a
    é´«j*  ã                   @   s  d Z ddlmZmZ ddlZddlZddlZddlZddl	Z	ddl
mZmZ ddlmZ e	j d¡Ze	j d¡Zdd„ Zed	d
„ ƒZee	j dddg¡e	j dg d¢¡dd„ ƒƒƒZdd„ Zdd„ Ze	j ddi eddgiƒfdddieddgiƒfdddgieddd giƒfd!dgd"d#œeddgiƒfd!dgd$d#œedejdgiƒfg¡d%d&„ ƒZd'd(„ Ze	j d)g d*¢¡d+d,„ ƒZe	j d-d"d$g¡d.d/„ ƒZd0d1„ Z e	j dg d2¢¡d3d4„ ƒZ!e	j dg d5¢¡d6d7„ ƒZ"d8d9„ Z#d:d;„ Z$e	j d<¡e	j d=d>d?g¡d@dA„ ƒƒZ%dS )BzZ
Tests encoding functionality during parsing
for all of the parsers defined in parsers.py
é    )ÚBytesIOÚTextIOWrapperN)Ú	DataFrameÚread_csvz=ignore:Passing a BlockManager to DataFrame:DeprecationWarningZpyarrow_skipc                 C   sL   d}| }t d |¡ƒ}|j|d|d�}tddggddgd	�}t ||¡ d S )
NÚcp1255u   ×©×œ×•×�:1234
562:123ú:)ÚsepÚencodingi2  é{   u   ×©×œ×•×�Ú1234©Úcolumns©r   Úencoder   r   ÚtmÚassert_frame_equal)Úall_parsersr	   ÚparserÚdataÚresultÚexpected© r   úc/var/www/html/laksyashrimp/venv/lib/python3.9/site-packages/pandas/tests/io/parser/test_encoding.pyÚtest_bytes_io_input   s    r   c                 C   s@   | }t d ¡ ƒ}|j|ddd d�}tddggƒ}t ||¡ d S )Nu   Å�aski, Jan;1ú;úutf-8)r   r	   Úheaderu   Å�aski, Jané   r   )r   r   r   r   r   r   r   r   Útest_read_csv_unicode(   s
    r   r   ú,ú	r	   )úutf-16zutf-16lezutf-16bec              	   C   s  | }d  d|¡}dt ¡ › d�}|ddœ}d}t |¡�¼}| |¡}t|dƒ�}	|	 |¡ W d   ƒ n1 sn0    Y  tt	| |¡ƒ|d	��<}
|j
|fd
|i|¤Ž}|j
|
fd
|i|¤Ž}W d   ƒ n1 sÐ0    Y  t ||¡ W d   ƒ n1 sú0    Y  d S )Nz)skip this
skip this too
A,B,C
1,2,3
4,5,6r   Ú__z__.csvé   )r   Zskiprowsr   Úwb©r	   r	   )ÚreplaceÚuuidÚuuid4r   Úensure_cleanr   ÚopenÚwriter   r   r   r   )r   r   r	   r   r   ÚpathÚkwargsÚutf8Ú
bytes_dataÚfZbytes_bufferr   r   r   r   r   Útest_utf16_bom_skiprows2   s    û

