o 9wi)!@sZddlZddlZddlZddlZddlmZmZmZm Z m Z ddZ e dkr+e dSdS)N) LANGUAGESTO_LANGUAGE_CODEoptional_float optional_intstr2boolc Cstjtjd}|jddtdd|jdddd |jd td d d |jdtddd |jdtjr4dnddd |jddt dd|jddt dd|jddtgddd|jd d!td"d#d |jd$d%td&gd'd(d)|jd*td+d,d |jd-td.d.d/gd0d)|jd1tdt t t d2d3t Dd4d)|jd5dd6d |jd7d8gd9d:d;|jdd?|jd@d=dAd?|jdBtdCdCdDgdEd)|jdFtdGdHd |jdItdJdKd |jdLt dMdNd |jdOd=dPd?|jdQdt dRd|jdSdt dTd|jdUdVtdWd|jdXd=dYd?|jdZtdd[d |jd\td]d^d |jd_td]d`d |jdatdbdcd |jddtdbded |jdftdgdhd |jdid=djd?|jdktddld |jdmtd dnd |jdotd+dpd |jdqtdrdsd |jdttdudvd |jdwtdxdyd |jdztd{d|d |jd}tdd~d |jdtddd |jdtd dd |jdtdddgd~d)|jdtddd |jdtddd |jdtd dd |jddddtjddd|jddddtdtddd|j}ddlm}|||dS)N)formatter_classaudio+zaudio file(s) to transcribe)nargstypehelpz--modelsmallz name of the Whisper model to use)defaultr z--model_cache_onlyFzZIf True, will not attempt to download models, instead using cached models from --model_dir)r rr z --model_dirz>the path to save model files; uses ~/.cache/whisper by defaultz--devicecudacpuz#device to use for PyTorch inferencez--device_indexrz/device index to use for FasterWhisper inference)rr r z --batch_sizez&the preferred batch size for inferencez--compute_typefloat16)rfloat32int8zcompute type for computation)rr choicesr z --output_dirz-o.zdirectory to save the outputsz--output_formatz-fall)rsrtvtttxttsvjsonaudzSformat of the output file; if not specified, all available formats will be produced)r rrr z --verboseTz4whether to print out the progress and debug messagesz--task transcribe translatezawhether to perform X->X speech recognition ('transcribe') or X->English translation ('translate')z --languagecSsg|]}|qS)title).0kr r N/home/ubuntu/sommelier/.venv/lib/python3.10/site-packages/whisperx/__main__.py szcli..zHlanguage spoken in the audio, specify None to perform language detectionz --align_modelz/Name of phoneme-level ASR model to do alignmentz--interpolate_methodnearest)r&linearignorezaFor word .srt, method to assign timestamps to non-aligned words, or merge them into neighbouring.)rrr z --no_align store_truez Do not perform phoneme alignment)actionr z--return_char_alignmentsz9Return character-level alignments in the output json filez --vad_methodpyannotesilerozVAD method to be usedz --vad_onsetg?zYOnset threshold for VAD (see pyannote.audio), reduce this if speech is not being detectedz --vad_offsetgZd;?z[Offset threshold for VAD (see pyannote.audio), reduce this if speech is not being detected.z --chunk_sizezYChunk size for merging VAD segments. Default is 30, reduce this if the chunk is too long.z --diarizez?Apply diarization to assign speaker labels to each segment/wordz--min_speakersz+Minimum number of speakers to in audio filez--max_speakersz+Maximum number of speakers to in audio filez--diarize_modelz pyannote/speaker-diarization-3.1z,Name of the speaker diarization model to usez--speaker_embeddingszEInclude speaker embeddings in JSON output (only works with --diarize)z --temperatureztemperature to use for samplingz --best_ofz token is higher than this value AND the decoding has failed due to `logprob_threshold`, consider the segment as silencez--max_line_widthzb(not possible with --no_align) the maximum number of characters in a line before breaking the linez--max_line_countzG(not possible with --no_align) the maximum number of lines in a segmentz--highlight_wordszQ(not possible with --no_align) underline each word as it is spoken in srt and vttz--segment_resolutionsentencechunkz --threadsz]number of threads used by torch for CPU inference; supercedes MKL_NUM_THREADS/OMP_NUM_THREADSz --hf_tokenz9Hugging Face Access Token to access PyAnnote gated modelsz--print_progresszFif True, progress will be printed in transcribe() and align() methods.z --versionz-Vversionz %(prog)s whisperxz*Show whisperx version information and exit)r*r1r z--python-versionz-PzPython z ()z(Show python version information and exit)transcribe_task)argparseArgumentParserArgumentDefaultsHelpFormatter add_argumentstrrtorchr is_availableintsortedrkeysrfloatrr importlibmetadatar1platformpython_versionpython_implementation parse_args__dict__whisperx.transcriber4)parserargsr4r r r$cli sj2"*  rJ__main__) r5importlib.metadatar@rBr:whisperx.utilsrrrrrrJ__name__r r r r$sM