o
    0 i;$                     @   s   d dl Z d dlZd dlmZmZ d dlmZ d dlmZm	Z	 d dl
Z
d dlmZ d dlmZ ddlmZ dd	lmZ dd
lmZmZ ddlmZmZmZmZ eeZee Z e!dd e D Z"eG dd dZ#G dd deZ$G dd deZ%dS )    N)	dataclassfield)Enum)OptionalUnion)FileLock)Dataset   )$MODEL_FOR_QUESTION_ANSWERING_MAPPING)PreTrainedTokenizer)check_torch_load_is_safelogging   )SquadFeaturesSquadV1ProcessorSquadV2Processor"squad_convert_examples_to_featuresc                 c   s    | ]}|j V  qd S N)
model_type).0conf r   l/home/app/PaddleOCR-VL-test/.venv_paddleocr/lib/python3.10/site-packages/transformers/data/datasets/squad.py	<genexpr>"   s    r   c                   @   s<  e Zd ZU dZedddde idZee	d< edddidZ
ee	d	< ed
ddidZee	d< ed
ddidZee	d< edddidZee	d< edddidZee	d< edddidZee	d< edddidZee	d< edddidZee	d< edddidZee	d< eddd idZee	d!< ed"dd#idZee	d$< dS )%SquadDataTrainingArgumentszb
    Arguments pertaining to what data we are going to input our model for training and eval.
    Nhelpz!Model type selected in the list: z, )defaultmetadatar   zFThe input data dir. Should contain the .json files for the SQuAD task.data_dir   zThe maximum total input sequence length after tokenization. Sequences longer than this will be truncated, sequences shorter will be padded.max_seq_lengthzVWhen splitting up a long document into chunks, how much stride to take between chunks.
doc_stride@   zkThe maximum number of tokens for the question. Questions longer than this will be truncated to this length.max_query_length   zThe maximum length of an answer that can be generated. This is needed because the start and end predictions are not conditioned on one another.max_answer_lengthFz1Overwrite the cached training and evaluation setsoverwrite_cachezDIf true, the SQuAD examples contain some that do not have an answer.version_2_with_negativeg        zIIf null_score - best_non_null is greater than the threshold predict null.null_score_diff_threshold   n_best_sizer   zjlanguage id of input for language-specific xlm models (see tokenization_xlm.PRETRAINED_INIT_CONFIGURATION)lang_id   z3multiple threads for converting example to featuresthreads)__name__
__module____qualname____doc__r   joinMODEL_TYPESr   str__annotations__r   r    intr!   r#   r%   r&   boolr'   r(   floatr*   r+   r-   r   r   r   r   r   %   sb   
 				r   c                   @   s   e Zd ZdZdZdS )SplittraindevN)r.   r/   r0   r:   r;   r   r   r   r   r9   h   s    r9   c                   @   s   e Zd ZU dZeed< ee ed< eed< e	ed< dej
dddfded	ed
ee deeef de	dee defddZdd Zdeeejf fddZdS )SquadDatasetzH
    This will be superseded by a framework-agnostic approach soon.
    argsfeaturesmodeis_language_sensitiveNFpt	tokenizerlimit_length	cache_dirdataset_formatc                 C   s&  || _ || _|jrt nt | _t|tr'zt| }W n t	y&   t	dw || _
|jr/dnd}tj|d ur:|n|jd|j d|jj d|j d| }	|	d }
t|
 tj|	r|jst }t  tj|	dd| _| jd	 | _| jd
d | _| jdd | _t d|	 dt |  | jd u s| jd u rt!d|	 d nX|tj"kr| j#|j| _n| j$|j| _t%| j||j|j&|j'|tj(k|j)|d\| _| _t }t*| j| j| jd|	 t d|	 dt | dd W d    d S W d    d S 1 sw   Y  d S )Nzmode is not a valid split namev2v1Zcached__z.lockT)Zweights_onlyr>   datasetexamplesz"Loading features from cached file z [took %.3f s]zDeleting cached file z; will allow dataset and examples to be cached in future run)rJ   rB   r    r!   r#   Zis_trainingr-   Zreturn_dataset)r>   rI   rJ   z!Saving features into cached file z [took z.3fz s])+r=   r@   r'   r   r   	processor
isinstancer4   r9   KeyErrorr?   ospathr2   r   value	__class__r.   r    r   existsr&   timer   torchloadZold_featuresr>   getrI   rJ   loggerinfowarningr;   Zget_dev_examplesZget_train_examplesr   r!   r#   r:   r-   save)selfr=   rB   rC   r?   r@   rD   rE   Zversion_tagZcached_features_fileZ	lock_pathstartr   r   r   __init__w   sp   

"


$ zSquadDataset.__init__c                 C   s
   t | jS r   )lenr>   )r[   r   r   r   __len__   s   
zSquadDataset.__len__returnc                 C   s4  | j | }tj|jtjd}tj|jtjd}tj|jtjd}tj|jtjd}tj|jtj	d}tj|j
tj	d}|||d}	| jjdv rJ|	d= | jjdv rx|	||d | jjrc|	d|i | jrx|	dtj|jtjd| jj i | jtjkrtj|jtjd}
tj|jtjd}|	|
|d	 |	S )
N)Zdtype)	input_idsattention_masktoken_type_ids)xlmZrobertaZ
distilbertZ	camembertrc   )Zxlnetrd   )	cls_indexp_maskis_impossibleZlangs)start_positionsend_positions)r>   rT   Ztensorra   longrb   rc   re   rf   r8   rg   r=   r   updater'   r@   ZonesshapeZint64r+   r?   r9   r:   Zstart_positionZend_position)r[   ifeaturera   rb   rc   re   rf   rg   Zinputsrh   ri   r   r   r   __getitem__   s0   
$zSquadDataset.__getitem__)r.   r/   r0   r1   r   r5   listr   r9   r7   r:   r   r   r6   r   r4   r]   r_   dictrT   ZTensorro   r   r   r   r   r<   m   s8   
 

Lr<   )&rN   rS   dataclassesr   r   enumr   typingr   r   rT   Zfilelockr   Ztorch.utils.datar   Zmodels.auto.modeling_autor
   Ztokenization_utilsr   utilsr   r   Zprocessors.squadr   r   r   r   Z
get_loggerr.   rW   rp   keysZMODEL_CONFIG_CLASSEStupler3   r   r9   r<   r   r   r   r   <module>   s&   
B