repo_name
stringlengths
1
62
dataset
stringclasses
1 value
lang
stringclasses
11 values
pr_id
int64
1
20.1k
owner
stringlengths
2
34
reviewer
stringlengths
2
39
diff_hunk
stringlengths
15
262k
code_review_comment
stringlengths
1
99.6k
timefold-solver
github_2023
java
171
TimefoldAI
rsynek
@@ -38,23 +41,60 @@ public static HardMediumSoftScore parseScore(String scoreString) { } public static HardMediumSoftScore ofUninitialized(int initScore, int hardScore, int mediumScore, int softScore) { + if (initScore == 0) { + return of(hardScore, mediumScore, softScore); + } ...
How often is this method called to express a single-level score that is either -1, 0, or 1? This `if-else-if` structure surely does not come for free. Proposal: do this optimization only for single-level score factory methods, where the chance we hit the constant is much higher. (applies to all Score types)
timefold-solver
github_2023
others
158
TimefoldAI
rsynek
@@ -88,31 +88,23 @@ jobs: - name: Set micro snapshot version on the release branch run: | + git checkout -B ${{ github.event.inputs.releaseBranch }}-bump
```suggestion git checkout -B ${{ github.event.inputs.releaseBranch }}-bump-to-next-micro-version ```
timefold-solver
github_2023
others
158
TimefoldAI
rsynek
@@ -88,31 +88,23 @@ jobs: - name: Set micro snapshot version on the release branch run: | + git checkout -B ${{ github.event.inputs.releaseBranch }}-bump mvn -Dfull versions:set -DnewVersion=${{ github.event.inputs.nextMicroVersion }}-SNAPSHOT - find . -name 'pom.xml' | xa...
```suggestion git checkout -B ${{ github.event.inputs.releaseBranch }}-bump-to-next-minor-version ```
timefold-solver
github_2023
others
158
TimefoldAI
rsynek
@@ -88,31 +88,23 @@ jobs: - name: Set micro snapshot version on the release branch run: | + git checkout -B ${{ github.event.inputs.releaseBranch }}-bump mvn -Dfull versions:set -DnewVersion=${{ github.event.inputs.nextMicroVersion }}-SNAPSHOT - find . -name 'pom.xml' | xa...
```suggestion git push origin ${{ github.event.inputs.releaseBranch }}-bump-to-next-micro-version ```
timefold-solver
github_2023
others
158
TimefoldAI
rsynek
@@ -88,31 +88,23 @@ jobs: - name: Set micro snapshot version on the release branch run: | + git checkout -B ${{ github.event.inputs.releaseBranch }}-bump mvn -Dfull versions:set -DnewVersion=${{ github.event.inputs.nextMicroVersion }}-SNAPSHOT - find . -name 'pom.xml' | xa...
```suggestion gh pr create --reviewer triceo,ge0ffrey --base ${{ github.event.inputs.releaseBranch }} --head ${{ github.event.inputs.releaseBranch }}-bump-to-next-micro-version --title "chore: move to ${{ github.event.inputs.nextMicroVersion }}-SNAPSHOT" --body-file .github/workflows/release-pr-body-stable.m...
timefold-solver
github_2023
others
158
TimefoldAI
rsynek
@@ -88,31 +88,23 @@ jobs: - name: Set micro snapshot version on the release branch run: | + git checkout -B ${{ github.event.inputs.releaseBranch }}-bump mvn -Dfull versions:set -DnewVersion=${{ github.event.inputs.nextMicroVersion }}-SNAPSHOT - find . -name 'pom.xml' | xa...
```suggestion git push origin ${{ github.event.inputs.releaseBranch }}-bump-to-next-minor-version ```
timefold-solver
github_2023
others
158
TimefoldAI
rsynek
@@ -88,31 +88,23 @@ jobs: - name: Set micro snapshot version on the release branch run: | + git checkout -B ${{ github.event.inputs.releaseBranch }}-bump mvn -Dfull versions:set -DnewVersion=${{ github.event.inputs.nextMicroVersion }}-SNAPSHOT - find . -name 'pom.xml' | xa...
