v0.15.0

DimensionalData v0.15.0

Diff since v0.14.0

Breaking changes:

• DimDataset is now DimStack. But there is a fallback const for const DimDataset = DimStack
• set will no longer set metadata with a Dict. It must be isa Metadata
• set will no longer set name with Nothing. it must be NoName()
• Minor changes to the output of view when no index is given, or when a matrix is indexed linearly. The original behaviour can be considered a bug.

Breaking changes to unexported methods:

• sel2indices/_sel2indices and dims2indices/_dims2indices have been refactored

Closed issues:

• Fix indexing with unordered vectors (#94)
• Index dimensions with Selectors (#129)
• @dim will raise an error if only import part of DimensionalData (#167)
• Clarify object equality (#182)
• Dataset is vague. Rename to Stack (#201)
• Would be nice if dimensions printed their length (#206)
• docstring for sel2indices and extending it (#207)

Merged pull requests:

• Fix typo in @dim example (#200) (@ali-ramadhan)
• fix dim macro when only import was used (#203) (@rafaqz)
• Metadata (#210) (@rafaqz)
• Renames and refactors (#211) (@rafaqz)
• use selectors in dims, and clean up dim methods (#212) (@rafaqz)

v0.14.0

DimensionalData v0.14.0

Contains breaking changes to reverse: both index and array are now reversed together. This shouldn't affect general use as IndexOrder sorts this out either way, but indexing directly into the dimension values will now be in reverse.

Diff since v0.13.3

Closed issues:

• Move to using only Dim{:X} style dimensions (#179)
• Should reverse also reverse the dimension? (#190)
• ArgumentError: mode step 1.0 does not match index step -1.0 (#194)
• Indexing with end fails in simple cases (#195)
• Indexing along non-existent dimensions should produce a warning (#196)

Merged pull requests:

• reverse reverses array and index, not relation (#191) (@rafaqz)
• cleanup tests (#192) (@rafaqz)
• MassInstallAction: Install the TagBot workflow on this repository (#193) (@rafaqz)
• warn when a indexing dimension is not found in dims (#198) (@rafaqz)

v0.13.3

DimensionalData v0.13.3

Diff since v0.13.2

Merged pull requests:

• reverse AbstractSampled, not Sampled (#188) (@rafaqz)

v0.13.2

DimensionalData v0.13.2

Diff since v0.13.1

Closed issues:

• add span for radian units and other coherent irregular spans (#132)
• Allow tupled dims argument in mean (and other Statistics reduction functions)? (#186)

Merged pull requests:

• Test more syntax for reducing dims methods, include any and all and refactor (#187) (@rafaqz)

v0.13.1

DimensionalData v0.13.1

Diff since v0.13.0

Closed issues:

• Incremental compilation may be fatally broken unique(A::AbstractDimArray) (#184)

Merged pull requests:

• Add diff method for DimArrays (#183) (@singularitti)
• fix method duplication warning in unique (#185) (@rafaqz)

v0.13.0

DimensionalData v0.13.0

Breaking Changes

This release contains a number of breaking changes.

• The dim macro is slightly changed - shortname is overkill and basically never different from the type name in practice.

Now the macro is limited to the form:

@dim Ti TimeDim "Time" # supertype is TimeDim, name is "Time"
@dim Ti "Time"         # supertype is Dimension, name is "Time"
@dim Ti TimeDim        # supertype is TimeDim, nameis "Ti"
@dim Ti                # supertype is Dimension, name is "Ti"

• The shortname method is also removed.

• The name field of Dimension and DimArray is now a Symbol - not a string. This is for simplicity and interop with other interfaces like Tables.jl where keys are Symbol. Constructors should still work, but warn if you use a String for name. The name method for all AbstractDimArray is also expected to return a Symbol. The @dim macro still uses a String - as a symbol is no different to a type in macro args, and doesn't help us with dispatch.

• Array types that defined dims but are not <: AbstractDimArray will have lost a number of Base methods. This is basically no longer supported - there are too many method ambiguities for it to be practical to dispatch on dims (I doubt this was actually in use anyway).

Diff since v0.12.1

Closed issues:

• Datasets? (#80)
• otherdims is type unstable (#164)
• Should val & data be merged? (#169)
• Is there a way to set a dim to a different type? (#170)
• Separate Forward and Reverse by purpose (#171)
• DimDataset (#172)
• Should A in abstract type AbstractDimArray{T,N,D<:Tuple,A} <: AbstractArray{T,N} end be constrained? (#174)
• What is A here? A bug? (#175)
• Does val supports val(A::AbstractDimArray, dims::Tuple) => Tuple now? (#177)

Merged pull requests:

• allow vector indexing (#156) (@rafaqz)
• add a tables interface (#162) (@rafaqz)
• add DimDataset (#173) (@rafaqz)
• Fix a DimArray constructor (#176) (@singularitti)
• Label (#181) (@rafaqz)

v0.12.1

DimensionalData v0.12.1

Diff since v0.12.0

Closed issues:

• What does the dim name Ti stand for? (#154)
• How to make a stable iterator over slices of otherdims? (#165)
• dimwise errors for dimension with eltype Date (#166)

Merged pull requests:

• Update course.md (#155) (@musm)
• Fix a typo in docs of dimwise (#159) (@singularitti)
• Update docs of dimnum, hasdim (#160) (@singularitti)
• Update examples of otherdims & commondims (#161) (@singularitti)
• Small fixes to README.md (#163) (@singularitti)
• Fix dimwise (#168) (@rafaqz)

v0.12.0

DimensionalData v0.12.0

Diff since v0.11.2

Closed issues:

• Implement * (or other unicode symbol) as in xarray (#41)
• Document dims types and dispatch for devs (#90)
• handle rot* methods (#138)
• How to select every "singular splice" of a given dimension? (#140)
• getindex with Tuple? (#143)
• Fix broadcast over wrapped arrays like CuArray (#146)

Merged pull requests:

• Remove auto atol for At (#145) (@rafaqz)
• Add Aqua.jl quality assurance (#148) (@rafaqz)
• handle rotations (#149) (@rafaqz)
• use DimArray and AbstractDimArray everywhere (#150) (@rafaqz)
• Dimwise function broadcast (#151) (@rafaqz)
• Refactor: @generated, selectors and docs (#153) (@rafaqz)

v0.11.2

DimensionalData v0.11.2

Diff since v0.11.1

Merged pull requests:

• Auto -> AutoMode, and other small fixes and refactors (#144) (@rafaqz)

v0.11.1

DimensionalData v0.11.1

Diff since v0.11.0

Merged pull requests:

• add missing Statistics for mean/std (#141) (@Datseris)
• minor updates and fixes for plotting (#142) (@rafaqz)