Tim Holy
This eliminates the inconsistency in `c_i = 0` vs `c_L
Consider [HS62](https://github.com/JuliaSmoothOptimizers/OptimizationProblems.jl/blob/65bda82047287f09d261b86c968256eb21ae8df0/src/PureJuMP/hs62.jl), for which the constraint is ```julia x[1] + x[2] + x[3] - 1 == 0 ``` and the initial value is `x0 = [0.7, 0.2, 0.1]` which sums...
Clustal Omega writes files like this: ``` # STOCKHOLM 1.0 #=GS Q8BX79|reviewed|Probable DE G-protein coupled receptor 21|taxID:10090 ... Q8BX79|reviewed|Probable -----MNSTWDGN---------QSSHPFCLLAL-------GYLETVRFCL ... ``` That "|reviewed|Probable" ended up getting slurped into the...
Formatting.jl is now archived. EDIT: This also adds [compat] bounds on stdlibs, and makes Julia 1.6 the minimum supported version.
This gives the package a bit more formal structure. It's easier to recover the original AST: fragments are interwoven with non-fragment statements. Constants are extracted in an attempt to increase...
This eliminates the common failures observed in https://github.com/dmetivie/ExpectationMaximization.jl/issues/11#issuecomment-1817926977 The strategy is to stop updating a component when non-finite values are encountered. Obviously, this only helps if `robust=true`, but I'm a...
In cases of poor initialization, some components of the mixture may drop out. For example, let's create a 2-component mixture that is very poorly initialized: ```julia julia> X = randn(10);...
This is related to recent work on handling control-flow properly. In brief, this PR contains an example for which our current approach fails and which seems to require a breaking...
`bodymethod` was the prototype for `Base.bodyfunction`. Now that is being updated for changes in lowering, so it's better to rely on it.
### Feature description For visualization of volumes that are complex or not very transparent, it would be nice to combine the advantages of `volume` and `volumeslices` through implementation of a...