Releases: zkat/big-brain
v0.12.0
v0.11.0
v0.10.0
This is my first pass at a significant API improvement. I iterated on it for a
while and this is what I settled on. I look forward to continuing to evolve
this API as I get more feedback and experience with it! Please let me know
what you think!
Breaking Changes
- thinker: stop cancelling actions when they go under Picker thresholds (4c72b3d1)
- stages: Strongly typed stages for BigBrainPlugin (65ca646e)
- builders: Blanket impls for ActionBuilder and ScorerBuilder when Clone (8bed75b5)
- api: rename attach and hide it from docs (6c64df4f)
- choice: return &Choice instead of cloning (a76dcbb6)
Features
- example: update thirst example (edecc4c9)
Bug Fixes
v0.9.0
v0.8.0
v0.7.0
0.6.0
0.5.0
Got a few goodies in this release, mainly focused around composite actions and
scorers, which were apparently broken.
Shout out again to @piedoomy for some of these
contributions!
Features
- actions: Add new Concurrently composite action (6c736374)
- evaluators: added inversed linear evaluator helper (#19) (f871d19e)
- scorers: Added WinningScorer composite scorer (#20) (748b30ae)
Breaking Changes
- scorers: Composite Scorers now all use
.push()
instead of a mixture of.push()
and.when()
. Please update any usages of composite scorers (63bad1fd
Bug Fixes
0.4.0
0.3.5
Previously, if a Picker
re-picked the same action, and that action had been
set to Success
or Failure
, it would just keep running the action in that
state until it was time to switch to a different one.
With this version, that behavior is changed, and Failure
and Success
actions that are re-picked will be respawned entirely (not even reused).
Cheers to @piedoomy on Discord for pointing out
this weird behavior!
Bug Fixes
- thinker: launch a new action when the current action is in an end state (80d23f2f)