mlplt 0.1.0

The complete ML toolkit — EDA, cleaning, training, explainability, deployment

mlpilot 🚀

The Complete Python ML Toolkit — Killing the Boilerplate.

mlpilot is the definitive "all-in-one" library for machine learning practitioners. What currently takes 40 hours of repetitive coding, mlpilot delivers in 5 minutes. No dependency hell. No boilerplate. Just results.

---

📢 A Note from the Author

Hi! I'm Anannya Vyas, a student developer. 🎓

I didn't build mlpilot just to add another library to the pile. I built it because I was tired of writing the same 200 lines of null-handling, scaling, and correlation plotting for every single project. I wanted a library that "thinks" like a data scientist—making smart defaults while giving you full override control.

This project is my graduation milestone. It’s a journey from learning how to build a package to implementing local-first AI intelligence. Whether you are a student like me or a professional looking to move faster, I hope mlpilot saves you from the "boilerplate tax."

Special thanks to my teacher Lovnish Verma for the inspiration and guidance.

---

🚀 The mlpilot Advantage

mlpilot isn't just a library; it's a replacement for half your requirements.txt.

Feature	mlpilot	Profiling	PyCaret	SHAP	Fairlearn
Smart EDA Report	✅	✅	❌	❌	❌
Auto-Cleaning	✅	❌	Partial	❌	❌
Model Leaderboard	✅	❌	✅	❌	❌
Bias & Fairness Audit	✅	❌	❌	❌	✅
AI Data Analyst	✅	❌	❌	❌	❌
API Deployment	✅	❌	❌	❌	❌

---

🛠️ Prerequisites & Installation

AI Features (Recommended)

mlpilot uses local-first AI to keep your data private.

1. Install Ollama.
2. Run ollama pull llama3.2 (or llama3).
3. Keep the Ollama app running to use ml.analyst() and ml.explain_data().

Installation

# Core installation
pip install mlplt

# Full suite (includes AI, NLP, and Deployment tools)
pip install mlplt[full]

---

⚡ Quick Start: The "Instant Analyst" Workflow

Solve a churn prediction problem from scratch in under 60 seconds of coding.

import mlpilot as ml
import pandas as pd

# 1. Load your data
df = pd.read_csv("customer_data.csv")

# 2. Let the AI explain it in 3 sentences
ml.explain_data(df)

# 3. Auto-generate a premium dashboard
viz = ml.visualize(df, target="churn")
viz.to_html("insights.html")

# 4. Clean and Engineer (leakage-safe)
clean = ml.clean(df, target="churn")
feats = ml.features(clean.df, target="churn")

# 5. Training & Auditing
X_train, X_test, y_train, y_test = ml.split(feats, test_size=0.2)
base = ml.baseline(X_train, y_train)

# 6. Audit for Bias & Stability
audit = ml.audit(base.best_model, X_test, y_test)
audit.print_summary()

# 7. Generate your final Project Story
ml.story([df, base, audit])

---

🧠 The AI Brain: Intelligent ML

🔍 AI Analyst (ml.analyst)

Ask questions about your data in plain English. No more searching for pandas syntax.

"Which city has the highest average churn, and is it correlated with revenue?" mlpilot generates the code, shows it to you for review, and runs it instantly.

⚖️ MLAudit (ml.audit)

Go beyond accuracy. Audit your models for:

• Technical Stability: How does the model react to noise?
• Social Bias: Check for Demographic Parity and Equalized Odds to ensure your model is fair to all groups.

📝 DataStory (ml.story)

Automatically synthesizes your EDA, Cleaning, Training, and Auditing results into a narrative report. Perfect for presentations!

---

📖 Complete API Reference

Module	Function	Description
EDA	ml.analyze(df)	12-section comprehensive report.
Visualizer	ml.visualize(df)	Intelligent auto-plotting dashboard.
Cleaner	ml.clean(df)	Auto-fix nulls, outliers, and dtypes.
Validator	ml.validate(df)	Check for leakage, drift, and schema errors.
FeatureForge	ml.features(df)	One-line encoding and scaling pipeline.
Baseline	ml.baseline(X,y)	15+ model comparison leaderboard.
Explainer	ml.explain(model,X)	Global and local SHAP explanations.
LaunchPad	ml.deploy(model)	FastAPI + Docker in 5 minutes.

---

📁 Project Structure

mlpilot/
├── ai/          # Analyst, Audit, Story (The Brain)
├── clean/       # Cleaner, Strategies, Diff
├── eda/         # Analyzer, Visualizer, Plots
├── train/       # Baseline, HyperX, Evaluate
├── validate/    # Schema, Drift, Checks
└── deploy/      # LaunchPad, FastAPI

---

🤝 Contributing

Contributions are welcome! This is a student project, and I'd love to see how you improve it.

1. Fork the repo.
2. Install dev dependencies: pip install -e ".[dev]"
3. Run tests: pytest tests/

---

📄 License

Licensed under the MIT License. Use it, fork it, make it yours.

---

⭐ Show Your Support

If mlpilot helped you crush your project boilerplate:

• ⭐ Star the repo on GitHub.
• 📢 Share it with your classmates.
• 🐛 Report bugs to help me learn and improve.

Built with ❤️ for the ML community.