mlplt 2.0.5

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

mlpilot 🚀

The complete, unbreakable machine learning toolkit — from raw data to production in seconds.

---

🌟 The Vision: "Unbreakable & Honest Data Science"

Most Machine Learning libraries fail when things get messy—or worse, they give you "perfect" results that fall apart in production. mlpilot is built for the real world. It is a self-healing, zero-config environment that ensures your models are both Unbreakable and Statistically Honest.

• Leakage Guard (v2.0.2): Advanced proxy detection. Automatically drops features with >0.98 correlation to the target to prevent "cheating" and ensure realistic hackathon winning scores.
• Self-Healing 2.0: Real-time syntax correction and "Hallucination Immunity" for AI-generated code.
• Global Encoding Shield: Revolutionary protection against UnicodeEncodeError on legacy Windows terminals.
• Silent Production: Absolute suppression of 3rd-party technical noise (HuggingFace, Transformers, etc.).

---

⚡ Quick Start: The "Omni-Pipeline"

Execute a professional, honest ML workflow in 10 lines of code. For a detailed step-by-step guide, see our 🚀 Hands-On Tutorial (Titanic).

import mlpilot as ml
import seaborn as sns

# 1. Load data
try:
    df = sns.load_dataset('titanic')
except:
    import pandas as pd
    df = pd.read_csv("your_data.csv")

# 2. The Unbreakable Pipeline
clean = ml.clean(df, target='survived')              # Hardens data & DROPS LEAKY PROXIES (like 'alive')
X, y  = ml.features(clean.df, target='survived')     # Zero-leakage engineering
blitz = ml.baseline(X, y)                            # Verified model tournament
tuned = ml.tune(blitz.best_name, X, y, time_budget=30) # Smart hyperparameter tuning

# 3. Defensible Verdict (Realistic accuracy, not 1.0)
print(f"Winner: {blitz.best_name} | Score: {blitz.best_score:.4f}")

---

📋 Module Encyclopedia

🧬 Phase 1: Data Foundations & Cleansing

• ml.analyze(df): 12-section SmartEDA report with quality scoring.
• ml.clean(df): The ultimate hardening tool. Handles nulls, outliers, dtypes, duplicates, and Data Leakage Detection.
• ml.validate(df, schema): Verifies your data against a strictly inferred industrial schema.

🧪 Phase 2: The Predictive Core

• ml.features(df): Leakage-safe fit_transform() engine. Statistics are learned from training data ONLY.
• ml.baseline(X, y): A high-speed model tournament. Compares 12+ model families to find the best baseline.
• ml.tune(name, X, y): Budget-aware hyperparameter tuning using a local optimization engine.
• ml.evaluate(model, X, y): Generates a 5-metric technical report with confusion matrices.

🔍 Phase 3: Specialized Insight

• ml.explain(model, X): Professional interpretability (SHAP). Tells you exactly why a prediction was made.
• ml.balance(df): Fixes class imbalance using intelligent oversampling (SMOTE).
• ml.forecast(series): TimeSense — Advanced multi-model time-series forecasting.

---

🧠 Advanced: The "Unbreakable" Engine

Smart Leakage Guard (v2.0.2)

In the real world, datasets often include "cheating" columns (proxies) that are direct synonyms for the target. mlpilot v2.0.2 introduces the Leakage Guard. It calculates feature-target correlation in real-time and drops variables with > 0.98 correlation. This prevents "Fake Perfect Scores" and ensures your hackathon results are trustworthy and defensible.

AST-Healing & Hallucination Immunity

mlpilot scans every line of AI-generated code against Python standards. It automatically purges hallucinations and redirects file-load errors to data already in memory, ensuring zero crashes during Natural Language interactions.

The Encoding Shield

Windows legacy terminals (CP1252) often crash when a library prints emojis. mlpilot features a Global Encoding Shield that automatically sanitizes all output for your specific terminal in real-time.

---

📦 Installation & Environments

Standard Build (v2.0.4):

pip install mlplt==2.0.4

GitHub Repository: https://github.com/Anannya-Vyas/mlplt-my-python-library-

---

🤝 Contributing & Support

MLPilot is a professional-grade machine learning toolkit. We welcome bug reports and collaborations.

• Developer: Anannya Vyas
• Email: [e-mail]
• GitHub: Anannya-Vyas/mlplt-my-python-library-

Current Version: v2.0.4 (The "Public Professional" Release)

MIT © 2026 mlpilot