Revisit BrainCollection for 0.6.1 against stabilized member contracts
- Dominant language
- Python
- Stars
- 130
- Forks
- 48
- Avg merge
- 6h 42m
- Merged PRs (30d)
- 34
Description
BrainCollection and its exclusive support were deferred from 0.6.0 to 0.6.1 while every other
planned 0.6.0 feature was retained.
Preserve the implementation reference at 5b0080be and
development/specs/braincollection.md. Before reintroduction, reconcile member ownership,
cached/in-memory equivalence, fit/predict/contrast semantics, archive versus image export,
cache lifetime and memory/concurrency with the stabilized BrainData, Glm, Ridge and
DesignMatrix.
Do not revive historical pipeline abstractions or automatically copy the old execution layer.
Reuse proven shared internals. Bring a design and release-scope proposal for approval before
implementation. No start date.
Note: development/execution-model.md preserves the deferred 0.6.1 execution design; it is not
an active 0.6.0 subsystem.
Contributor guide
Research direction
Start with development/specs/braincollection.md and implementation reference 5b0080be, then compare the stated contracts with BrainData, Glm, Ridge and DesignMatrix. Prepare a design and 0.6.1 release-scope proposal covering ownership, equivalence, semantics, export, caching, memory and concurrency; the issue is done when that proposal is approved.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100