The Linux Foundation Projects
Skip to main content

Linux Foundation Energy has announced the 1.10 release series of Power Grid Model (PGM), including support for columnar data and a generic branch component type. PGM is a high-performance Python/C++ library for steady-state distribution power system analysis.

Details of all changes in this release follow, and the release notes are also available.

Highlights

  • ✅ Added native support for columnar data formats (see #548 and the documentation).
  • ✅ Added a generic branch component type (see #729#752 and the documentation).
  • ✅ Update data IDs are now optional in certain cases (see #548 and the documentation).
  • 💥 Multiple bug fixes (e.g., see #670#671#672#690#703).
  • 💥 Explicitly forbid certain cases of unsupported usages (see below for details, and #774 and #784).
  • 🚀 Improved numeric stability of source injection calculation (see #458).
  • 🚀 Various improvements to the user experience (e.g., see #668#674 and #697).
  • 🚀 Improvements and fixes to the documentation.
  • 🎉 Various improvements to the internal implementation and compiler support.

Support for Python 3.9 was dropped

As agreed upon during the last Power Grid Model meet-up, support for Python 3.9 was dropped. Please refer to our Release Strategy on details.

Explicitly disallowing unsupported behavior

The following changes affect previously undefined or unsupported behavior. Migration strategies are provided below.

  • The power_grid_model.core library was renamed to power_grid_model._core (see #787). The core module was never intended to be directly used by users as it contains the core of the power_grid_model implementation. The same is true for several functions in the modules in power_grid_model.validation (see #822).
    • Migration: if you do use the power_grid_model.core module in your workflow, please migrate to other functions exposed in the public modules in power_grid_model that are not prefixed with a leading underscore (_). Idem for usage of functions in power_grid_model.validation.
  • Providing data types (dtypes) that are not consistent with ones internally used in the Python library is now explicitly forbidden (see #811). This behavior was already undefined.

Please contact us on the issue board or discussion board if you find missing features or need help otherwise.

What’s Changed

Full Changelogv1.9.0...v1.10.0