(4r1   c                 C   s6   t j |d¡}| }|j|ddd�}t|ƒdks2J ‚d S )Nzutf16_ex.txtr!   r    )r	   r   é2   )Úosr,   Újoinr   Úlen)r   Úcsv_dir_pathr,   r   r   r   r   r   Útest_utf16_exampleO   s    r7   c                 C   sL   t j |d¡}| }|j|d dd�}| d¡}|d d }d}||ksHJ ‚d S )Núunicode_series.csvúlatin-1)r   r	   r   r   i`  u$   Ã� kÃ¶ldum klaka (Cold Fever) (1994))r3   r,   r4   r   Z	set_index)r   r6   r,   r   r   Úgotr   r   r   r   Útest_unicode_encodingV   s    
r;   zdata,kwargs,expectedza
1Úar   z"a"
1Ú	quotecharú"zb
1ÚnamesÚbÚ1ú
1T)r?   Úskip_blank_linesFc                    sn   | }d‰ d‰‡ ‡fdd„}|j dkrD|dkrD| dd¡rDtjd	d
� |j||ƒfdˆi|¤Ž}t ||¡ d S )Nu   ï»¿r   c                    s   ˆ |    ˆ¡}t|ƒS )N)r   r   )Ú_dataZbom_data©Úbomr.   r   r   Ú_encode_data_with_bom{   s    z,test_utf8_bom.<locals>._encode_data_with_bomÚpyarrowrB   rC   Tz,https://github.com/apache/arrow/issues/38676)Úreasonr	   )ÚengineÚgetÚpytestÚskipr   r   r   )r   r   r-   r   Úrequestr   rG   r   r   rE   r   Útest_utf8_bomb   s    ÿþ
ýrO   c                 C   sL   t dgdgdœƒ}| }| |¡}d |¡}|jt|ƒ|d�}t ||¡ d S )Ng333333@Útest)Zmb_numZ	multibytezmb_num,multibyte
4.8,testr%   )r   Úformatr   r   r   r   r   )r   Ú	utf_valueÚencoding_fmtr   r   r	   r   r   r   r   r   Útest_read_csv_utf_aliases‹   s    

rT   zfile_path,encoding)))Úior   Úcsvz	test1.csvr   ))rU   r   r   r8   r9   ))rU   r   r   zsauron.SHIFT_JIS.csvÚshiftjisc           
      C   s  | }||Ž }|j ||d�}t||d��$}|  |¡}|jr<J ‚W d   ƒ n1 sP0    Y  t ||¡ t|dd��(}	|j |	|d�}|	jrŒJ ‚W d   ƒ n1 s 0    Y  t ||¡ t|ddd��(}	|j |	|d�}|	jrÞJ ‚W d   ƒ n1 sò0    Y  t ||¡ d S )Nr%   Úrb)Úmoder   )rY   Ú	buffering)r   r*   Úclosedr   r   )
r   Ú	file_pathr	   Údatapathr   Zfpathr   Úfar   Zfbr   r   r   Útest_binary_mode_file_buffers—   s    
(((r_   Úpass_encodingc           	      C   sª   | }|  |¡}|jdkr2|du r2|dv r2t d¡ tddgiƒ}tjd|dd��F}| d	¡ | d
¡ |j	||rt|nd d�}t 
||¡ W d   ƒ n1 sœ0    Y  d S )NrH   T)é   é    zThese cases freezeZfooÚbarzw+)rY   r	   Zreturn_filelikezfoo
barr   r%   )rQ   rJ   rL   rM   r   r   r)   r+   Úseekr   r   )	r   rR   rS   r`   r   r	   r   r0   r   r   r   r   Útest_encoding_temp_file·   s    



re   c                 C   s’   | }d}d}d}t ||giƒ}t ¡ �X}| |› d|› � |¡¡ | d¡ |j||d�}t ||¡ |j	rpJ ‚W d   ƒ n1 s„0    Y  d S )Nz	shift-jisu	   ã�¦ã�™ã�¨u   ã�“ã‚€Ú
r   r%   )
r   ÚtempfileÚNamedTemporaryFiler+   r   rd   r   r   r   r[   )r   r   r	   Útitler   r   r0   r   r   r   r   Útest_encoding_named_temp_fileË   s    