```suggestion gh pr create --reviewer triceo,ge0ffrey --base ${{ github.event.inputs.sourceBranch }} --head ${{ github.event.inputs.releaseBranch }}-bump-to-next-minor-version --title "chore: move to ${{ github.event.inputs.nextVersion }}-SNAPSHOT" --body-file .github/workflows/release-pr-body.md ```
timefold-solver
github_2023
others
156
TimefoldAI
rsynek
@@ -112,63 +113,77 @@ As you can see in the examples, most instances have a lot more possible solution Timefold supports several optimization algorithms to efficiently wade through that incredibly large number of possible solutions. Depending on the use case, some optimization algorithms perform better than others, b...
Should we clearly say that every feature apart from scalability features is included?
timefold-solver
github_2023
others
146
TimefoldAI
triceo
@@ -16,7 +16,7 @@ with https://timefold.ai[Timefold]'s constraint solving Artificial Intelligence You will build a REST application that optimizes a school timetable for students and teachers: -image::quickstart/school-timetabling/schoolTimetablingScreenshot.png[] +image::quickstart/school-timetabling/schoolTimeta...
Why do we keep and reference an *old* version of the app?
timefold-solver
github_2023
others
138
TimefoldAI
rsynek
@@ -46,32 +46,30 @@ jobs: distribution: 'temurin' cache: 'maven' + # Need Maven 3.9.0+ to recognize MAVEN_ARGS. + - name: Set up Maven + uses: stCarolas/setup-maven@v4.5 + with: + maven-version: 3.9.3 + # We skip tests in dry run, to make the process faste...
There is a dedicated GHA? Interesting.
timefold-solver
github_2023
others
138
TimefoldAI
rsynek
@@ -242,7 +242,7 @@ public class CloudBalance ... { [[cloudBalancingMainMethod]] == Run the cloud balancing Hello World -. xref:introduction/introduction.adoc#runTheExamplesInAnIDE[Download and configure the examples in your preferred IDE.] +. xref:introduction/introduction.adoc#buildFromSource[Download and configu...
Nice catch.
timefold-solver
github_2023
others
115
TimefoldAI
rsynek
@@ -1,16 +1,27 @@ -# TODO release XSD to timefold.ai/xsd/ -# TODO reset timefold-quickstarts stable branch -# TODO adjust website pom.yml release date and version -# TODO tag github issues with version name: Release on: workflow_dispatch: inputs: version: - description: 'Release version' + ...
Let's move these two before the "Next version..." parameters to define these branches before referring to them.
timefold-solver
github_2023
others
115
TimefoldAI
rsynek
@@ -19,79 +30,94 @@ jobs: - name: Checkout timefold-solver uses: actions/checkout@v3 with: - path: 'timefold-solver' - repository: 'TimefoldAI/timefold-solver' fetch-depth: 0 - - name: Checkout timefold-quickstarts - uses: actions/checkout@v3 - ...
Something for consideration: define some MAVEN_OPTIONS (?) env var to keep the `-no-transfer-progress --batch-mode` in a single place.
timefold-solver
github_2023
others
115
TimefoldAI
rsynek
@@ -19,79 +30,94 @@ jobs: - name: Checkout timefold-solver uses: actions/checkout@v3 with: - path: 'timefold-solver' - repository: 'TimefoldAI/timefold-solver' fetch-depth: 0 - - name: Checkout timefold-quickstarts - uses: actions/checkout@v3 - ...
In the last stage of the release, this branch should be deleted. Apart from just having a stale branch in the git repository, if you want to do a micro release from already existing release branch, there will likely be a conflict with the existing release-bump branch.
timefold-solver
github_2023
others
120
TimefoldAI
triceo
@@ -18,8 +18,7 @@ A _realistic contest_ is _an official, independent contest_: Realistic contests provide an objective comparison of Timefold with competitive software and academic research. -The source code of all these examples is available in the distribution zip under [path]_examples/sources_ -and also in git ...
Maybe link to Github instead?
timefold-solver
github_2023
others
108
TimefoldAI
rsynek
@@ -25,63 +25,73 @@ </constructionHeuristic> <localSearch> <unionMoveSelector> - <listChangeMoveSelector> - <valueSelector id="1"/> - <destinationSelector> - <nearbySelection> - <originValueSelector mimicSelectorRef="1"/> - <nearbyDistanceMeterClass>ai.timefo...
