spotanomaly2 2.1.1

Streamlined anomaly detection system for time series data

SpotAnomaly

Streamlined anomaly detection system for time series data with dual training paradigms (batch and online learning).

Requirements

• Python 3.10 or higher
• uv package manager (recommended)

Installation

Using uv (Recommended)

Install uv:

# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh

# Homebrew
brew install uv

# Windows
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

Set up the project:

uv sync

This command automatically installs Python 3.11, creates a virtual environment, and installs all dependencies.

Using pip

python3.11 -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install -e .

Usage

Configuration

Choose your training paradigm in config/default.yaml:

model:
  paradigm: "batch"  # or "online"

• batch: Traditional batch training with sklearn/LightGBM (faster, trains once)
• online: Streaming online learning with river (incremental updates)

Running the Pipeline

Run all steps a first time:

uv run python -m src all

Run all steps but skip the download:

uv run python -m src all --skip-download

Or run individual steps:

uv run python -m src process  # Process raw data
uv run python -m src train    # Train forecasting models
uv run python -m src detect   # Detect anomalies

Run predictions with a pre-trained model:

uv run python -m src all --predict-only --model 20251229_172126

Viewing the result

You can explore, chart, and analyze the processed and modeled data interactively in notebooks/analyze_combined.ipynb.

Running live mode

uv run python -m src live --interval 5

Then open the live report

Development

Install with development dependencies:

uv sync --all-extras