Adapt / Bump to ManifoldsBase 1.0 (#437)

* adapt version. * A bit of work on typos. * Update metadata. * Fix tests. * Runs formatter. * Make sure all docs and tutorials run on the new versions. * Rename optimize! to getstarted and bump version. * Fix make.jl links and others in the docs that can be md links by now.
JuliaManifolds · Feb 10, 2025 · c0f89f3 · c0f89f3 · kellertuer · Feb 10, 2025
1 parent b0e5ac0
commit c0f89f3
Show file tree

Hide file tree

Showing 40 changed files with 159 additions and 120 deletions.
diff --git a/.gitignore b/.gitignore
@@ -18,7 +18,7 @@ tutorials/img
 docs/src/tutorials/*/
 docs/src/tutorials/*.md
 docs/.CondaPkg
-docs/src/tutorials/Optimize!_files
+docs/src/tutorials/getstarted_files
 docs/src/tutorials/*.html
 docs/src/changelog.md
 docs/styles/Google

diff --git a/.vale.ini b/.vale.ini
@@ -7,12 +7,16 @@ Packages = Google
 [formats]
 # code blocks with Julia in Markdown do not yet work well
 qmd = md
+jl = md
 
 [docs/src/*.md]
 BasedOnStyles = Vale, Google
 
 [docs/src/contributing.md]
-BasedOnStyles =
+BasedOnStyles = Vale, Google
+Google.Will = false ; given format and really with intend a _will_
+Google.Headings = false ; some might jeally ahabe [] in their headers
+Google.FirstPerson = false ; we pose a few contribution points as first-person questions
 
 [Changelog.md, CONTRIBUTING.md]
 BasedOnStyles = Vale, Google
@@ -39,12 +43,14 @@ TokenIgnores = \$(.+)\$,\[.+?\]\(@(ref|id|cite).+?\),`.+`,``.*``,\s{4}.+\n
 Google.Units = false #wto ignore formats= for now.
 TokenIgnores = \$(.+)\$,\[.+?\]\(@(ref|id|cite).+?\),`.+`,``.*``,\s{4}.+\n
 
-[tutorials/*.md] ; actually .qmd for the first, second autogenerated
+[tutorials/*.qmd] ; actually .qmd for the first, second autogenerated
 BasedOnStyles = Vale, Google
 ; ignore (1) math (2) ref and cite keys (3) code in docs (4) math in docs (5,6) indented blocks
 TokenIgnores = (\$+[^\n$]+\$+)
 Google.We = false # For tutorials we want to address the user directly.
 
-[docs/src/tutorials/*.md]
- ; ignore since they are derived files
-BasedOnStyles =
+[docs/src/tutorials/*.md] ; actually .qmd for the first, second autogenerated
+BasedOnStyles = Vale, Google
+; ignore (1) math (2) ref and cite keys (3) code in docs (4) math in docs (5,6) indented blocks
+TokenIgnores = (\$+[^\n$]+\$+)
+Google.We = false # For tutorials we want to address the user directly.
diff --git a/.zenodo.json b/.zenodo.json
@@ -25,6 +25,11 @@
             "name": "Riemer, Tom-Christian",
             "type": "ProjectMember"
         },
+        {
+            "affiliation": "NTNU Trondheim",
+            "name": "Oddsen, Sander Engen",
+            "type": "ProjectMember"
+        },
         {
             "name": "Schilly, Harald",
             "type": "Other"

diff --git a/Changelog.md b/Changelog.md
@@ -5,7 +5,13 @@ All notable Changes to the Julia package `Manopt.jl` will be documented in this
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [0.5.5] Januaey 4, 2025
+## [0.5.6] February 10, 2025
+
+### Changed
+
+* bump dependencies of all JuliaManifolds ecosystem packages to be consistent with ManifoldsBase 1.0
+
+## [0.5.5] January 4, 2025
 