I see a lot of combinations got commented out as expected. However, this impacts: a) performance tests in timefold-examples (if they run these configurations) b) turtle tests (assuming there are some turtle tests)
timefold-solver
github_2023
java
102
TimefoldAI
rsynek
@@ -461,4 +468,27 @@ Constraint talkUndesiredRoomTags(ConstraintFactory factory) { .asConstraint(TALK_UNDESIRED_ROOM_TAGS); } + Constraint speakerMakespan(ConstraintFactory factory) { + return factory.forEach(Speaker.class) + .join(Talk.class, + fi...
The next line should cover this case too. Is the extra `if` worth it?
timefold-solver
github_2023
java
99
TimefoldAI
rsynek
@@ -718,19 +718,130 @@ private static <Result> Runnable innerSum(MutableReference<Result> resultContain */ public static <A, Mapped extends Comparable<? super Mapped>> UniConstraintCollector<A, ?, Mapped> min( Function<A, Mapped> groupValueMapping) { - return min(groupValueMapping, Compar...
```suggestion * {@code .groupBy(min(Person::name, Person::age))} returns {@code Ann} or {@code Eric}, ```
timefold-solver
github_2023
java
90
TimefoldAI
rsynek
@@ -35,11 +37,34 @@ public int countMutations(Solution_ a, Solution_ b) { Object bEntity = bIt.next(); for (GenuineVariableDescriptor<Solution_> variableDescriptor : entityDescriptor .getGenuineVariableDescriptorList()) { - // TODO broken if ...
Maybe slightly more readable: ``` int shorterListSize = Math.min(aSize, bSize); for (int i = 0; i < shorterListSize; i++) { ```
timefold-solver
github_2023
others
69
TimefoldAI
rsynek
@@ -0,0 +1,17 @@ +# https://github.com/Ezard/semantic-prs +enabled: true +titleAndCommits: true +allowMergeCommits: false +types: + - feat + - fix + - docs + - style + - refactor + - perf + - test + - build + - ci + - chore
What's the difference between style, refactor and chore? And what about build vs. ci? The fewer keywords we use, the bigger chance we use them right.
timefold-solver
github_2023
others
69
TimefoldAI
rsynek
@@ -0,0 +1,74 @@ +== Contributing to Timefold Solver + +This is an open source project, and you are more than welcome to contribute! + +* Found an issue? https://github.com/TimefoldAI/timefold-solver/issues[Submit an issue.] +* Want to fix an issue or contribute an improvement? https://github.com/TimefoldAI/timefold-so...
```suggestion . https://github.com/TimefoldAI/timefold-solver/fork[Fork it.] ```
timefold-solver
github_2023
others
69
TimefoldAI
rsynek
@@ -0,0 +1,74 @@ +== Contributing to Timefold Solver + +This is an open source project, and you are more than welcome to contribute! + +* Found an issue? https://github.com/TimefoldAI/timefold-solver/issues[Submit an issue.] +* Want to fix an issue or contribute an improvement? https://github.com/TimefoldAI/timefold-so...
```suggestion . https://github.com/TimefoldAI/timefold-solver/compare[Create a new Pull Request.] ```
timefold-solver
github_2023
java
68
TimefoldAI
rsynek
@@ -103,6 +123,29 @@ public void update(InTuple_ tuple) { } } + private List<FlattenedItem_> iterableToList(Iterable<FlattenedItem_> iterable) { + if (iterable instanceof Collection<FlattenedItem_> collection) { + // Optimization for Collection, where we know the size. + ...
In this case, it is safe to cast the `iterable` to a `Collection` and pass it here directly. It could bring an additional minor perf improvement, if we skip the `forEach` and the method reference (no proof).
timefold-solver
github_2023
java
28
TimefoldAI
ge0ffrey
@@ -131,8 +130,7 @@ public BenchmarkAggregatorFrame(BenchmarkAggregator benchmarkAggregator) { this.benchmarkAggregator = benchmarkAggregator; benchmarkResultIO = new BenchmarkResultIO(); plannerBenchmarkResultList = Collections.emptyList(); - Locale locale = benchmarkAggregator.getBen...
