pyflowml 1.1.0

AutoML library for the full ML lifecycle — train, evaluate, visualise, and save

🚀 PyFlowML

A completely automated, zero-code Machine Learning library that handles your entire data lifecycle instantly from your terminal.

---

⚡ Why PyFlowML?

Stop manually importing Scikit-Learn tools, cleaning datasets, writing loops, and crashing your Jupyter Notebooks. PyFlowML automatically:

1. Optimizes your dataset memory (saving up to 80% RAM).
2. Cleans text outliers, NaNs, and duplicates automatically.
3. Spawns an intelligently parallelized AutoML Engine.
4. Searches across XGBoost, LightGBM, Random Forests, SVM, etc.
5. Generates a stunning dark-mode Visual Dashboard!

---

🛠️ Installation

Simply install the heavily optimized library via pip:

pip install pyflowml

---

🔥 Quickstart: The Magic CLI

The easiest way to use PyFlowML is right from your terminal without writing a single line of code! Just navigate to the folder with your CSV data and type:

pyflowml

Our beautiful interactive menu will guide you through picking your target column, selecting a time budget, and automating the rest!

Example Terminal Output:

  ✔  Loaded  2,000 rows × 17 columns
  🔍  Profiling dataset…
  📦  Optimising memory… Memory: 1.3 MB → 0.2 MB (saved 87%)
  🧹  Cleaning data…
  🤖  Training AutoML models (budget=60s)…
  ✅  Best: KNN | f1=0.5098
  📈  Generating visualisation dashboard…

---

💻 Zero-Boilerplate Code (Pro Mode)

If you strictly want to integrate PyFlowML into your Python backend or Jupyter Notebooks, it's as simple as three lines of code:

AutoML Classification

import pandas as pd
from pyflowml.models.auto import AutoClassifier

# 1. Load Data
df = pd.read_csv("my_dataset.csv")
X_train = df.drop(columns=["target"])
y_train = df["target"]

# 2. Launch your AutoML Engine!
engine = AutoClassifier(metric="f1", time_limit=60)
engine.fit(X_train, y_train)

# 3. View Results
print(f"🥇 Best Model: {engine.best_model_name_}")
engine.leaderboard()

(For Regression, simply swap AutoClassifier for AutoRegressor!)

---

🌟 Feature Breakdown

Feature	Description
🧠 AutoML Search	Safely threads 6+ state-of-the-art architectures without deadlocking
📊 Intelligent Profiler	Detects classification vs regression & profiles correlation/skewness
⚙️ Smart Pipeline	Auto Label-Encoding & OneHot features instantly
📦 Memory Optimizer	Downcasts int64/float64 seamlessly for massive datasets
⏱️ Time Budget	time_limit=60 caps the search — models predicted to exceed the remaining budget are skipped, so training stays close to the limit
📈 One-Figure Dashboard	Dark-themed ROC, Confusion Matrix, & Leaderboards plotted entirely in one frame!
💾 Safe Versioning	Instantly saves your .pkl and .json metadata on successful train

---

🤝 Contributing

We welcome contributions! Have an idea to make PyFlowML faster? Open an Issue or submit a Pull Request.

📄 License

This open-source project is heavily protected under the MIT License.