 ### Added
 
@@ -122,7 +128,7 @@ In general we introduce a few factories, that avoid having to pass the manifold
   * the previous `stabilize=true` is now set with `(project!)=embed_project!` in general,
     and if the manifold is represented by points in the embedding, like the sphere, `(project!)=project!` suffices
   * the new default is `(project!)=copyto!`, so by default no projection/stabilization is performed.
-* the positional argument `p` (usually the last or the third to last if subsolvers existed) has been moved to a keyword argument `p=` in all State constructors
+* the positional argument `p` (usually the last or the third to last if sub solvers existed) has been moved to a keyword argument `p=` in all State constructors
 * in `NelderMeadState` the `population` moved from positional to keyword argument as well,
 * the way to initialise sub solvers in the solver states has been unified In the new variant
   * the `sub_problem` is always a positional argument; namely the last one

diff --git a/Project.toml b/Project.toml
@@ -1,7 +1,7 @@
 name = "Manopt"
 uuid = "0fc0a36d-df90-57f3-8f93-d78a9fc72bb5"
 authors = ["Ronny Bergmann <[email protected]>"]
-version = "0.5.5"
+version = "0.5.6"
 
 [deps]
 ColorSchemes = "35d6a980-a343-548e-a6ea-1d62b119f2f4"
@@ -51,7 +51,7 @@ LineSearches = "7.2.0"
 LinearAlgebra = "1.10"
 ManifoldDiff = "0.3.8, 0.4"
 Manifolds = "0.9.11, 0.10"
-ManifoldsBase = "0.15.18"
+ManifoldsBase = "0.15.18, 1.0"
 ManoptExamples = "0.1.10"
 Markdown = "1.10"
 Plots = "1.30"

diff --git a/Readme.md b/Readme.md
@@ -36,7 +36,7 @@ In Julia you can get started by just typing
 using Pkg; Pkg.add("Manopt");
 ```
 
-and then checkout the [Get started: optimize!](https://manoptjl.org/stable/tutorials/Optimize/) tutorial.
+and then checkout the [🏔️ Get started with Manopt.jl](https://manoptjl.org/stable/tutorials/getstarted/) tutorial.
 
 ## Related packages
 

diff --git a/docs/Project.toml b/docs/Project.toml
@@ -40,9 +40,9 @@ JuMP = "1"
 LRUCache = "1"
 LineSearches = "7"
 Literate = "2"
-Manifolds = "0.8.81, 0.9, 0.10"
-ManifoldsBase = "0.14.12, 0.15"
-Manopt = "0.4, 0.5"
+Manifolds = "0.10"
+ManifoldsBase = "1"
+Manopt = "0.5"
 Plots = "1"
 QuadraticModels = "0.9.6"
 RecursiveArrayTools = "2, 3"

diff --git a/docs/make.jl b/docs/make.jl
@@ -21,7 +21,7 @@ Arguments
   Then you can spare time in the rendering by not passing this argument.
   If quarto is not run, some tutorials are generated as empty files, since they
   are referenced from within the documentation.
-  These are currently `Optimize.md` and `ImplementOwnManifold.md`.
+  These are currently `getstarted.md` and `ImplementOwnManifold.md`.
 """
     )
     exit(0)
@@ -35,7 +35,7 @@ tutorials_in_menu = !("--exclude-tutorials" ∈ ARGS)
 # (a) setup the tutorials menu – check whether all files exist
 tutorials_menu =
     "How to..." => [
-        "🏔️ Get started: optimize." => "tutorials/Optimize.md",
+        "🏔️ Get started with Manopt.jl" => "tutorials/getstarted.md",
         "Speedup using in-place computations" => "tutorials/InplaceGradient.md",
         "Use automatic differentiation" => "tutorials/AutomaticDifferentiation.md",
         "Define objectives in the embedding" => "tutorials/EmbeddingObjectives.md",
@@ -54,8 +54,9 @@ for (name, file) in tutorials_menu.second
         global all_tutorials_exist = false
         if !run_quarto
             @warn "Tutorial $name does not exist at $fn."
-            if (!isfile(fn)) &&
-                (endswith(file, "Optimize.md") || endswith(file, "ImplementOwnManifold.md"))
+            if (!isfile(fn)) && (
+                endswith(file, "getstarted.md") || endswith(file, "ImplementOwnManifold.md")
+            )
                 @warn "Generating empty file, since this tutorial is linked to from the documentation."
                 touch(fn)
             end

diff --git a/docs/src/about.md b/docs/src/about.md
@@ -28,8 +28,8 @@ to clone/fork the repository or open an issue.
 * [ExponentialFamilyProjection.jl](https://github.com/ReactiveBayes/ExponentialFamilyProjection.jl) package uses `Manopt.jl` to project arbitrary functions onto the closest exponential family distributions. The package also integrates with [`RxInfer.jl`](https://github.com/ReactiveBayes/RxInfer.jl) to enable Bayesian inference in a larger set of probabilistic models.
 * [Caesar.jl](https://github.com/JuliaRobotics/Caesar.jl) within non-Gaussian factor graph inference algorithms
 
-Is a package missing? [Open an issue](https://github.com/JuliaManifolds/Manopt.jl/issues/new)!
-It would be great to collect anything and anyone using Manopt.jl
+If you are missing a package, that uses `Manopt.jl`, please [open an issue](https://github.com/JuliaManifolds/Manopt.jl/issues/new).
+It would be great to collect anything and anyone using Manopt.jl in this list.
 
 ## Further packages
 

diff --git a/docs/src/index.md b/docs/src/index.md
@@ -20,7 +20,7 @@ or in other words: find the point ``p`` on the manifold, where ``f`` reaches its
 It belongs to the “Manopt family”, which includes [Manopt](https://manopt.org) (Matlab) and [pymanopt.org](https://www.pymanopt.org/) (Python).
 
 If you want to delve right into `Manopt.jl` read the
-[🏔️ Get started: optimize.](tutorials/Optimize.md) tutorial.
+[🏔️ Get started with Manopt.jl](tutorials/getstarted.md) tutorial.
 
 `Manopt.jl` makes it easy to use an algorithm for your favourite
 manifold as well as a manifold for your favourite algorithm. It already provides
@@ -94,7 +94,7 @@ The notation in the documentation aims to follow the same [notation](https://jul
 ### Visualization
 
 To visualize and interpret results, `Manopt.jl` aims to provide both easy plot functions as well as [exports](helpers/exports.md). Furthermore a system to get [debug](plans/debug.md) during the iterations of an algorithms as well as [record](plans/record.md) capabilities, for example to record a specified tuple of values per iteration, most prominently [`RecordCost`](@ref) and
-[`RecordIterate`](@ref). Take a look at the [🏔️ Get started: optimize.](tutorials/Optimize.md) tutorial on how to easily activate this.
+[`RecordIterate`](@ref). Take a look at the [🏔️ Get started with Manopt.jl](tutorials/getstarted.md) tutorial on how to easily activate this.
 
 ## Literature
 

diff --git a/docs/src/solvers/index.md b/docs/src/solvers/index.md
@@ -98,7 +98,7 @@ For these you can use
 * [Steihaug-Toint Truncated Conjugate-Gradient Method](truncated_conjugate_gradient_descent.md) a solver for a constrained problem defined on a tangent space.
 
 
-## Alphabetical list List of algorithms
+## Alphabetical list of algorithms
 
 | Solver   | Function        | State   |
 |:---------|:----------------|:---------|
@@ -202,7 +202,7 @@ also use the third (lowest level) and just call
 solve!(problem, state)
 ```
 
