v1.4.1

What's Changed

• version bump by @ahalev in #59
• add shape reward key to microgrid step to only calc when necessary by @ahalev in #61
• grid marginal cost fix by @ahalev in #60
• reward shaping updates by @ahalev in #62
• log net load by @ahalev in #63
• net load position in obs space by @ahalev in #64
• return np arrays for actions in mpc for consistency w action space by @ahalev in #65
• envs updates by @ahalev in #66
• fix bug where pre-normalized actions are not clipped by @ahalev in #67
• reward shaping funcs by @ahalev in #68
• docs updates by @ahalev in #74
• reward shaping by @ahalev in #75
• reward shaping updates by @ahalev in #77
• update genset status to bool in pymgrid25 files by @ahalev in #78
• remove analyze_shaper.py by @ahalev in #79
• efficiency by @ahalev in #80
• efficiency by @ahalev in #81
• convert action fix by @ahalev in #82
• fix absolute action clipping by @ahalev in #83
• revert to analyze_shaper.py removal by @ahalev in #88
• test net load by @ahalev in #87
• revert + cherry pick by @ahalev in #89
• add DecayCycleTransitionModel by @ahalev in #91
• add dynamic clip test by @ahalev in #92
• add reset to decay model by @ahalev in #93
• check for increments of 1 by @ahalev in #94
• fix decay model issue and add tests by @ahalev in #95
• genset slack by @ahalev in #96
• docs updates by @ahalev in #21
• deprecate Microgrid.run in favor of Microgrid.step by @ahalev in #86
• update errors and finish Microgrid.run deprecation by @ahalev in #99
• deprecate new by @ahalev in #100
• convert fixes by @ahalev in #102
• standardization shaper by @ahalev in #103

Full Changelog: v1.4.0...v1.4.1

v1.4.0

Lots of small fixes and updates; this release is just a bit overdue.

What's New

• Net load environment: A continuous RL environment where actions are scaled to the net load in each step. See #49, #50, #54, #57
• Clipping functionality: MicrogridSpace and ModuleSpace now have .clip() methods that allow clipping to the bounds of the spaces. See #25, #51
• RescaleShaper: a reward shaping function that scales the original reward by a constant factor. See #58

What's Changed

• MPC API compatibility with environments in #22, #24
• Allow normalization to bounds that are different than [-1, 1] in #19, #20
• Add verbose equality to give a verbose explanation of microgrid/module differences in #28, #34, #35
• Logging updates. Actions, both normalized and denormalized, are logged in environments in #31, #32, #33
• Various bug fixes
• Various documentation and testing updates

Full Changelog: v1.3.0...v1.4.0

v1.3.0

What's Changed

• slow tests by @ahalev in #1
• battery tests by @ahalev in #2
• mpc: update cycling through solvers by @ahalev in #4
• flex module logging by @ahalev in #3
• dry run context manager by @ahalev in #5
• fix num iter in rbc by @ahalev in #7
• custom batteries by @ahalev in #6
• setattrs updates by @ahalev in #9
• manifest update by @ahalev in #11
• battery by @ahalev in #10
• to python microgrid by @ahalev in #12
• version bump by @ahalev in #13

Full Changelog: https://github.com/ahalev/python-microgrid/commits/v1.3.0