = Diarization(auth_token=args.auth_token, num_speakers=args.num_speakers, min_speakers=args.min_speakers, max_speakers=args.max_speakers)
|
| 170 |
+
diarization_result = list(diarization.run(args.audio_file))
|
| 171 |
+
|
| 172 |
+
# Print result
|
| 173 |
+
print("Diarization result:")
|
| 174 |
+
for entry in diarization_result:
|
| 175 |
+
print(f" start={entry.start:.1f}s stop={entry.end:.1f}s speaker_{entry.speaker}")
|
| 176 |
+
|
| 177 |
+
marked_whisper_result = diarization.mark_speakers(diarization_result, whisper_result)
|
| 178 |
+
|
| 179 |
+
# Write output JSON to file
|
| 180 |
+
_write_file(args.whisper_file, args.output_json_file, ".json",
|
| 181 |
+
lambda f: json.dump(marked_whisper_result, f, indent=4, ensure_ascii=False))
|
| 182 |
+
|
| 183 |
+
# Write SRT
|
| 184 |
+
_write_file(args.whisper_file, args.output_srt_file, ".srt",
|
| 185 |
+
lambda f: write_srt(marked_whisper_result["segments"], f, maxLineWidth=args.max_line_width))
|
| 186 |
+
|
| 187 |
+
if __name__ == "__main__":
|
| 188 |
+
main()
|
src/diarization/requirements.txt
ADDED
|
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
intervaltree
|
| 2 |
+
srt
|
| 3 |
+
torch
|
| 4 |
+
ffmpeg-python==0.2.0
|
| 5 |
+
https://github.com/pyannote/pyannote-audio/archive/refs/heads/develop.zip
|
src/diarization/transcriptLoader.py
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
import json
|
| 3 |
+
from pathlib import Path
|
| 4 |
+
|
| 5 |
+
def load_transcript_json(transcript_file: str):
|
| 6 |
+
"""
|
| 7 |
+
Parse a Whisper JSON file into a Whisper JSON object
|
| 8 |
+
|
| 9 |
+
# Parameters:
|
| 10 |
+
transcript_file (str): Path to the Whisper JSON file
|
| 11 |
+
"""
|
| 12 |
+
with open(transcript_file, "r", encoding="utf-8") as f:
|
| 13 |
+
whisper_result = json.load(f)
|
| 14 |
+
|
| 15 |
+
# Format of Whisper JSON file:
|
| 16 |
+
# {
|
| 17 |
+
# "text": " And so my fellow Americans, ask not what your country can do for you, ask what you can do for your country.",
|
| 18 |
+
# "segments": [
|
| 19 |
+
# {
|
| 20 |
+
# "text": " And so my fellow Americans, ask not what your country can do for you, ask what you can do for your country.",
