TimeNet

Find, build, and load time-series datasets through one shared format.

Time-series datasets use many incompatible layouts. TimeF gives them one storage format. TimeNet provides one API to find, build, download, and load them.

TimeNet handles the data layer. Model training, inference, and evaluation stay in your chosen framework.

Quick start

TimeNet needs Python 3.11 or newer.

uv add 'timenet[cli]'
uv tool install timenet-connectors
timenet-build build timenet/hello-world --out .timenet-registry

Load the local build with Python:

from timenet.client import TimeNet

dataset = TimeNet(".timenet-registry").load("timenet/hello-world")
dataset.describe()

signal = dataset.records[0].signals[0]
values = signal.to_numpy()

This example uses a local registry, so it works without registry credentials or network access.

See the Get started guide for installation options and a longer example.

How it fits together

A connector converts a raw source into a TimeF version. The version contains:

• manifest.json, which identifies the dataset and lists its files
• control.duckdb, which stores the hierarchy and relationships
• a Parquet or Zarr values plane, which stores Signal values

A registry stores immutable TimeF versions. The client reads a version without importing its connector.

The repository is a uv workspace with two packages:

Read the architecture guide for the complete design.

License

TimeNet uses the MIT License.

This license covers the TimeNet code, not the datasets that connectors fetch. Each dataset keeps its upstream license and access terms. Read the dataset licensing guide before redistribution.