-### Closed-form subsolvers
+### Closed-form sub solvers
 
 If a subsolver solution is available in closed form, `ClosedFormSubSolverState` is used to indicate that.
 

diff --git a/docs/src/tutorials/InplaceGradient.md b/docs/src/tutorials/InplaceGradient.md
@@ -5,7 +5,7 @@ When it comes to time critical operations, a main ingredient in Julia is given b
 mutating functions, that is those that compute in place without additional memory
 allocations. In the following, we illustrate how to do this with `Manopt.jl`.
 
-Let’s start with the same function as in [Get started: optimize!](https://manoptjl.org/stable/tutorials/Optimize!.html)
+Let’s start with the same function as in [🏔️ Get started with Manopt.jl](https://manoptjl.org/stable/tutorials/getstarted.html)
 and compute the mean of some points, only that here we use the sphere $\mathbb S^{30}$
 and $n=800$ points.
 
@@ -32,7 +32,7 @@ p[2] = 1.0
 data = [exp(M, p, σ * rand(M; vector_at=p)) for i in 1:n];
 ```
 
-## Classical Definition
+## Classical definition
 
 The variant from the previous tutorial defines a cost $f(x)$ and its gradient $\operatorname{grad}f(p)$
 ““”
@@ -58,26 +58,26 @@ We can also benchmark this as
 @benchmark gradient_descent($M, $f, $grad_f, $p0; stopping_criterion=$sc)
 ```
 
