- Changed metadata for repeated experiments to retain number of repetitions, added the index of each result - Eagerly expanded parameter ranges in labs to avoid issues with repeated traversal of, for example, ranges
0.13.3
- Updated some docstrings that were misleading - Tightened-up type checking for variable results argument - Re-introduced the use of Dill by default
0.13.2
- Fixed a bug in adding repeated experiments to notebooks
0.13.1
- Got rid of the chunking as there seems to be a race condition in ipyparallel or ZeroMQ. A small delay fixes it (for now) - Changed the way repeated experiments work, from returning a list of results dicts to returning a single one whose results were a list of results dicts (so embedding the list one level down) - Corresponding changes to summary experiments - Added more summary information to summary experiments - Removed the default use of dill in ClusterLab - ADded documentation for ExperimentCombinator
0.12.1
- Chunked downloading of pending results from clusters to work around some undocumented limits in some database backends - Added slightly improved error handling in summary experiments
0.11.2
- Maintenance release changing the build system slightly to allow for auto-generation of documentation on readthedocs.io