predict.py

from sklearn import preprocessing
from sklearn import ensemble
from sklearn import metrics
import pandas as pd
import numpy as np
import os
import joblib

from . import dispatcher

TEST_DATA = os.environ.get("TEST_DATA")

MODEL = os.environ.get("MODEL")

def predict():
   
    df = pd.read_csv(TEST_DATA)
    test_idx = df['id'].values
    predictions = None

    for FOLD in range(5):
    
        print(FOLD)
        df = pd.read_csv(TEST_DATA)
        encoders = joblib.load(os.path.join('models',f"{MODEL}_{FOLD}_label_encoders.pkl"))
        cols = joblib.load(os.path.join('models',f"{MODEL}_{FOLD}_columns.pkl"))
        for c in cols:
            lbl = encoders[c]
            df.loc[:, c] = lbl.transform(df[c].values.tolist())
            
        #data is ready
        #testing
        clf = joblib.load(os.path.join('models',f"{MODEL}_{FOLD}.pkl"))
        df = df[cols]
        preds = clf.predict_proba(df)[:, 1]

        if FOLD == 0 :
            predictions = preds
        else:
            predictions += preds

    predictions /= 5 

    sub = pd.DataFrame(np.column_stack((test_idx,predictions)),columns=['id','target'])
    return sub


if __name__== '__main__':

    submission = predict()
    submission.id = submission.id.astype(int)
    submission.to_csv(f'models/{MODEL}.csv',index=False)