adamops 0.1.1

A comprehensive MLOps library for end-to-end machine learning workflows

🚀 AdamOps

AdamOps is a comprehensive MLOps library for end-to-end machine learning workflows. It provides a unified interface for data processing, model training, evaluation, deployment, and monitoring.

✨ Features

📊 Data Module (DataOps)

• Loaders: CSV, Excel, JSON, SQL, API, compressed files with auto-encoding detection
• Validators: Type validation, missing values, duplicates, shape, statistical checks
• Preprocessors: Missing value imputation, outlier handling, text cleaning
• Feature Engineering: Encoding, scaling, feature selection, auto feature generation
• Splitters: Train/test, time-series, K-Fold, stratified splitting

🤖 Model Module (ModelOps)

• Regression: Ridge, Lasso, ElasticNet, XGBoost, LightGBM
• Classification: Decision Tree, Gradient Boosting, XGBoost, LightGBM, Naive Bayes, KNN
• Clustering: K-Means, DBSCAN, Hierarchical, GMM
• Ensembles: Voting, Stacking, Blending, Weighted averaging
• AutoML: Model selection, hyperparameter tuning (Grid, Random, Bayesian)

📈 Evaluation Module

• Metrics: Classification, regression, and clustering metrics
• Visualization: Confusion matrices, ROC curves, feature importance plots
• Explainability: SHAP and LIME explanations
• Reports: HTML/PDF report generation

🚀 Deployment Module

• Exporters: ONNX, PMML, TFLite, CoreML
• APIs: FastAPI, Flask, Streamlit
• Containerization: Docker, Kubernetes
• Cloud: AWS, GCP, Azure

📡 Monitoring Module

• Drift Detection: Data and concept drift
• Performance Tracking: Model metrics over time
• Alerts: Performance degradation notifications
• Dashboards: Real-time monitoring dashboards

🔄 Pipelines Module

• Workflows: End-to-end ML workflows as DAGs
• Orchestration: Scheduling and pipeline execution

🛠️ Installation

Basic Installation

pip install adamops

Development Installation

git clone https://github.com/adamops/adamops.git
cd adamops
pip install -e ".[dev]"

Full Installation (all extras)

pip install adamops[all]

🚀 Quick Start

Data Loading

from adamops.data import loaders

# Load CSV with auto-encoding detection
df = loaders.load_csv("data.csv")

# Load from SQL database
df = loaders.load_sql("SELECT * FROM table", "sqlite:///database.db")

Data Validation

from adamops.data import validators

# Create validation report
report = validators.validate(df)
print(report.summary())

Data Preprocessing

from adamops.data import preprocessors

# Handle missing values
df = preprocessors.handle_missing(df, strategy="knn")

# Handle outliers
df = preprocessors.handle_outliers(df, method="iqr")

Feature Engineering

from adamops.data import feature_engineering

# Encode categorical variables
df = feature_engineering.encode(df, method="onehot", columns=["category"])

# Scale numerical features
df = feature_engineering.scale(df, method="standard", columns=["value"])

Model Training

from adamops.models import modelops

# Train a model
model = modelops.train(
    X_train, y_train,
    task="classification",
    algorithm="xgboost"
)

# Predict
predictions = model.predict(X_test)

AutoML

from adamops.models import automl

# Run AutoML
best_model = automl.run(
    X_train, y_train,
    task="classification",
    tuning="bayesian",
    time_limit=3600
)

Evaluation

from adamops.evaluation import metrics

# Compute metrics
results = metrics.evaluate(y_true, y_pred, task="classification")
print(results)

CLI Usage

# Train a model
adamops train --data data.csv --target y --algorithm xgboost

# Evaluate a model
adamops evaluate --model model.pkl --data test.csv

# Deploy as API
adamops deploy --model model.pkl --type api --port 8000

📁 Project Structure

adamops/
├── adamops/
│   ├── __init__.py
│   ├── cli.py
│   ├── data/
│   │   ├── loaders.py
│   │   ├── validators.py
│   │   ├── preprocessors.py
│   │   ├── feature_engineering.py
│   │   └── splitters.py
│   ├── models/
│   │   ├── modelops.py
│   │   ├── registry.py
│   │   ├── ensembles.py
│   │   └── automl.py
│   ├── evaluation/
│   │   ├── metrics.py
│   │   ├── visualization.py
│   │   ├── explainability.py
│   │   ├── comparison.py
│   │   └── reports.py
│   ├── deployment/
│   │   ├── exporters.py
│   │   ├── api.py
│   │   ├── containerize.py
│   │   └── cloud.py
│   ├── monitoring/
│   │   ├── drift.py
│   │   ├── performance.py
│   │   ├── alerts.py
│   │   └── dashboard.py
│   ├── pipelines/
│   │   ├── workflows.py
│   │   └── orchestrators.py
│   └── utils/
│       ├── config.py
│       ├── logging.py
│       └── helpers.py
├── tests/
├── examples/
├── docs/
├── setup.py
├── requirements.txt
└── README.md

📚 Documentation

Full documentation is available at https://adamops.readthedocs.io

🤝 Contributing

Contributions are welcome! Please see our Contributing Guide for details.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• scikit-learn team for their excellent ML library
• XGBoost and LightGBM teams for gradient boosting implementations
• SHAP and LIME teams for explainability tools
• The entire open-source ML community

---

Made with ❤️ by the AdamOps Team