🚀 MLPilot

---

✨ Features

• 📂 Automatic CSV & Parquet loading
• 🤖 Regression & Classification detection
• 📊 Statistical EDA & data quality reports
• 🧹 Missing value, duplicate & outlier handling
• ⚙️ Feature engineering & preprocessing pipeline
• 🏆 Compare 14+ ML models with Cross Validation
• 🎯 Bayesian hyperparameter tuning using Optuna
• 🔍 SHAP feature importance & explainability
• 📦 Export complete inference pipeline as .joblib

---

⚡ Installation

pip install mlpilot

For development:

python -m pip install -e ".[dev]"

---

🚀 Quick Start

Run the complete pipeline:

mlpilot run --data dataset.csv

Specify the target column:

mlpilot run --data dataset.csv --target price --task regression

Custom output directory:

mlpilot run --data dataset.csv --output outputs/

---

🔄 Pipeline Workflow

Raw Dataset
     │
     ▼
 Load Data
     │
     ▼
 Statistical EDA
     │
     ▼
 Data Cleaning
     │
     ▼
 Feature Engineering
     │
     ▼
 Train/Test Split
     │
     ▼
 Preprocessing
     │
     ▼
 Compare Models
     │
     ▼
 Optuna Tuning
     │
     ▼
 SHAP Explainability
     │
     ▼
 Model Evaluation
     │
     ▼
 Export Pipeline (.joblib)

---

🐍 Python API

from ml_pilot import PipelineRunner
from ml_pilot.config import load_config

config = load_config()
runner = PipelineRunner(config)

context = runner.run(
    data_path="dataset.csv",
    target="price"
)

print(context.best_model_name)
print(context.metrics)

---

📁 Generated Artifacts

outputs/
├── mlpilot_pipeline.joblib
├── leaderboard.csv
├── metrics.json
├── model_comparison.json
├── feature_importances.json
├── run_metadata.json
├── serving_schema.json
├── predict_snippet.py
└── shap/
    ├── shap_summary.png
    ├── shap_dependence.png
    └── shap_waterfall.png

---

🧠 Supported Models

Category	Models
Linear	Ridge, Lasso, ElasticNet, Linear Regression
Tree	Decision Tree, Random Forest, Extra Trees
Boosting	Gradient Boosting, HistGradientBoosting
Instance	KNN, SVR
Neural	MLP Regressor / Classifier

---

📂 Project Structure

mlpilot/
├── src/ml_pilot/
│   ├── cli.py
│   ├── config/
│   ├── core/
│   ├── stages/
│   └── utils/
├── tests/
├── examples/
└── pyproject.toml

---

🛠 Tech Stack

• Python 3.12+
• Scikit-learn
• Pandas & NumPy
• Optuna
• SHAP
• Typer + Rich
• Joblib
• Plotly

---

📄 License

Released under the MIT License.

#\x00 \x00M\x00L\x00_\x00P\x00i\x00l\x00o\x00t\x00 \x00 \x00