octopyx 1.0.0

A modular Python library for automated data preprocessing, feature selection, model evaluation, and report generation.

                   OCTOPY

A Modular Python Library for Machine Learning Automation

Author : Sahil Kewat Version: 1.0.0 License: MIT Language: Python 3.12+

1. OVERVIEW

Octopy is a modular machine learning support library that automates data preprocessing, feature selection, model evaluation, and report generation. It is designed to simplify the ML workflow by providing a collection of plug-and-play Python modules.

The library is built for developers, data scientists, and researchers who want a fast, reproducible, and well-structured way to prepare, train, and analyze ML models.

2. MODULES INCLUDED

Octopy contains the following main Python modules:

1. pipeline.py - Handles the creation and execution of ML pipelines.

2. prep.py - Cleans and preprocesses raw datasets.

3. selector.py - Performs feature selection and ranking.

4. smart_eda.py - Generates visual and statistical exploratory data analysis.

5. report.py - Loads a trained model, evaluates it, and creates JSON reports.

6. MODULE DETAILS AND FUNCTIONS

---

A. pipeline.py

Purpose: Automates machine learning pipeline creation from preprocessing to model training and saving.

Key Functions: • build_pipeline(model, preprocess_steps) - Combines preprocessing and model into a single pipeline. • train_pipeline(pipeline, X_train, y_train) - Fits the pipeline to training data. • save_pipeline(pipeline, filename) - Saves the pipeline to disk using pickle/joblib. • load_pipeline(filename) - Loads an existing pipeline for inference or retraining.

---

B. prep.py

Purpose: Cleans, encodes, and scales data for model training.

Key Functions: • handle_missing_values(df) - Fills or removes missing values automatically. • encode_categorical(df) - Converts categorical variables into numeric form. • scale_features(df) - Applies standard or min-max scaling to numeric features. • preprocess_data(df) - Combines all preprocessing operations into one function.

---

C. selector.py

Purpose: Selects the most important features for model training.

Key Functions: • select_k_best_features(X, y, k) - Selects top k features based on statistical tests. • feature_importance(model, X, y) - Displays or returns feature importance scores. • recursive_feature_elimination(model, X, y) - Uses RFE to iteratively eliminate less important features.

---

D. smart_eda.py

Purpose: Automates exploratory data analysis (EDA) and visualization.

Key Functions: • describe_data(df) - Provides summary statistics of the dataset. • plot_distributions(df) - Plots histograms and distribution graphs for numeric features. • correlation_heatmap(df) - Displays correlation between numeric variables. • detect_outliers(df) - Identifies outliers using z-score or IQR method.

---

E. report.py

Purpose: Evaluates trained ML models and generates automated reports.

Key Functions: • load_model(model_path) - Loads a trained model from .pkl, .sav, or joblib files. • load_test_data(x_path, y_path) - Loads X_test and y_test from CSV files. • evaluate_model(model, X_test, y_test) - Computes MAE, MSE, RMSE, and R² metrics. • extract_hyperparameters(model) - Extracts key hyperparameters from scikit-learn compatible models. • generate_report(model_path, x_test_path=None, y_test_path=None) - Generates a structured JSON report containing model name, hyperparameters, and evaluation metrics.

4. INSTALLATION

5. Clone or download the repository: got clone https://github.com/Sahilkewat80085/OctoPy.git cd Octopy

6. Install dependencies and package: pip install -e .