-    BenchmarkTools.Trial: 106 samples with 1 evaluation.
-     Range (min … max):  46.774 ms …  50.326 ms  ┊ GC (min … max): 2.31% … 2.47%
-     Time  (median):     47.207 ms               ┊ GC (median):    2.45%
-     Time  (mean ± σ):   47.364 ms ± 608.514 μs  ┊ GC (mean ± σ):  2.53% ± 0.25%
+    BenchmarkTools.Trial: 89 samples with 1 evaluation per sample.
+     Range (min … max):  52.976 ms … 104.222 ms  ┊ GC (min … max): 8.05% … 5.55%
+     Time  (median):     55.145 ms               ┊ GC (median):    9.99%
+     Time  (mean ± σ):   56.391 ms ±   6.102 ms  ┊ GC (mean ± σ):  9.92% ± 1.43%
 
-         ▄▇▅▇█▄▇                                                    
-      ▅▇▆████████▇▇▅▅▃▁▆▁▁▁▅▁▁▅▁▃▃▁▁▁▁▁▁▁▁▁▁▁▁▃▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▅ ▃
-      46.8 ms         Histogram: frequency by time         50.2 ms <
+        ▅██▅▃▁                                                      
+      ▅███████▁▅▇▅▁▅▁▁▅▅▁▁▁▅▅▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▅ ▁
+      53 ms         Histogram: log(frequency) by time      81.7 ms <
 
-     Memory estimate: 182.50 MiB, allocs estimate: 615822.
+     Memory estimate: 173.54 MiB, allocs estimate: 1167348.
 
-## In-place Computation of the Gradient
+## In-place computation of the gradient
 
 We can reduce the memory allocations by implementing the gradient to be evaluated in-place.
 We do this by using a [functor](https://docs.julialang.org/en/v1/manual/methods/#Function-like-objects).
 The motivation is twofold: on one hand, we want to avoid variables from the global scope,
 for example the manifold `M` or the `data`, being used within the function.
 Considering to do the same for more complicated cost functions might also be worth pursuing.
 
-Here, we store the data (as reference) and one introduce temporary memory in order to avoid
+Here, we store the data (as reference) and one introduce temporary memory to avoid
 reallocation of memory per `grad_distance` computation. We get
 
 ``` julia
@@ -116,16 +116,16 @@ We can again benchmark this
 ) setup = (m2 = deepcopy($p0))
 ```
 
-    BenchmarkTools.Trial: 176 samples with 1 evaluation.
-     Range (min … max):  27.358 ms … 84.206 ms  ┊ GC (min … max): 0.00% … 0.00%
-     Time  (median):     27.768 ms              ┊ GC (median):    0.00%
-     Time  (mean ± σ):   28.504 ms ±  4.338 ms  ┊ GC (mean ± σ):  0.60% ± 1.96%
+    BenchmarkTools.Trial: 130 samples with 1 evaluation per sample.
+     Range (min … max):  36.646 ms … 64.781 ms  ┊ GC (min … max): 0.00% … 0.00%
+     Time  (median):     37.559 ms              ┊ GC (median):    0.00%
+     Time  (mean ± σ):   38.658 ms ±  3.904 ms  ┊ GC (mean ± σ):  0.73% ± 2.68%
 
-        ▂█▇▂ ▂                                                     
-      ▆▇████▆█▆▆▄▄▃▄▄▃▃▃▁▃▃▃▃▃▃▃▃▃▄▃▃▃▃▃▃▁▃▁▁▃▁▁▁▁▁▁▃▃▁▁▃▃▁▁▁▁▃▃▃ ▃
-      27.4 ms         Histogram: frequency by time        31.4 ms <
+      ██▅▅▄▂▁ ▂                                                    
+      ███████▁██▁▅▁▁▁▅▁▁▁▁▅▅▅▁▁▁▅▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▅▁▁▁▁▁▁▁▁▁▅ ▅
+      36.6 ms      Histogram: log(frequency) by time        61 ms <
 
-     Memory estimate: 3.83 MiB, allocs estimate: 5797.
+     Memory estimate: 3.59 MiB, allocs estimate: 6863.
 
 which is faster by about a factor of 2 compared to the first solver-call.
 Note that the results `m1` and `m2` are of course the same.
@@ -134,7 +134,7 @@ Note that the results `m1` and `m2` are of course the same.
 distance(M, m1, m2)
 ```
 
-    2.4669338186126805e-17
+    4.8317610992693745e-11
 
 ## Technical details
 
@@ -146,21 +146,24 @@ Pkg.status()
 ```
 