A) Either the benchmark report has a locale. B) Or the locale is the eye of the beholder. Basically, the browser's locale determines if it's 1'000,05 (BE) or 1,000.05 (US). If B) is possible, I prefer that. A big advantage over the JfreeChart approach is that the locale isn't hard coded in the report output. An...
timefold-solver
github_2023
others
28
TimefoldAI
ge0ffrey
@@ -56,6 +56,10 @@ <version.surefire.plugin>3.0.0-M8</version.surefire.plugin> <version.versions.plugin>2.15.0</version.versions.plugin> + <version.org.webjars.webjars-locator>0.46</version.org.webjars.webjars-locator> + <version.org.webjars.bootstrap>5.2.3</version.org.webjars.bootstrap> + <versio...
Should be at line 35 in this file, with the other dependency version, not with the plugin versions.
timefold-solver
github_2023
others
28
TimefoldAI
ge0ffrey
@@ -148,6 +152,24 @@ <artifactId>jcl-over-slf4j</artifactId> <version>${version.org.slf4j}</version> </dependency> + <dependency> + <groupId>org.webjars</groupId> + <artifactId>webjars-locator</artifactId> + <version>${version.org.webjars.webjars-locator}</version> + ...
comment should be above these artifacts?
timefold-solver
github_2023
java
1,493
TimefoldAI
zepfred
@@ -27,14 +27,17 @@ import ai.timefold.solver.core.config.localsearch.decider.acceptor.LocalSearchAcceptorConfig; import ai.timefold.solver.core.config.localsearch.decider.forager.LocalSearchForagerConfig; import ai.timefold.solver.core.config.phase.PhaseConfig; +import ai.timefold.solver.core.config.solver.PreviewF...
Should it be named `moveProviderClass`?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -155,12 +157,12 @@ protected void pickMove(ConstructionHeuristicStepScope<Solution_> stepScope) { protected <Score_ extends Score<Score_>> void doMove(ConstructionHeuristicMoveScope<Solution_> moveScope) { InnerScoreDirector<Solution_, Score_> scoreDirector = moveScope.getScoreDirector(); - sc...
This variable seems off to me. I think it should be within the scope of `scoreDirector.assertExpectedUndoMoveScore`.
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -24,17 +25,17 @@ public class DefaultConstructionHeuristicPhase<Solution_> implements ConstructionHeuristicPhase<Solution_> { protected final ConstructionHeuristicDecider<Solution_> decider; - protected final EntityPlacer<Solution_> entityPlacer; + protected final PlacerBasedMoveRepository<Solu...
Should we retrieve it from the builder?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -131,29 +126,32 @@ public void expandNode(ExhaustiveSearchStepScope<Solution_> stepScope) { @SuppressWarnings("unchecked") private <Score_ extends Score<Score_>> void doMove(ExhaustiveSearchStepScope<Solution_> stepScope, ExhaustiveSearchNode moveNode) { - InnerScoreDirector<Solution_, ...
Let's make the name consistent. We have used `executeTemporaryMove` in the CH decider.
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -131,29 +126,32 @@ public void expandNode(ExhaustiveSearchStepScope<Solution_> stepScope) { @SuppressWarnings("unchecked") private <Score_ extends Score<Score_>> void doMove(ExhaustiveSearchStepScope<Solution_> stepScope, ExhaustiveSearchNode moveNode) { - InnerScoreDirector<Solution_, ...
Same as before. The scope seems incorrect to me.
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -113,21 +111,20 @@ public void decideNextStep(LocalSearchStepScope<Solution_> stepScope) { @SuppressWarnings("unchecked") protected <Score_ extends Score<Score_>> void doMove(LocalSearchMoveScope<Solution_> moveScope) { - InnerScoreDirector<Solution_, Score_> scoreDirector = moveScope.getScoreDire...
Same as before
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -113,21 +111,20 @@ public void decideNextStep(LocalSearchStepScope<Solution_> stepScope) { @SuppressWarnings("unchecked") protected <Score_ extends Score<Score_>> void doMove(LocalSearchMoveScope<Solution_> moveScope) { - InnerScoreDirector<Solution_, Score_> scoreDirector = moveScope.getScoreDire...
