Skip to content

Latest commit

 

History

History
164 lines (102 loc) · 4.24 KB

CHANGELOG.md

File metadata and controls

164 lines (102 loc) · 4.24 KB

Change Log

Version 0.6.0

Features

  • Added a mechanism for the server to provide a list of available options (with associated types).
  • Created a general implementation of the explicit discipline client for OpenMDAO. The client creates an OpenMDAO ExplicitComponent which can be added to any OpenMDAO model.

Bug Fixes

  • None

Version 0.5.3

Features

  • None

Bug Fixes

  • Added missing function arguments to explicit discipline.

Version 0.5.2

Features

  • None

Bug Fixes

  • Lowered the dependency versions (they were far too stringent and new)
  • Change PyPI deployment to source only. It is not practical to distribute a platform-specific wheel. The wheel must be platform-specific, because gRPC has C underpinnings.

Version 0.5.1

Features

  • Transitioned away from setuptools and setup.py to a pyproject.toml and poetry-based package.
  • gRPC and protobuf stubs are now automatically compiled during installation.
  • Added test coverage report generation that is uploaded to coveralls.
  • Added action to upload to PyPI when a release is published.

Bug Fixes

  • Lowered the dependency versions (they were far too stringent and new)
  • Change PyPI deployment to source only. It is not practical to distribute a platform-specific wheel. The wheel must be platform-specific, because gRPC has C underpinnings.

Version 0.5.0

  • yanked due to source distribution issues. All features present in 0.5.1

Version 0.4.0

Features

  • General documentation updates.

Bug Fixes

  • None

Version 0.3.0

This release is one of the biggest changes to the code to date. It contains a fundamental reorganization and adds a number of features. Notably, it adds unit and integration testing of almost all the code.

Features

  • Reorganized codebase to reduce code duplication. The clients and servers now use base classes.
  • Protobuf/gRPC files are now generated at build time and not committed to the repository. This requires grpc-tools and protoletariat to be installed. See the readme for details.
  • Added a change log file to the repository.
  • Updated API and logic to conform with newer Philote definition.
  • Added unit testing suite.
  • Added integration test suite (based on examples).
  • Completed implicit discipline functionality and testing.
  • Fixed unit tests for GetVariableDefinitions and GetPartialsDefinitions.
  • Added edge case handling for partials of variables that are scalar.

Bug Fixes

  • Corrected the preallocate_inputs function for the implicit case to resolve variable copy issues.
  • Fixed typo in discrete input parsing.
  • Moved to setup.py, as setuptools is still in beta for pyproject.toml.
  • Added jupyter book for documentation.
  • Added a quick start guide.

Version 0.2.1

This is purely a bugfix release. Thanks to Alex Xu for finding these bugs and fixing them.

Features

  • None

Bug Fixes

  • Fixed bug that prevented proper chunking of array data
  • Fixed flat view of arrays used during variable transfer

Version 0.2.0

This version augments the Philote MDO version to 0.3.0.

Features

  • Moved to Philote version 0.3.0
  • Renamed RPC function from Compute to Functions for Philote 0.3.0 compatibility
  • Renamed RPC function from ComputePartials to Gradient for Philote 0.3.0 compatibility

Bug Fixes

  • Added flattened views for the ndarrays received. The previous version would error for n-dimensional arrays, as the slices would not work unless the array was flattened.

Version 0.1.0

Initial release of the Philote MDO Python bindings. Includes working remote explicit disciplines. Only the generic API currently works, so there is no framework support for OpenMDAO or CSDL.

Features

  • Implemented a remote explicit discipline analysis server API.
  • Implemented a corresponding client for explicit analyses.
  • Added a simple parabaloid example to demonstrate the server/client in action.

Bug Fixes

  • None, as this is the first release.

Note

All versions starting with a 0 as the major version number should be considered pre-release. While they may work in production environments, it is expected that bugs may surface and that several features are still missing. Because of this, the API may still change frequently before version 1.0.0 is released.