rj   )r   r!   z	utf-16-bez	utf-16-lezutf-32c                 C   sR   d}t | | ¡ƒ}t|d| d�}tddgddgdd	ggd
dgd�}t ||¡ d S )Nu   a	b
ï¼šfoo	0
bar	1
baz	2r    )Ú	delimiterr	   u   ï¼šfoor   rc   r   Zbazr#   r<   r@   )r   r   r   )r	   r   Zencoded_datar   r   r   r   r   Ú%test_parse_encoded_special_charactersß   s    þrl   )r   Nr!   r   r9   c              	   C   sÔ   | }t g d¢g d¢g d¢dœƒ}t ¡ �Œ}|j|d|d� |jdkr–d}tjt|d	��  |j||d
d� W d   ƒ n1 s|0    Y  W d   ƒ d S |j||d
d�}W d   ƒ n1 sº0    Y  t 	||¡ d S )N)ZRaphaelZ	DonatellozMiguel AngelZLeonardo)ÚredÚpurpleZorangeÚblue)Zsaizbo staffZnunchunkZkatana)ÚnameÚmaskZweaponF)Úindexr	   rH   úBThe 'memory_map' option is not supported with the 'pyarrow' engine©ÚmatchT)r	   Ú
memory_map)
r   r   r)   Úto_csvrJ   rL   ÚraisesÚ
ValueErrorr   r   )r   r	   r   r   ÚfileÚmsgÚdfr   r   r   Útest_encoding_memory_mapð   s     ýÿ

..r}   c              	   C   sÖ   | }t dgd d�}d|jd< t d¡�Ž}|j|dddd	� |jd
kr˜d}tjt|d��  |j	|ddd� W d  ƒ n1 s~0    Y  W d  ƒ dS |j	|ddd�}W d  ƒ n1 s¼0    Y  t 
||¡ dS )zO
    Chunk splits a multibyte character with memory_map=True

    GH 43540
    Zaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaai   )r   u�   aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaÄ…iÿ  zbug-gh43540.csvFr   ©rr   r   r	   rH   rs   rt   NT)r   rv   )r   Zilocr   r)   rw   rJ   rL   rx   ry   r   r   )r   r   r|   Úfnamer{   Údfrr   r   r   Ú test_chunk_splits_multibyte_char  s    

..r�   c              	   C   sH  g }d}d}d}t t|ƒt|ƒ|ƒD ]V}d dd„ t ||d ƒD ƒ¡d }z| d¡ W n tyn   Y q$Y n0 | |¡ q$| }t|ƒ}t d	¡�”}	|j	|	d
d
dd� |j
dk�rd}
tjt|
d��" |j|	dddd� W d  ƒ n1 sì0    Y  W d  ƒ dS |j|	dddd�}W d  ƒ n1 �s.0    Y  t ||¡ dS )zg
    GH 43787

    Test correct handling of UTF-8 chars when memory_map=True and encoding is UTF-8
    é€   ú u   ð�‚€Ú c                 S   s   g | ]}t |ƒ‘qS r   )Úchr)Ú.0Úcr   r   r   Ú
<listcomp>/  ó    z,test_readcsv_memmap_utf8.<locals>.<listcomp>rf   r   zutf8test.csvFr~   rH   rs   rt   NT)r   rv   r	   )ÚrangeÚordr4   r   ÚUnicodeEncodeErrorÚappendr   r   r)   rw   rJ   rL   rx   ry   r   r   )r   ÚlinesÚline_lengthZ
start_charZend_charÚlnumÚliner   r|   r   r{   r€   r   r   r   Útest_readcsv_memmap_utf8"  s,    "
02r’   Zpyarrow_xfailrY   zw+bzw+tc                 C   s~   | }d}d|v rd}t j|dd��.}| |¡ | d¡ | |¡}W d   ƒ n1 sV0    Y  tg dgd�}t ||¡ d S )Ns   abcdÚtÚabcdr   )rY   r	   r   r   )rg   ÚSpooledTemporaryFiler+   rd   r   r   r   r   )r   rY   r   ÚcontentÚhandler|   r   r   r   r   Útest_not_readableD  s    

(r˜   )&Ú__doc__rU   r   r   r3   rg   r'   ÚnumpyÚnprL   Zpandasr   r   Zpandas._testingZ_testingr   ÚmarkÚfilterwarningsZ
pytestmarkZusefixturesZskip_pyarrowr   r   Zparametrizer1   r7   r;   ÚnanrO   rT   r_   re   rj   rl   r}   r�   r’   r˜   r   r   r   r   Ú<module>   sl   ÿ
	
ýöþ
þ

ÿ

"
