U ,-e!@sHdZddlmZmZddlZddlmZddlm Z GdddeZ dS) z) Text/audio processor class for MusicGen )ListOptionalN)ProcessorMixin)to_numpycsbeZdZdZdZdZfddZddd Zd d Zd d Z ddZ de e e jdddZZS)MusicgenProcessoral Constructs a MusicGen processor which wraps an EnCodec feature extractor and a T5 tokenizer into a single processor class. [`MusicgenProcessor`] offers all the functionalities of [`EncodecFeatureExtractor`] and [`TTokenizer`]. See [`~MusicgenProcessor.__call__`] and [`~MusicgenProcessor.decode`] for more information. Args: feature_extractor (`EncodecFeatureExtractor`): An instance of [`EncodecFeatureExtractor`]. The feature extractor is a required input. tokenizer (`T5Tokenizer`): An instance of [`T5Tokenizer`]. The tokenizer is a required input. ZEncodecFeatureExtractor)Z T5TokenizerZT5TokenizerFastcs t|||j|_d|_dS)NF)super__init__feature_extractorcurrent_processor_in_target_context_manager)selfr tokenizer __class__q/var/www/html/Darija-Ai-Train/env/lib/python3.8/site-packages/transformers/models/musicgen/processing_musicgen.pyr +szMusicgenProcessor.__init__NTcCs|jj|||dS)N)tasklanguage no_timestamps)rget_decoder_prompt_ids)r rrrrrrr0sz(MusicgenProcessor.get_decoder_prompt_idscOs|jr|j||S|dd}|dd}|dd}t|dkrV|d}|dd}|dkrn|dkrntd|dk r|j|f|}|dk r|j|f|d|i|}|dkr|S|dkr|S|d|d<d |kr|d |d <|SdS) a Forwards the `audio` argument to EncodecFeatureExtractor's [`~EncodecFeatureExtractor.__call__`] and the `text` argument to [`~T5Tokenizer.__call__`]. Please refer to the doctsring of the above two methods for more information. audioN sampling_ratetextrzAYou need to specify either an `audio` or `text` input to process.Z input_values padding_mask)r r poplen ValueErrorrr )r argskwargsrrrinputsZ audio_inputsrrr__call__3s,        zMusicgenProcessor.__call__cOs`|dd}|dd}t|dkr8|d}|dd}|dk rN|j||dS|jj||SdS)ai This method is used to decode either batches of audio outputs from the MusicGen model, or batches of token ids from the tokenizer. In the case of decoding token ids, this method forwards all its arguments to T5Tokenizer's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring of this method for more information. rNrrr)r)rr _decode_audior batch_decode)r rr audio_valuesrrrrr$Ys    zMusicgenProcessor.batch_decodecOs|jj||S)z This method forwards all its arguments to T5Tokenizer's [`~PreTrainedTokenizer.decode`]. Please refer to the docstring of this method for more information. )rdecode)r rr rrrr&kszMusicgenProcessor.decode)rreturnc Cst|}|j\}}}|dkr$t|St|}||jd}d|jj}tj|dd|ffd|d}|}t|D]>}t ||||dddf|jjk} | |d||<qp|S)zn This method strips any padding from the audio values to return a list of numpy audio arrays. Nr)rrrZconstant)Zconstant_values) rshapelistr padding_valuenppadtolistrangeZasarrayZreshape) r r%rZbszZchannelsZseq_len differencer+iZ sliced_audiorrrr#rs    zMusicgenProcessor._decode_audio)NNT)N)__name__ __module__ __qualname____doc__Zfeature_extractor_classZtokenizer_classr rr"r$r&rrr,Zndarrayr# __classcell__rrrrrs   &r) r5typingrrnumpyr,Zprocessing_utilsrutilsrrrrrrs