De-Time

De-Time is workflow-oriented research software for reproducible time-series decomposition. It gives you one public package, one decomposition contract, and one docs surface across univariate, multivariate, native-backed, and machine-facing workflows.

What it is

• A canonical Python package with import path detime.
• A stable decompose() entrypoint plus DecompositionConfig and DecompResult.
• A documentation set centered on practical workflows rather than benchmark scoreboards.
• A package whose flagship methods are SSA, STD, STDR, and MSSA.
• A machine-facing surface with schemas, recommendations, and a minimal MCP server.

What it is not

• Not a new decomposition algorithm.
• Not a benchmark-paper artifact disguised as a library.
• Not a replacement for every specialized upstream implementation.
• Not a promise that every wrapper has the same maturity as the flagship path.

Start here

• Install for package installation and extras.
• Quickstart for the first successful Python and CLI runs.
• Choose a Method for picking a starting workflow.
• ML Workflows for the package's machine-learning-facing use cases.
• Methods for method family details.
• Migration from tsdecomp if you are upgrading existing code.

Package boundary

This repository ships the software package itself. Companion benchmark artifacts live in the separate systems-mechanobiology/de-time-bench repository. The main package no longer exposes benchmark orchestration, leaderboard helpers, or benchmark-derived methods.

The legacy tsdecomp import and CLI still resolve to De-Time, but only as a deprecated compatibility alias.

Visual reference