Did you notice any improvements in performance after removing this lambda?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -57,12 +58,12 @@ phaseTermination, buildDecider(phaseConfigPolicy, phaseTermination)) private LocalSearchDecider<Solution_> buildDecider(HeuristicConfigPolicy<Solution_> configPolicy, PhaseTermination<Solution_> termination) { - var moveSelector = buildMoveSelector(configPolicy); + ...
Let's use `String::format`
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -2,64 +2,82 @@ import java.util.List; import java.util.Objects; +import java.util.function.BiFunction; +import ai.timefold.solver.core.api.score.Score; import ai.timefold.solver.core.impl.domain.solution.descriptor.DefaultPlanningListVariableMetaModel; import ai.timefold.solver.core.impl.domain.solution.desc...
```suggestion if (EphemeralMoveDirector.class.isAssignableFrom(getClass())) { ```
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -0,0 +1,116 @@ +package ai.timefold.solver.core.impl.move.generic; + +import java.util.Arrays; +import java.util.Collection; +import java.util.List; +import java.util.Objects; +import java.util.Set; +import java.util.stream.Collectors; + +import ai.timefold.solver.core.api.domain.solution.PlanningSolution; +import a...
Should we use `LinkedIdentityHashSet`?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -0,0 +1,116 @@ +package ai.timefold.solver.core.impl.move.generic; + +import java.util.Arrays; +import java.util.Collection; +import java.util.List; +import java.util.Objects; +import java.util.Set; +import java.util.stream.Collectors; + +import ai.timefold.solver.core.api.domain.solution.PlanningSolution; +import a...
Same as before
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -0,0 +1,12 @@ +package ai.timefold.solver.core.impl.move.streams; + +import java.util.Iterator; +import java.util.Random; + +import ai.timefold.solver.core.preview.api.move.Move; + +public interface MoveIterable<Solution_> extends Iterable<Move<Solution_>> {
I'm not sure what the advantage is of creating this abstraction
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -0,0 +1,34 @@ +package ai.timefold.solver.core.impl.move; + +import ai.timefold.solver.core.impl.phase.event.PhaseLifecycleListener; +import ai.timefold.solver.core.preview.api.move.Move; + +import org.jspecify.annotations.NullMarked; + +/** + * This is a shared abstraction for all three types of move iterators curr...
If we plan to use only this repository in the future, we need to consider that both CH and the Exact methods require a finite repository.
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -190,6 +182,7 @@ protected <Score_ extends Score<Score_>> void predictWorkingStepScore(AbstractSt @Override public void stepEnded(AbstractStepScope<Solution_> stepScope) { if (!isNested()) { + var solver = stepScope.getPhaseScope().getSolverScope().getSolver();
To ensure that these solver changes do not cause score corruption, we need to avoid sharing the score director in nested phases, while the solver itself can be shared. Given that the previous statement is correct, the changes will not cause any issues. Does that make sense?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -229,25 +226,42 @@ public void setWorkingSolution(Solution_ workingSolution) { * possibly many thousands of objects. * Providing the init score and genuine entity count requires another pass over the entities. * The following code does all of those operations in a single pass. + ...
I wonder if `moveRepository.initialize` will be called twice.
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -260,25 +274,30 @@ private void assertInitScoreZeroOrLess() { } @Override - public Score_ doAndProcessMove(Move<Solution_> move, boolean assertMoveScoreFromScratch, Consumer<Score_> moveProcessor) { - if (trackingWorkingSolution) { + public void executeMove(Move<Solution_> move) {
Is it safe to skip the assertions here?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -90,6 +92,19 @@ static <Score_ extends Score<Score_>> ConstraintAnalysis<Score_> getConstraintAn */ void setWorkingSolution(Solution_ workingSolution); + /**
This comment answers one of my questions
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -26,34 +23,10 @@ public interface InnerScoreDirectorFactory<Solution_, Score_ extends Score<Score */ ScoreDefinition<Score_> getScoreDefinition(); - @Override - default InnerScoreDirector<Solution_, Score_> buildScoreDirector(boolean lookUpEnabled, - ConstraintMatchPolicy constraintMatc...
