fungtk/VIP_sequence
0
1�
2 ���d� � �T � d dl Zd� Zd� Zd dlmZ d dlZd dlmZ d dl Z d� Z3 ej � � 5 Z ej
� � 5 ej � � 5 ej ddd g�4� � Zddd� � n # 1 swxY w Y ej � � 5 ej d d � � � Z ej d
d
� � � Z ej dd� � � Zddd� � n # 1 swxY w Y ddd� � n # 1 swxY w Y ej
� � 5 ej d� � Zddd� � n # 1 swxY w Y ej
� � 5 ej d�� � Ze� e5eeeeged�� � ddd� � n # 1 swxY w Y ddd� � n # 1 swxY w Y e � � � dS )� Nc � � | � d� � � } | j j � � � }|| _ | d j � ddd�� � j � ddd�� � | d< t j | d d� � � � d6� � � d d7� � | d< | S ) Nc �N � | j dk r| j � � � n| S )N�object)�dtype�str�strip)�xs �5C:\Users\fungtk\PycharmProjects\data analysis\main.py�<lambda>z clean_data.<locals>.<lambda> s � �Q�W��-@�-@�A�E�K�K�M�M�M�a� � �GMV�,� T)�regex�$�coerce)�errors�����r )�apply�columnsr r �replace�pd�8to_numeric�fillna)�dfr s r9 �10clean_datar s� � � ���G�G� H� H�B��j�n�"�"�$�$�G��B�J� �5� �
�%�%�c�2�D�%�9�9�=�E�E�c�2�TX�E�Y�Y�B�u�I� �
�b��i��9�9�9�@�@��D�D�L�L�Q�PR�S�S�B�u�I�
�Ir c �V � �� t dt � � � � � D �]�t � �fd�t d� � D � � � � r։dz }|t � � � k rh� j | d � j � d k r@|dz
}|t � � � k r(� j | d � j � d k �@|t � � � k r|dz }� j � � � � }� j | � j �< |� j |< �dk r�t � �fd�t d� � D � � � � r։dz }|t � � � k rh� j | d � j � d k r@|dz
}|t � � � k r(� j | d � j � d k �@|t � � � k r|dz }� j � � � � }� j | � j �< |� j |< �� � S )N� c 3 �z �K � | ]5}�j �|z11 d �j �|z12 dz13 d k V � �6dS )�Main Cat� N��iloc��.0�jr �is ��r14 � <genexpr>z%cat_brand_checking.<locals>.<genexpr> sL �� � � �^�^�PQ�r�w�q�1�u�~�j�)�R�W�Q��U�Q�Y�-?�15�-K�K�^�^�^�^�^�^r � r! r c 3 �z �K � | ]5}�j �|z16 d �j �|z17 dz18 d k V � �6dS )�19Brand Namer! Nr" r$ s ��r20 r( z%cat_brand_checking.<locals>.<genexpr>% sL �� � � �m�m�_`�"�'�!�a�%�.� �6�"�'�!�a�%�!�)�:L�\�:Z�Z�m�m�m�m�m�mr r+ )�range�len�allr# �copy)r �next_cat_index�temp�next_brand_indexr' s ` @r21 �cat_brand_checkingr3 sC ��� �
�1�c�"�g�g�
�
� -� -�� �^�^�^�^�^�UZ�[\�U]�U]�^�^�^� ^� ^� +� ��U�N� �3�r�7�7�*�*�r�w�~�/F�z�/R�VX�V]�^_�V`�ak�Vl�/l�/l� �!�#�� !�3�r�7�7�*�*�r�w�~�/F�z�/R�VX�V]�^_�V`�ak�Vl�/l�/l� ��R���(�(� �!�#���7�1�:�?�?�$�$�D��� �0�B�G�A�J�&*�B�G�N� #�
��6�6�c�m�m�m�m�m�di�jk�dl�dl�m�m�m�m�m�6� !�1�u� �"�S��W�W�,�,���9I�1J�<�1X�\^�\c�de�\f�gs�\t�1t�1t� �A�%� � #�S��W�W�,�,���9I�1J�<�1X�\^�\c�de�\f�gs�\t�1t�1t� �3�r�7�7�*�*� �A�%� ��7�1�:�?�?�$�$�D���!1�2�B�G�A�J�(,�B�G�$� %�� �Ir )�random)�shufflec �� � | j } t j | |t �� � }t |� � }|}t j | |t �� � }|d � t � � |d< |d j � d� � }|| }t |� � � d�� � � d�� � }t |� � }t j | |t �� � }|d � t � � |d< |d j � d� � } || }t |� � }ddd d22t d � � g}23g d �} t j |d