     Status `~/Repositories/Julia/Manopt.jl/tutorials/Project.toml`
-      [6e4b80f9] BenchmarkTools v1.5.0
-      [5ae59095] Colors v0.12.11
-      [31c24e10] Distributions v0.25.108
-      [26cc04aa] FiniteDifferences v0.12.31
-      [7073ff75] IJulia v1.24.2
+      [47edcb42] ADTypes v1.13.0
+      [6e4b80f9] BenchmarkTools v1.6.0
+    ⌃ [5ae59095] Colors v0.12.11
+      [31c24e10] Distributions v0.25.117
+      [26cc04aa] FiniteDifferences v0.12.32
+      [7073ff75] IJulia v1.26.0
       [8ac3fa9e] LRUCache v1.6.1
-      [af67fdf4] ManifoldDiff v0.3.10
-      [1cead3c2] Manifolds v0.9.18
-      [3362f125] ManifoldsBase v0.15.10
-      [0fc0a36d] Manopt v0.4.63 `..`
-      [91a5bcdd] Plots v1.40.4
+      [af67fdf4] ManifoldDiff v0.4.2
+      [1cead3c2] Manifolds v0.10.13
+      [3362f125] ManifoldsBase v1.0.1
+      [0fc0a36d] Manopt v0.5.5 `..`
+      [91a5bcdd] Plots v1.40.9
+      [731186ca] RecursiveArrayTools v3.29.0
+    Info Packages marked with ⌃ have new versions available and may be upgradable.
 
 ``` julia
 using Dates
 now()
 ```
 
-    2024-05-26T13:52:05.613
+    2025-02-10T13:22:51.002
diff --git a/docs/styles/config/vocabularies/Manopt/accept.txt b/docs/styles/config/vocabularies/Manopt/accept.txt
@@ -23,6 +23,7 @@ Cartis
 canonicalization
 canonicalized
 Constantin
+[Cc]ubics
 Dai
 deactivatable
 Diepeveen
@@ -76,9 +77,11 @@ Munkvold
 [Mm]ead
 [Nn]elder
 Nesterov
+Nesterovs
 Newton
 nonmonotone
 nonpositive
+[Nn]onsmooth
 [Pp]arametrising
 Parametrising
 [Pp]ock
@@ -110,9 +113,11 @@ Stephansen
 Stokkenes
 [Ss]ubdifferential
 [Ss]ubgradient
+[Ss]ubgradients
 subsampled
 [Ss]ubsolver
 summand
+summands
 superlinear
 supertype
 th

diff --git a/ext/ManoptLineSearchesExt.jl b/ext/ManoptLineSearchesExt.jl
@@ -39,7 +39,7 @@ function (cs::Manopt.LineSearchesStepsize)(
     end
     function ϕdϕ(α)
         # TODO: optimize?
-        retract!(M, p_tmp, p, η, α, cs.retraction_method)
+        ManifoldsBase.retract_fused!(M, p_tmp, p, η, α, cs.retraction_method)
         get_gradient!(mp, X_tmp, p_tmp)
         vector_transport_to!(M, Y_tmp, p, η, p_tmp, cs.vector_transport_method)
         phi = f(M, p_tmp)

diff --git a/src/Manopt.jl b/src/Manopt.jl
@@ -53,6 +53,7 @@ using ManifoldDiff:
     riemannian_gradient!,
     riemannian_Hessian,
     riemannian_Hessian!
+using ManifoldsBase
 using ManifoldsBase:
     AbstractBasis,
     AbstractDecoratorManifold,

diff --git a/src/helpers/checks.jl b/src/helpers/checks.jl
@@ -50,7 +50,7 @@ function check_differential(
     #
     T = exp10.(log_range)
     # points `p_i` to evaluate the error function at
-    points = map(t -> retract(M, p, Xn, t, retraction_method), T)
+    points = map(t -> ManifoldsBase.retract_fused(M, p, Xn, t, retraction_method), T)
     costs = [F(M, pi) for pi in points]
     # linearized
     linearized = map(t -> F(M, p) + t * dF(M, p, Xn), T)
@@ -297,7 +297,7 @@ function check_Hessian(
     #
     T = exp10.(log_range)
     # points `p_i` to evaluate error function at
-    points = map(t -> retract(M, p, X_n, t, retraction_method), T)
+    points = map(t -> ManifoldsBase.retract_fused(M, p, X_n, t, retraction_method), T)
     # corresponding costs
     costs = [f(M, pi) for pi in points]
     # linearized