Pureml SDK & CLI can be directly installed using pip.

pip install pureml pureml_evaluate

For additional project requirements we will need to install the following packages

You can use the following command to install the packages.

pip install pandas lightgbm xlrd


you can create a requirements.txt file with the following contents


and run the following command

pip install -r requirements.txt

Download and load your dataset

Download your dataset from here.

Start by creating a function to load the dataset into a DataFrame. We will use the @load_data() decorator from PureML SDK.

import pureml
from pureml.decorators import load_data,transformer,dataset,model
import numpy as np
import pandas as pd
import lightgbm as lgb
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
import warnings
import random

rand_seed = 1234

def load_dataset():
    df = pd.read_csv('default of credit card clients.csv', header=1)

    return df

Preprocess the data

We can add a few more functions to preprocess the data. We will use the @transformer() decorator from PureML SDK.

with transformer functions. We specify the parent of the functions using the parent argument. This will ensure that the functions are executed in the order specified.

def remove_columns(df):
    return df.drop(['ID'],axis =1)

def rename_columns(df):
    return df.rename(columns={"PAY_0": "PAY_1","default payment next month":"default", "SEX":"sex"})    

def dataset_imbalances(df):
    categorical_features = ["sex", "EDUCATION", "MARRIAGE"]

    for col_name in categorical_features:
        df[col_name] = df[col_name].astype("category")

    Y, A = df.loc[:, "default"], df.loc[:, "sex"]
    X = pd.get_dummies(df.drop(columns=["default", "sex"]))

    A_str ={1: "male", 2: "female"})

    # Generate "Interest" column as a DataFrame
    interest_values = np.random.normal(loc=2 * Y, scale=A)
    interest_column = pd.DataFrame(interest_values, columns=["Interest"])

    # Concatenate "Interest" column with X DataFrame
    X = pd.concat([X, interest_column], axis=1)

    return {'X':X,'Y':Y,'A_str':A_str}

def resample_training_data(X_train, Y_train, A_train):
    negative_ids = Y_train[Y_train == 0].index
    positive_ids = Y_train[Y_train == 1].index
    balanced_ids = positive_ids.union(
        np.random.choice(a=negative_ids, size=len(positive_ids)))

    X_train = X_train.loc[balanced_ids, :]
    Y_train = Y_train.loc[balanced_ids]
    A_train = A_train.loc[balanced_ids]
    return  {"X_train": X_train, "Y_train":Y_train, "A_train": A_train}

def add_new_column(sensitive_features):
    values = ['Indian', 'African', 'American']

    list_length = sensitive_features.shape[0]
    full_list = values * (list_length // len(values))
    full_list += values[:list_length % len(values)]

    full_list = np.array(full_list)

    s_feat = pd.concat([sensitive_features.reset_index(drop=True), pd.DataFrame(full_list, columns=['race'])], axis=1)

    return s_feat

A transformer can have multiple parents. In this case, the transformer will be executed after all the parents have been executed. The output of the parents will be passed as input to the transformer.

Creating a dataset

We can now create a dataset from the pipeline. The dataset will be created by executing the pipeline and saving the output of the last transformer in the pipeline. The dataset can be created by using the @dataset decorator. The decorator takes the following arguments:

  • label: The name of the dataset
  • parent: The name of the transformer whose output will be saved as the dataset
  • upload: If True, the dataset will be uploaded to the cloud. If False, the dataset will be saved locally.
@dataset(label='Credit Loan Dataset4',upload=True)
def create_dataset():
    df = load_dataset()
    df = remove_columns(df)
    df = rename_columns(df)
    data  = dataset_imbalances(df)
    X,Y,A_str = data['X'],data['Y'],data['A_str']
    X_train, X_test, y_train, y_test, A_train, A_test = train_test_split(X, Y, A_str, test_size=0.35, stratify=Y)
    data = resample_training_data(X_train, y_train, A_train)
    X_train, y_train, A_train = data['X_train'],data['Y_train'],data['A_train']

    A_test = add_new_column(sensitive_features=A_test)

    return {"x_train":X_train,"y_train":y_train.to_numpy(),"x_test":X_test,"y_test":y_test.to_numpy(),"sensitive_features" : A_test}


Creating a model to classify the dataset

With the PureML model module, you can perform a variety of actions related to creating and managing models. PureML assists you with training and tracking all of your machine learning project information, including ML models and datasets, using semantic versioning and full artifact logging.

We can make a separate python file for the model. The model file will contain the model definition and the training code.

The model training function can be created by using the @model decorator. The decorator takes the model name as the argument in the format model_name.

from pureml.decorators import model
@model(label='Credit Model Underwriting')
def create_model():
    data = pureml.dataset.fetch('Credit Loan Dataset4:v1')
    x_train = data['x_train']
    y_train = data['y_train']
    lgb_params = {
    "objective": "binary",
    "metric": "auc",
    "learning_rate": 0.412,
    "num_leaves": 10,
    "max_depth": 3,
    "random_state": rand_seed,
    "n_jobs": 1,}

    estimator = Pipeline(
            ("preprocessing", StandardScaler()),
            ("classifier", lgb.LGBMClassifier(**lgb_params)),
    ), y_train)
    return estimator

model_lgb = create_model()

Once our training is complete our model will be ready to rock and roll🎸✨. But that’s too much of a hassle. So for now, let’s just do some predictions

Add prediction to your model

For registered models, prediction function along with its requirements and resources can be logged to be used for further processes like evaluating and packaging.

PureML predict module has a method add. Here we are using the following arguments:

  • label: The name of the model (model_name:version)
  • paths: The path to the file and requirements.txt file.

Our file has the script to load the model and make predictions. The requirements.txt file has the dependencies required to run the file.

You can know more about the prediction process here

from pureml import BasePredictor, Input, Output
import pureml
from typing import Any

class Predictor(BasePredictor):
    label:Any = "Credit Model Underwriting:v1"
    input:Any = Input(type="numpy ndarray")
    output:Any = Output(type="numpy ndarray")

    def load_models(self):
        self.model = pureml.model.fetch(self.label)

    def predict(self, data):
        predictions = self.model.predict(data)

        return predictions
store the above python code as The predict file is specific for this example
import pureml

                   paths={"predict": "./", "requirements":"./requirements.txt"})

Create your first Evaluation

from pureml_policy import policy_eval
results = policy_eval.eval(
            label_model='Credit Model Underwriting:v1',
            label_dataset='Credit Loan Dataset4:v1')

Congrats! You have successfully created your first evaluation. You can now apply polices in Dashboard.

To know more about applying policies, you can refer to the documentation here

To know more about the Documents upload, you can refer to the documentation here

To know more about the Questionaire, you can refer to the documentation here

To know more about the Forms, you can refer to the documentation here

To know more about Report Generation, you can refer to the documentation here