|24| d�� � |d< g d�} t j |d | d�� � |d< |�
d� � }
g }|
D ])\ }}t |� � }|� |� � �*t j |� � }t j |||g� � }|� dd�� � }|d
� dd� � |j d d �d
f< d}|� |d�� � |S )N)�25sheet_namer �#zN-r! )�fracT)�dropr iP� i�� �inf)�No GMV�$1 - 50K� $50k - 100k�>$100Kr
)�bins�labels�include_lowest� GMV_group)r? r> r= r< )�26categories�orderedz27SKU ID (Y)�first)�subset�keepr z result.xlsxF)�index)�namer �28read_excelr r �astype�contains�sample� reset_indexr3 �float�cut� Categorical�groupby�append�concat�drop_duplicatesr �loc�to_excel)�filename�sheet1�sheet2�sheet3r �df_highlight_final�29df_top_150� yn_mask_150�df_gmv_ranking�yn_mask_gmv_rankingr@ rA �ordered_labels�30grouped_df�shuffled_groupsrJ �group�shuffled_group�shuffled_df_GMV_ranking� df_combined� unique_df�files r31 �sort_excel_filerk 8 s� � ��m�H� 32 � �x�F�#� >� >� >�B� �B� � �B� �� � �x�F�#�F�F�F�J� ��o�,�,�S�1�1�J�s�O� �S�/�%�.�.�t�4�4�K��[�L�)�J��J�'�'�.�.�A�.�6�6�B�B��B�M�M�J�#�J�/�/�J��]�8��c�J�J�J�N�(��-�4�4�S�9�9�N�3��(��-�1�:�:�4�@�@��#�%8�$8�9�N� ��/�/�N� ��5�&�%��,�,� /�D�
<�
<�
<�F�"$�&� ��)>�T�RX�im�"n�"n�"n�N�;� �D�D�D�N�"$�.� � �1L�Yg�qu�"v�"v�"v�N�;� � �'�'� �4�4�J��O�!� /� /� ��e� ,�E�2�2��� � �~�.�.�.�.� �i�� 8� 8���)� /�33�<S� T�U�U�K� �+�+�<�W�+�M�M�I� %�e� ,� 4� 4�b�� ;� ;�I�M�!�!�!�U�(�� �D�
���t�5��)�)�)� �Kr z File Namez.xlsxz.csv)�label�34file_types�Sheet1)rl �value�Sheet2�Sheet3z BB sequencez35Output Box)rl rk )�fn�inputs�outputs�api_name)�gradio�grr r3 �numpyr4 �pandasr r5 �osrk �Blocks�demo�Row�Column�File� File_Name�Textboxrn rp rq �Button� greet_btn�output�click�launch� r r36 �<module>r� s� �� � � � � 37� 38� 39� � � �B � � � � � � � � � � � � � � � � � � � �@� @� @�D �R�Y�[�[� 4�D� ����� @� @�
�R�Y�[�[� Q� Q� ���k�w��>O�P�P�P�I� Q� Q� Q� Q� Q� Q� Q� Q� Q� Q� Q���� Q� Q� Q� Q�
�R�Y�[�[� @� @� �R�Z�h�h�?�?�?�F� �R�Z�h�h�?�?�?�F� �R�Z�h�h�?�?�?�F� @� @� @� @� @� @� @� @� @� @� @���� @� @� @� @�@� @� @� @� @� @� @� @� @� @� @���� @� @� @� @� 40����� -� -� �B�I�m�,�,� �-� -� -� -� -� -� -� -� -� -� -���� -� -� -� -� 41����� 4� 4����|�,�,�,�����?�I�v�v�v�3V�`f�!2� � 4� 4� 4�4� 4� 4� 4� 4� 4� 4� 4� 4� 4� 4���� 4� 4� 4� 4�4� 4� 4� 4� 4� 4� 4� 4� 4� 4� 4���� 4� 4� 4� 4�, � � �
�
�
�
�
s� �F �C&�A2�& C&�2A6 �6C&�9A6 �:C&� 7C� C&�C �C&�C �C&� F �&C* �*F �-C* �.F � D � F � D! �!F �$D! �%F �7/E2�& F �2E6 �6F �9E6 �:F � F
�F