Releases: baban9/timeseriesflow
v1.1.0
TimeSeriesFlow v0.2.0
TimeSeriesFlow v0.2.0
TimeSeriesFlow v0.1.0
TimeSeriesFlow v0.1.0
First public release of TimeSeriesFlow and AdaptiveForecast.
Install
From git (PyPI publish planned for v0.2):
pip install git+https://github.com/baban9/timeseriesflow.git@v0.1.0
Development install:
git clone https://github.com/baban9/timeseriesflow.git
cd timeseriesflow
git checkout v0.1.0
pip install -e ".[dev]"
Requires Python 3.10+.
Quick start
from timeseriesflow import EntityContext, entity_flow
@entity_flow(entity_key="device_id", time_key="timestamp")
def process_device(df, ctx: EntityContext) -> dict[str, object]:
return {"device_id": ctx.entity_id, "rows": len(df)}
result = process_device.run(df)
print(result.outputs)
Production pipeline:
tsflow run examples/basic_pipeline.py
Highlights
TimeSeriesFlow: @entity_flow, EntityRunner, checkpoints, tsflow CLI
AdaptiveForecast: profile-aware architecture selection (7 recipes)
Combined workflow: per-entity profiling in examples/advise_and_process.py
Golden path docs: recommended API for new projects
Documentation
Combined workflow (TS + AdaptiveForecast)
Full changelog
See CHANGELOG.md.