|
| 21 |
+
# "start": 0.0,
|
| 22 |
+
# "end": 10.36,
|
| 23 |
+
# "words": [
|
| 24 |
+
# {
|
| 25 |
+
# "start": 0.0,
|
| 26 |
+
# "end": 0.56,
|
| 27 |
+
# "word": " And",
|
| 28 |
+
# "probability": 0.61767578125
|
| 29 |
+
# },
|
| 30 |
+
# {
|
| 31 |
+
# "start": 0.56,
|
| 32 |
+
# "end": 0.88,
|
| 33 |
+
# "word": " so",
|
| 34 |
+
# "probability": 0.9033203125
|
| 35 |
+
# },
|
| 36 |
+
# etc.
|
| 37 |
+
|
| 38 |
+
return whisper_result
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
def load_transcript_srt(subtitle_file: str):
|
| 42 |
+
import srt
|
| 43 |
+
|
| 44 |
+
"""
|
| 45 |
+
Parse a SRT file into a Whisper JSON object
|
| 46 |
+
|
| 47 |
+
# Parameters:
|
| 48 |
+
subtitle_file (str): Path to the SRT file
|
| 49 |
+
"""
|
| 50 |
+
with open(subtitle_file, "r", encoding="utf-8") as f:
|
| 51 |
+
subs = srt.parse(f)
|
| 52 |
+
|
| 53 |
+
whisper_result = {
|
| 54 |
+
"text": "",
|
| 55 |
+
"segments": []
|
| 56 |
+
}
|
| 57 |
+
|
| 58 |
+
for sub in subs:
|
| 59 |
+
# Subtitle(index=1, start=datetime.timedelta(seconds=33, microseconds=843000), end=datetime.timedelta(seconds=38, microseconds=97000), content='地球上只有3%的水是淡水', proprietary='')
|
| 60 |
+
segment = {
|
| 61 |
+
"text": sub.content,
|
| 62 |
+
"start": sub.start.total_seconds(),
|
| 63 |
+
"end": sub.end.total_seconds(),
|
| 64 |
+
"words": []
|
| 65 |
+
}
|
| 66 |
+
whisper_result["segments"].append(segment)
|
| 67 |
+
whisper_result["text"] += sub.content
|
| 68 |
+
|
| 69 |
+
return whisper_result
|
| 70 |
+
|
| 71 |
+
def load_transcript(file: str):
|
| 72 |
+
# Determine file type
|
| 73 |
+
file_extension = Path(file).suffix.lower()
|
| 74 |
+
|
| 75 |
+
if file_extension == ".json":
|
| 76 |
+
return load_transcript_json(file)
|
| 77 |
+
elif file_extension == ".srt":
|
| 78 |
+
return load_transcript_srt(file)
|
| 79 |
+
else:
|
| 80 |
+
raise ValueError(f"Unsupported file type: {file_extension}")
|
src/utils.py
CHANGED
|
@@ -102,17 +102,26 @@ def write_srt(transcript: Iterator[dict], file: TextIO,
|
|
| 102 |
|
| 103 |
def __subtitle_preprocessor_iterator(transcript: Iterator[dict], maxLineWidth: int = None, highlight_words: bool = False):
|
| 104 |
for segment in transcript:
|
| 105 |
-
words = segment.get('words', [])
|
|
|
|
|
|
|
|
|
|
| 106 |
|
| 107 |
if len(words) == 0:
|
| 108 |
# Yield the segment as-is or processed
|
| 109 |
-
if maxLineWidth is None or maxLineWidth < 0:
|
| 110 |
yield segment
|
| 111 |
else:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 112 |
yield {
|
| 113 |
'start': segment['start'],
|
| 114 |
'end': segment['end'],
|
| 115 |
-
'text': process_text(
|
| 116 |
}
|
| 117 |
# We are done
|
| 118 |
continue
|
|
@@ -120,9 +129,17 @@ def __subtitle_preprocessor_iterator(transcript: Iterator[dict], maxLineWidth: i
|
|
| 120 |
subtitle_start = segment['start']
|
| 121 |
subtitle_end = segment['end']
|
| 122 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 123 |
text_words = [ this_word["word"] for this_word in words ]
|
| 124 |
subtitle_text = __join_words(text_words, maxLineWidth)
|
| 125 |
-
|
| 126 |
# Iterate over the words in the segment
|
| 127 |
if highlight_words:
|
| 128 |
last = subtitle_start
|
|
|
|
| 102 |
|
| 103 |
def __subtitle_preprocessor_iterator(transcript: Iterator[dict], maxLineWidth: int = None, highlight_words: bool = False):
|
| 104 |
for segment in transcript:
|
| 105 |
+
words: list = segment.get('words', [])
|
| 106 |
+
|
| 107 |
+
# Append longest speaker ID if available
|
| 108 |
+
segment_longest_speaker = segment.get('longest_speaker', None)
|
| 109 |
|
| 110 |
if len(words) == 0:
|
| 111 |
# Yield the segment as-is or processed
|
| 112 |
+
if (maxLineWidth is None or maxLineWidth < 0) and segment_longest_speaker is None:
|
| 113 |
yield segment
|
| 114 |
else:
|
| 115 |
+
text = segment['text'].strip()
|
| 116 |
+
|
| 117 |
+
# Prepend the longest speaker ID if available
|
| 118 |
+
if segment_longest_speaker is not None:
|
| 119 |
+
text = f"({segment_longest_speaker}) {text}"
|
| 120 |
+
|
| 121 |
yield {
|
| 122 |
'start': segment['start'],
|
| 123 |
'end': segment['end'],
|
| 124 |
+
'text': process_text(text, maxLineWidth)
|
| 125 |
}
|
| 126 |
# We are done
|
| 127 |
continue
|
|
|
|
| 129 |
subtitle_start = segment['start']
|
| 130 |
subtitle_end = segment['end']
|
| 131 |
|
| 132 |
+
if segment_longest_speaker is not None:
|
| 133 |
+
# Add the beginning
|
| 134 |
+
words.insert(0, {
|
| 135 |
+
'start': subtitle_start,
|
| 136 |
+
'end': subtitle_start,
|
| 137 |
+
'word': f"({segment_longest_speaker})"
|
| 138 |
+
})
|
| 139 |
+
|
| 140 |
text_words = [ this_word["word"] for this_word in words ]
|
| 141 |
subtitle_text = __join_words(text_words, maxLineWidth)
|
| 142 |
+
|
| 143 |
# Iterate over the words in the segment
|
| 144 |
if highlight_words:
|
| 145 |
last = subtitle_start
|