stackexchange-tutorial/tutorial/shared.py

import os

data_dir = os.path.join(os.path.dirname(__file__), '../data/')
outputs_dir = os.path.join(os.path.dirname(__file__), '../outputs/')
raw_data = os.path.join(data_dir, 'CrossValidated-Posts.csv')
train_data = os.path.join(data_dir, 'train-raw.csv')
test_data = os.path.join(data_dir, 'test-raw.csv')
train_processed = os.path.join(data_dir, 'train-processed.pkl')
test_processed = os.path.join(data_dir, 'test-processed.pkl')
classifier_pkl = os.path.join(outputs_dir, 'classifier.pkl')
pipeline_pkl = os.path.join(outputs_dir, 'pipeline.pkl')

col_id = 'Id'
col_text = 'Text'
col_title = 'Title'
col_body = 'Body'
col_tags = 'Tags'
col_label = 'IsTaggedML'
extra_feature_cols = ['Score','ViewCount','AnswerCount','CommentCount','FavoriteCount']
text_cols = [col_title, col_body]
text_len_col = 'Text_Len'
all_raw_cols = ['Id','Title','Body','Tags','CreationDate','Score','ViewCount','AnswerCount','CommentCount','FavoriteCount','IsTaggedML']


def save(obj, path):
    import pickle
    with open(path, 'wb') as f:
        pickle.dump(obj, f)


def load(path):
    import pickle
    with open(path, 'rb') as f:
        return pickle.load(f)


def load_labels(path=train_data):
    import pandas as pd
    return pd.read_csv(path, usecols=[col_label])[col_label]


def compute_metrics(clf, X, y, prefix):
    from sklearn.metrics import accuracy_score, f1_score, recall_score, precision_score, roc_auc_score, balanced_accuracy_score, auc, precision_recall_curve
    X = X.sparse.to_coo()
    preds = clf.predict(X)
    probas = clf.predict_proba(X)[:,1]
    pr_curve = precision_recall_curve(y, probas)
    return {
        f"{prefix}_accuracy_score": accuracy_score(y, preds),
        f"{prefix}_f1_score": f1_score(y, preds),
        f"{prefix}_recall_score": recall_score(y, preds),
        f"{prefix}_precision_score": precision_score(y, preds),
        f"{prefix}_roc_auc_score": roc_auc_score(y, probas),
        f"{prefix}_pr_auc_score": auc(pr_curve[1], pr_curve[0]),
        f"{prefix}_balanced_accuracy_score": balanced_accuracy_score(y, preds)
    }


def load_data_and_labels(processed_path):
    X = load(processed_path)
    y = X[col_label]
    X = X.drop(columns=[col_id,col_label])
    return X, y