Format: application/x-python-code
1 o
2
3 ?�h� � @ s d dl ZG dd� d�ZdS )� Nc @ sr e Zd Zdefdd�Zdejdejfdd�Zdejdejfdd �Zdejdejfd
4 d �Z dejdejfd d
5 �Z
6 dS )�Sampler�sampling_configc C s
7 || _ d S )N)r )�selfr � r �<C:\Users\shubh\Videos\recall\recall_audit_sampler\sampler.py�__init__ s
8 zSampler.__init__�df�returnc C sP | j �d�}|dkr| �|�S |dkr| �|�S |dkr!| �|�S td|� ���)N�type�random�
9 stratifiedzstratified-proportionalz Unsupported sampling strategy: )r �get�
10 random_sample�stratified_sample� stratified_proportional_sample�
11 ValueError)r r �strategyr r r �sample s
12
13
14 zSampler.samplec C s4 | j �d�}|d u rtd��|jt|t|��dd�S )N� sample_sizez4`sample_size` must be specified for random sampling.�* ��n� random_state)r r
15 r r �min�len)r r r r r r r s zSampler.random_samplec C s@ | j �d�}| j �d�}|std��|�|�}g }|d u rI| j �dd�}td|� d�� |D ]\}}t|t|��} | dkrG|�|j| d d
16 �� q.nG|D ]D\}}t |t
17 �r\d � t t
18 |��nt
19 |�}
20 t |t�rk|�|
21 d�n|} t| t|��} | dkr�|�|j| d d
22 �� td |
23 � d
24 | � d| � �� qK|r�tj|dd�S tj|jd�S )N�by�ratioz/`by` must be specified for stratified sampling.Zdefault_per_group�
25 z1[INFO] No ratio provided. Using default ratio of z per group.r r r �:z
26 [STRATIFIED] u → requested: z , sampled: T�� ignore_index��columns)r r
27 r �groupby�printr r �appendr �
28 isinstance�tuple�join�map�str�dict�pd�concat� DataFramer" ) r r r r �grouped�sampled_groupsZ
29 default_ratio� group_key�group_dfr �
30 lookup_keyZgroup_nr r r r s0
31 �� "� zSampler.stratified_samplec C sN | j �d�}| j �d�}|std��|�|�}g }|d u rMdt|� }tdt|d �� d�� |D ]\}}tt|�| �} | dkrK|�|j | d d
32 �� q1nJ|D ]G\}}t
33 |t �r`d � t
34 t|��nt|�}
35 t
36 |t�ro|�|
37 d�n|} tt|�| �} | dkr�|�|j | d d
38 �� td |
39 � d
40 | d � d| � �� qO|r�tj|dd�S tj|jd�S )Nr �
41 proportionz<`by` must be specified for stratified-proportional sampling.� z+[INFO] No proportion provided. Using equal �d z % per group.r r r r z
42 [STRAT-PROP] u → u % → sampled: Tr r! )r r
43 r r# r r$ �round�intr% r r& r' r( r) r* r+ r, r- r. r" ) r r r r4 r/ r0 Zequal_pr1 r2 r r3 �pr r r r 6 s0
44 �� " � z&Sampler.stratified_proportional_sampleN) �__name__�
45 __module__� __qualname__r+ r r, r. r r r r r r r r r s r )�pandasr, r r r r r �<module> s