Indexed arrays for performance and safety (#17)

* Fix varnames (no extra tuple) Remove test Manifest * Update Readme * Testing with typed dictionaries for cache * * Fix performance issue * Started working on storing legal values * Convenience constructors and safe_insert! * Formatting * Add exports, consistent return from safe_insertvar * Add tests for IndexedVarArray * Compatibility fix for julia < 1.8 * Update JuliaFormatter * Formatting fixes (not using VS Code) * Formatting * Improve test coverage * Fix clear_cache! * Drop slices compiletime for performance * Cleanup, delete unused functions + formatting * Fix corner case for tuple concatenation. Delete unused function _has_index * Formatting fix * Fix indexing on string with slicing * Switch names of inservar! for IndexedVarArrays Default to safe insertion * Cleanup, more informative function name * Updated benchmarks from talk * Formatting fixes * Prep for release Co-authored-by: Lars Hellemo <[email protected]>
sintefore · Sep 19, 2022 · 21cf913 · 21cf913
1 parent d2eb9e4
commit 21cf913
Show file tree

Hide file tree

Showing 13 changed files with 3,760 additions and 236 deletions.
diff --git a/.github/workflows/format_check.yml b/.github/workflows/format_check.yml
@@ -19,7 +19,7 @@ jobs:
         run: |
           using Pkg
           # If you update the version, also update the style guide docs.
-          Pkg.add(PackageSpec(name="JuliaFormatter", version="0.22.7"))
+          Pkg.add(PackageSpec(name="JuliaFormatter", version="1.0.8"))
           using JuliaFormatter
           format("."; verbose = true)
           out = String(read(Cmd(`git diff`)))

diff --git a/NEWS.md b/NEWS.md
@@ -1,6 +1,10 @@
 TimeStructures release notes
 ===================================
 
+Version 0.6.2 (2022-09-19)
+--------------------------
+* Add IndexedVarArray that checks for valid indices on insert and has improved performance
+
 Version 0.6.1 (2022-06-29)
 --------------------------
 * Release under MIT License

diff --git a/Project.toml b/Project.toml
@@ -1,11 +1,12 @@
 name = "SparseVariables"
 uuid = "2749762c-80ed-4b14-8f33-f0736679b02b"
 authors = ["Truls Flatberg <[email protected]>", "Lars Hellemo <[email protected]>"]
-version = "0.6.1"
+version = "0.6.2"
 
 [deps]
 Dictionaries = "85a47980-9c8c-11e8-2b9f-f7ca1fa99fb4"
 JuMP = "4076af6c-e467-56ae-b986-b466b2749572"
+LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 Requires = "ae029012-a4dd-5104-9daa-d747884805df"
 Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 
@@ -14,4 +15,4 @@ Dictionaries = "0.3"
 JuMP = "1"
 Requires = "1.3"
 Tables = "1.7"
-julia = "1.6"
+julia = "1.6"
diff --git a/README.md b/README.md
@@ -10,6 +10,13 @@ Watch the JuliaCon/JuMP-dev 2022 lightning talk and check out the [notebook with
 
 [![SparseVariables - Efficient sparse modelling with JuMP](https://img.youtube.com/vi/YuDvfZo9W5A/3.jpg)](https://youtu.be/YuDvfZo9W5A)
 
+2022-09: Updated benchmarks of time spent on model construction with different number of variables (see [benchmark notebook for details](benchmark/benchmark.jl)) with additional types `IndexedVarArray` (model_indexed) and `SparseAxisArray` (model_sparse_aa) on current julia master:
+
+![](benchmark/res.svg)
+
+Benchmarks with time spent on model construction with different level of sparsity:
+
+![](benchmark/sparsity.svg)
 
 ## Usage
 
@@ -57,6 +64,22 @@ end
 @constraint(m, sum(z[endswith("a"), iseven]) >= 1)
 ```
 
+## IndexedVarArray
+
+Use IndexedVarArrays to check for valid indices and to warn against duplicate 
+indices, as well as improved performance:
+
+```julia
+    w = IndexedVarArray(m, "w", (car=cars, year=years))
+    m[:w] = w
+
+    for c in cars, y in years
+        insertvar!(m, c, y)
+    end
+```
+
+
+
 ## Solution information
 
 SparseVariables.jl provides `SolutionTable` that supports the [Tables.jl](https://github.com/JuliaData/Tables.jl) interface, allowing