o NiV@sdZddlmZddlmZddlmZddlZddlZddlZddlZ ddl m m Z ddlmZdddd d d d d ZgdZdZdZddZddZGdddejjZdS)zSAVEE dataset.)absolute_import)division)print_functionNangerdisgustfear happinessneutralsadnesssurprise)adfhnsasu)DCJEJKKLaG @inproceedings{Vlasenko_combiningframe, author = {Vlasenko, Bogdan and Schuller, Bjorn and Wendemuth, Andreas and Rigoll, Gerhard}, year = {2007}, month = {01}, pages = {2249-2252}, title = {Combining frame and turn-level information for robust recognition of emotions within speech}, journal = {Proceedings of Interspeech} } a SAVEE (Surrey Audio-Visual Expressed Emotion) is an emotion recognition dataset. It consists of recordings from 4 male actors in 7 different emotions, 480 British English utterances in total. The sentences were chosen from the standard TIMIT corpus and phonetically-balanced for each emotion. This release contains only the audio stream from the original audio-visual recording. The data is split so that the training set consists of 2 speakers, and both the validation and test set consists of samples from 1 speaker, respectively. cCst||krtdjt||dtdd|D}td|dkr)td|g}d}|D]\}}|}||7}||t||t||fq/|d d |d d|f|d <|S) aComputes boundary indices for each of the splits in split_probs. Args: split_probs: List of (split_name, prob), e.g. [('train', 0.6), ('dev', 0.2), ('test', 0.2)] n_items: Number of items we want to split. Returns: The item indices of boundaries between different splits. For the above example and n_items=100, these will be [('train', 0, 60), ('dev', 60, 80), ('test', 80, 100)]. z]Not enough items for the splits. There are {splits} splits while there are only {items} items)splitsitemscs|]\}}|VqdSN).0nameprrS/home/ubuntu/.local/lib/python3.10/site-packages/tensorflow_datasets/audio/savee.py Rz,_compute_split_boundaries..g:0yE>z"Probs should sum up to 1. probs={}gr)len ValueErrorformatsumabsappendint) split_probsn_items total_probssplit_boundariessum_prrprevrrr_compute_split_boundariesAs$  "r1cCsttdd|D}tj|}||t|t|}i}|D]\}}} t|| D]} |||| <q+q!t t} |D]\} } || }| | | q<| S)aSplit items to train/dev/test, so all items in group go into same split. Each group contains all the samples from the same speaker ID. The samples are splitted so that all each speaker belongs to exactly one split. Args: items_and_groups: Sequence of (item_id, group_id) pairs. split_probs: List of (split_name, prob), e.g. [('train', 0.6), ('dev', 0.2), ('test', 0.2)] split_number: Generated splits should change with split_number. Returns: Dictionary that looks like {split name -> set(ids)}. csrrr)ritem_idgroup_idrrrr rr!z-_get_inter_splits_by_group..) sortedsetnprandom RandomStateshuffler1r$range collections defaultdictadd)items_and_groupsr+ split_numbergroupsrngr.group_id_to_split split_namei_starti_endi split_to_idsr2r3splitrrr_get_inter_splits_by_groupcs    rIc@s8eZdZdZejdZdZddZ ddZ dd Z d S) Saveez8The audio part of SAVEE dataset for emotion recognition.z1.0.0a2 manual_dir should contain the file AudioData.zip. This file should be under Data/Zip/AudioData.zip in the dataset folder provided upon registration. You need to register at http://personal.ee.surrey.ac.uk/Personal/P.Jackson/SAVEE/Register.html in order to get the link to download the dataset. c CsHtjj|ttjtjjdddtjjtt dt j dddt dS) NwaviD) file_format sample_rate)namesaudiolabel speaker_id)rPrQz%http://kahlan.eps.surrey.ac.uk/savee/)builder descriptionfeaturessupervised_keyshomepagecitation)tfdscore DatasetInfo _DESCRIPTIONrU FeaturesDictAudio ClassLabellist LABEL_MAPvaluestfstring _CITATION)selfrrr_infosz Savee._infoc Cstj|jd}tjj|std || |}g}tjj d |D]}tj |\}}tj |\}}| ||fq'gd} t|| d} tjjtjjd| didtjjtjjd| d idtjjtjjd| d idgS) zReturns SplitGenerators.z AudioData.zipz`SAVEE requires manual download of the data. Please download the audio data and place it into: {}z{}/AudioData/*/*.wav))traing333333?) validation皙?)testrjr file_namesrh)r gen_kwargsrirk)ospathjoin manual_dirrciogfileexistsAssertionErrorr&extractglobrHr)rIrYrZSplitGeneratorSplitTRAIN VALIDATIONTEST) rf dl_managerzip_path extract_pathr>fnamefolder_rRr+rrrr_split_generatorss6     zSavee._split_generatorsc csr|D]3}tj|\}}tj|\}}td|d}t|}d||dd} | |||dfVqdS)zYields examples.z ^([a-zA-Z]+)r"z{}_{}.rrON)rnrorHrematchgrouprar&) rfrlrrwavnamerrR label_abbrevrQkeyrrr_generate_examplesszSavee._generate_examplesN) __name__ __module__ __qualname____doc__rYrZVersionVERSIONMANUAL_DOWNLOAD_INSTRUCTIONSrgrrrrrrrJs  #rJ)r __future__rrrr;rnrnumpyr6tensorflow.compat.v2compatv2rctensorflow_datasets.public_api public_apirYraSPEAKERSrer\r1rIrZGeneratorBasedBuilderrJrrrrs0       "!