v0.7.7

Unfold v0.7.7

Diff since v0.7.6

• new structured backbone for the solver. Added ~6-7 new solvers. Split up the solver into prepare & solve functions that can be chained / modified easily. Added a reference document for exactly this.
• major bug: likelihoodratiotest was calculating wrong results (it was calculating a wrong likelihood). This lead to a breaking change in API, where now the data need to be provided as well.
• Bumped JLD to 0.5 - this means one cannot save files now and load it in older Unfold versions, but old unfold version files can still be loaded
• several small docstring fixes
• new CUDAextension, allowing to provide CuArray to automatically choose a better solver (still needs to be tested more though)
• fit now has an argument that can be set fit=false to not actually run the ´fit!` function, to do it later
• benchmarks now available comparing ~7 solver types with and without GPU, some with and without Float32. Improvements of up to 10x of the SOA-GPU solver were achieved - fit a large model with 128 channels in wopping 8 seconds! Further tests are needed though to describe the accuracy drop that is achieved by solving X'Xb = X'y instead of Xb = y directly.

Merged pull requests:

• predict tutorial (#219) (@behinger)
• better fit docs (#221) (@behinger)
• docs: add maxvanmigem as a contributor for bug (#222) (@allcontributors[bot])
• update to JLD 0.5 (#224) (@behinger)
• docs: add ReneSkukies as a contributor for code, and ideas (#226) (@allcontributors[bot])
• new solver system (#227) (@behinger)
• better benchmark structure (#228) (@behinger)

Closed issues:

• documentation open issues (#195)
• LM overlap correction tutorial still uses old Unfold syntax (#199)
• Inconsistency in predict( ) input; better docstring (#216)
• Julia 1.11 crash (#223)
• slight differences between MixedModels.jl and Unfold.jl (#225)

docs-test-release

v0.7.6+docs1

Rename ci.yml to unittests.yml

v0.7.6d

This is only a release to test the "stable" documentation problem

v0.7.6c

this release is only to test the "stable" documenter problem

v0.7.6_zenodo

new release specifically to fix a bug with zenodo

v0.7.6

Unfold v0.7.6

Diff since v0.7.5

• bugfix for predict @ReneSkukies
• small documentation fixes
• fix Zenodo release metadata

Merged pull requests:

• Update juliacall_unfold.jl to include MNE Python warning (#211) (@carmenamme)
• docs: add carmenamme as a contributor for doc (#212) (@allcontributors[bot])
• Remove spaces between docstring and effect function (#215) (@jschepers)
• Update predict (#217) (@ReneSkukies)
• Grammar and readability for nonlinear_effects.jl (#218) (@vladdez)

v0.7.5

Unfold v0.7.5

Diff since v0.7.4

• improve tutorials
• a first fix for Term.jl output in Pluto.jl - current bug: no output via Documenter.jl

Merged pull requests:

• fixed/hacked pluto printing, doc fix (#210) (@behinger)

v0.7.4

Unfold v0.7.4

Diff since v0.7.3

• internal function get_basis_names now works for UnfoldLinearModel too

Merged pull requests:

• Update juliacall_unfold.jl (#206) (@behinger)
• add get_basis_names for massunivariate models (#208) (@behinger)

v0.7.3

Unfold v0.7.3

Diff since v0.7.2

• fixed major bug in UnfoldLinearModel introduced in 0.7.0. If you used for MassUnivariate models with Any this bug did not exist. But if you subset your data e.g. via [:fixation=>(@formula(0~1),times)] or similar then you have to recalculate your results! In this case the subsetting of the data/epochs went wrong due to a nasty, but very tiny bug. A SubDataFrame was silently cast to a DataFrame, removingthe information which events belonged to which index.
• Added small feature to be able to call effects(Dict(),model) which adds a dummy effect to the dict, to be able to get the pure intercept.

Merged pull requests:

• emptyeffects (#198) (@behinger)
• fix mass univariate bug (#205) (@behinger)

Closed issues:

• Allow empty dict in effects (#167)

v0.7.2

Unfold v0.7.2

Diff since v0.7.1

• bugfix for missing data & krylov solver
• bugfix with non-integer event onsets #200
• new internal functionality _residuals now allows to calculate residuals directly cut to an event, not requiring to generate the whole y = X*b

Merged pull requests:

• fix200 (#201) (@behinger)
• fixMissing (#202) (@behinger)
• residual (#203) (@behinger)

Closed issues:

• Bug: basisfunction with non-integer onsets, Float instead of Int (#200)