� ���g� ����dZddlZddlZddlZejd��Zejd��Zejd��Zejd��ZdZ dZ d �Z d �Z d �Z d �Zdd �Zdd�ZdS)zUtilities for file names.�Nz([A-Z]+)([A-Z][a-z])z([a-z\d])([A-Z])z (?<!_)_(?!_)z(_{2,})z ^\w+(\.\w+)*$z<>:/\|?*c��t�d|��}t�d|��}|���S)z(Convert camel-case string to snake-case.z\1_\2)�_uppercase_uppercase_re�sub�_lowercase_uppercase_re�lower��names �_/home/asafur/pinokio/api/open-webui.git/app/env/lib/python3.11/site-packages/datasets/naming.py�camelcase_to_snakecaser "s:�� "� &� &�x�� 6� 6�D� "� &� &�x�� 6� 6�D� �:�:�<�<��c���t�|��}d�|D��}d�d�tj�|��D����S)z/Convert snake-case string to camel-case string.c�B�g|]}t�|����S�)�_multiple_underscores_re�split��.0�ns r � <listcomp>z*snakecase_to_camelcase.<locals>.<listcomp>,s'�� <� <� <�!� $� *� *�1� -� -� <� <� <r �c3�JK�|]}|dk�|���V��dS)rN)� capitalizers r � <genexpr>z)snakecase_to_camelcase.<locals>.<genexpr>-s6����Z�Z�a�RS�WY�RY�RY�1�<�<�>�>�RY�RY�RY�RY�Z�Zr )�_single_underscore_rer�join� itertools�chain� from_iterablers r �snakecase_to_camelcaser)sY�� � &� &�t� ,� ,�D� <� <�t� <� <� <�D� �7�7�Z�Z�9�?�+H�+H��+N�+N�Z�Z�Z� Z� Z�Zr c��tj�|��|krtd|�����t |��S)N�&Should be a dataset name, not a path: )�os�path�basename� ValueErrorr rs r �filename_prefix_for_namer&0sB�� �w�������%�%��H�$�H�H�I�I�I� !�$� '� '�'r c���tj�|��|krtd|�����t jt |��stdt �d|�d����t|���d|��S)Nr!zSplit name should match 'z '' but got 'z'.�-)r"r#r$r%�re�match� _split_rer&)r rs r �filename_prefix_for_splitr,6s��� �w�������%�%��H�$�H�H�I�I�I� �8�I�u� %� %�W��U�Y�U�U�E�U�U�U�V�V�V�&�t�,�,� 6� 6�u� 6� 6�6r c��t||��}|r|d|��z }tj�||��}|�d�S)N�.�*)r,r"r#r)� dataset_namer�data_dir�filetype_suffix�prefix�filepaths r �filepattern_for_dataset_splitr5>sM�� &�|�U� ;� ;�F��(��'�o�'�'�'���w�|�|�H�f�-�-�H��>�>�>�r c�����t||���tj�|����|r=t |�����fd�t ���D��}�r�fd�|D��}|S�}�r|d���z }|gS)Nc�(��g|]}��d|d�d�d����S)r(�05dz-of-r)r�shard_id� num_shardsr3s ��r rz/filenames_for_dataset_split.<locals>.<listcomp>Ls5���f�f�f���C�C��C�C�C�:�C�C�C�f�f�fr c� ��g|] }|d���z�� S)r.r)r�filenamer2s �r rz/filenames_for_dataset_split.<locals>.<listcomp>Ns'���T�T�T�h��$9��$9�$9�9�T�T�Tr r.)r,r"r#r�len�range) r#r0rr2� shard_lengths� filenamesr<r:r3s ` @@r �filenames_for_dataset_splitrAFs������ &�|�U� ;� ;�F� �W�\�\�$�� '� '�F�� ���'�'� �f�f�f�f�f�TY�Zd�Te�Te�f�f�f� � � U�T�T�T�T�)�T�T�T�I����� � .� �-�O�-�-� -�H��z�r )N)NN)�__doc__rr"r)�compilerrrrr+�"INVALID_WINDOWS_CHARACTERS_IN_PATHr rr&r,r5rArr r �<module>rEs���  ������ � � � � � � � �%�"�*�%<�=�=��$�"�*�%8�9�9��"�� �?�3�3��%�2�:�j�1�1�� � �%0�"����[�[�[�(�(�(� 7�7�7����������r
Memory