Skip to content
Documentation

Installation

To use the Volue Data & Forecasts Timeseries Python library you need to have Python installed. The library supports Python 3.9 or newer, and is tested against Python 3.9, 3.10, 3.11, 3.12, 3.13, and 3.14. See Supported Python versions.

Install or update the latest 3.x version with pip:

Terminal window
pip install -U "volue-insight-timeseries~=3.0"

Always limit the library to one major version. ~=3.0 installs the latest 3.x release, but not 4.0 or later. We follow Semantic Versioning, so a new major version can contain breaking changes. Upgrade to it only after you have checked the release notes and tested your code.

In a requirements.txt file:

volue-insight-timeseries~=3.0 # Good - latest 3.x
volue-insight-timeseries # Bad - unpinned, may break on the next major version

We want the library to work on as many Python versions as we can, so you can upgrade Python on your own schedule:

  • No upper limit. The library does not cap the Python version. A new Python release is expected to work, and we add it to our tests when it comes out. The tested versions are listed above and in the package classifiers on PyPI.
  • Dropping a version needs a major release. We drop support for a Python version only when a new feature needs a newer Python, and only for versions that have reached end of life. Dropping a version is a breaking change, so it always comes in a new major version of the library, and the release notes say so.
  • pip picks a release that fits. pip installs the newest release that supports your Python version. If you pin the major version as shown above, a dropped Python version never changes the release you get within that major version.

pandas is optional. The plain install above does not install it. These functions need pandas:

  • TS.to_pandas() and TS.from_pandas()
  • TS.sum(), TS.mean() and TS.median()
  • PairedTS.to_pandas()
  • util.tags_to_DF()

To use them, install the library with the pandas extra:

Terminal window
pip install -U "volue-insight-timeseries[pandas]~=3.0"

If pandas is already installed, the plain install is enough. The library needs pandas 1.5.0 or newer. Without the pandas extra, pip does not check the pandas version, so an older pandas raises an ImportError when you call one of these functions.

Each TS object keeps its data in TS.points, a list of [timestamp_ms, value] pairs. You can use this list directly, or load it into any other dataframe library, for example polars.

© 2026 Volue.Version v0.6.38017 November 2026