Installation¶

LavenderTown is available on PyPI and can be installed using pip.

Basic Installation¶

pip install lavendertown

Optional Dependencies¶

LavenderTown supports several optional dependencies for extended functionality:

Polars Support¶

For better performance with large datasets:

pip install lavendertown[polars]

Ecosystem Integrations¶

Export rules to Pandera or Great Expectations:

pip install lavendertown[pandera]
pip install lavendertown[great_expectations]

Note: LavenderTown is compatible with both altair 4.x and 5.x. Installing Great Expectations will automatically install altair 4.x (which is compatible with LavenderTown).

Enhanced CLI¶

For the best CLI experience with beautiful terminal output:

pip install lavendertown[cli]

This includes: - Rich: Progress bars, formatted tables, and color-coded messages - python-dotenv: Configuration management via .env files - orjson: Fast JSON serialization (2-3x faster than standard library)

ML and Time-Series Features¶

For machine learning-based anomaly detection and time-series analysis:

pip install lavendertown[ml]          # PyOD + scikit-learn for 40+ ML anomaly detection algorithms
pip install lavendertown[timeseries]  # Ruptures + statsmodels + tsfresh for time-series analysis

Phase 6 Features: - ML (lavendertown[ml]): Includes PyOD library with 40+ additional ML anomaly detection algorithms (ABOD, CBLOF, HBOS, KNN, MCD, PCA, and more) beyond scikit-learn's Isolation Forest, LOF, and One-Class SVM - Time-Series (lavendertown[timeseries]): Includes Ruptures library for change point detection in time-series data

Phase 7 Features (v0.7.0): - Time-Series (lavendertown[timeseries]): Now also includes tsfresh for advanced time-series feature extraction (700+ features)

Data Profiling¶

Generate comprehensive HTML profiling reports:

pip install lavendertown[profiling]   # ydata-profiling for advanced data profiling

Parquet Export¶

Export findings to efficient Parquet format:

pip install lavendertown[parquet]     # PyArrow for Parquet export/import

Statistical Tests¶

Enhanced drift detection with statistical tests:

pip install lavendertown[stats]       # scipy.stats for Kolmogorov-Smirnov and chi-square tests

Phase 7 Features (v0.7.0)¶

Advanced integrations for enhanced visualizations, UI components, and database storage:

pip install lavendertown[plotly]      # Plotly for interactive visualizations (zoom, pan, 3D charts)
pip install lavendertown[ui]           # Streamlit Extras for enhanced UI components (metric cards, badges)
pip install lavendertown[database]     # SQLAlchemy for database backend (SQLite and PostgreSQL)

Phase 7 Features: - Plotly (lavendertown[plotly]): Interactive visualization backend with zoom, pan, and 3D charts - UI (lavendertown[ui]): Streamlit Extras integration for enhanced metric cards, badges, and improved layouts - Database (lavendertown[database]): SQLAlchemy support for scalable collaboration features with SQLite and PostgreSQL

All Optional Dependencies¶

Install everything at once:

pip install lavendertown[all]

This includes: - Polars support - Pandera and Great Expectations exports - Enhanced CLI (Rich, python-dotenv, orjson, Typer) - ML anomaly detection (PyOD + scikit-learn) - Time-series analysis (Ruptures + statsmodels + tsfresh) - Data profiling (ydata-profiling) - Parquet export (PyArrow) - Statistical tests (scipy.stats) - Plotly interactive visualizations - Streamlit Extras UI components - SQLAlchemy database backend

Development Installation¶

For contributing to LavenderTown:

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

This installs LavenderTown in editable mode with development dependencies including pytest, mypy, ruff, and black.

Requirements¶

Python 3.10 or higher
Streamlit 1.28.0 or higher
Pandas 1.5.0 or higher
Altair 4.2.1 or higher (compatible with both 4.x and 5.x)

Verification¶

After installation, verify that LavenderTown is working correctly:

from lavendertown import Inspector
import pandas as pd

# Create a simple test DataFrame
df = pd.DataFrame({"value": [1, 2, 3, None, 5]})

# Create inspector
inspector = Inspector(df)

# Get findings
findings = inspector.detect()
print(f"Found {len(findings)} data quality issues")

If this runs without errors, LavenderTown is installed correctly!