Should this method be part of `ScoreDirectorFactory`?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -12,20 +10,4 @@ */ public interface ScoreDirectorFactory<Solution_> {
Does it make sense to maintain this interface?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -80,21 +77,21 @@ void oldAndNewTrailing() { var inverseVariableSupply = SelectorTestUtils.mockSingletonInverseVariableSupply( new TestdataChainedEntity[] { a1, a2, a3, a4, a5, b1, b2, b3 }); - try (var ephemeralMoveDirector = new MoveDirector<>(innerScoreDirector).ephemeral()) { - ...
I observed in several places that we are not invoking `executeTemporaryMove` when replacing the ephemeral director. Can this be a problem?
timefold-solver
github_2023
java
1,476
TimefoldAI
zepfred
@@ -133,48 +148,47 @@ void doMoveInSameChain() { SelectorTestUtils.assertChain(a0, a1, a2, a3, a4, a5, a6, a7); var moveDirector = new MoveDirector<>(innerScoreDirector); - try (var ephemeralMoveDirector = moveDirector.ephemeral()) { - var scoreDirector = ephemeralMoveDirector.getS...
This one you're calling the temporary action
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -57,6 +60,7 @@ public static <Solution_> VariableListenerSupport<Solution_> create(InnerScoreDi .toList(); this.listVariableDescriptor = scoreDirector.getSolutionDescriptor().getListVariableDescriptor(); this.listVariableEventList = new ArrayList<>(); + this.unassignedValueW...
Arguably this should be an identity hash set; no entities should be considered equal here. (The class doesn't exist in the JDK, but it can be created by converting `IdentityHashMap`.)
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -0,0 +1,126 @@ +package ai.timefold.solver.core.impl.util; + +import java.util.Collection; +import java.util.IdentityHashMap; +import java.util.Iterator; +import java.util.Objects; +import java.util.Set; + +public class IdentityHashSet<V> implements Set<V> {
Take a look at the Javadoc of `AbstractSet` - it explains how to write a minimal set implementation. All the other methods should only be implemented if you can make them faster in your implementation, and arguably in this case, we don't really need it. Less is more.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -0,0 +1,126 @@ +package ai.timefold.solver.core.impl.util; + +import java.util.Collection; +import java.util.IdentityHashMap; +import java.util.Iterator; +import java.util.Objects; +import java.util.Set; + +public class IdentityHashSet<V> implements Set<V> { + + private ElementAwareList<V> delegate; + private ...
Let's make them final. Clearing is cheap, final is fast.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -0,0 +1,126 @@ +package ai.timefold.solver.core.impl.util; + +import java.util.Collection; +import java.util.IdentityHashMap; +import java.util.Iterator; +import java.util.Objects; +import java.util.Set; + +public class IdentityHashSet<V> implements Set<V> { + + private ElementAwareList<V> delegate; + private ...
Should be called `LinkedIdentityHashSet`. Also, it should be final.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -57,6 +60,7 @@ public static <Solution_> VariableListenerSupport<Solution_> create(InnerScoreDi .toList(); this.listVariableDescriptor = scoreDirector.getSolutionDescriptor().getListVariableDescriptor(); this.listVariableEventList = new ArrayList<>(); + this.unassignedValueW...
We should add a comment here, explaining why we chose this particular type. (It's not obvious.)
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -240,61 +261,83 @@ public void beforeListVariableChanged(ListVariableDescriptor<Solution_> variable public void afterListVariableChanged(ListVariableDescriptor<Solution_> variableDescriptor, Object entity, int fromIndex, int toIndex) { var notifiables = notifiableRegistry.get(variableDescr...
Both of these separate iterations have been merged together to just one iteration. If you can iterate once, iterate once; useful for CPU caches.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -240,61 +261,83 @@ public void beforeListVariableChanged(ListVariableDescriptor<Solution_> variable public void afterListVariableChanged(ListVariableDescriptor<Solution_> variableDescriptor, Object entity, int fromIndex, int toIndex) { var notifiables = notifiableRegistry.get(variableDescr...
I only split this into a separate method so that I can see its impact in the profiler individually. This is not actually a perf optimization, just helps with profiling.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -0,0 +1,77 @@ +package ai.timefold.solver.core.impl.util; + +import java.util.AbstractSet; +import java.util.IdentityHashMap; +import java.util.Iterator; + +import org.jspecify.annotations.NullMarked; + +/** + * This set does not support null keys. + * + * @param <V> + */ +@NullMarked +public final class LinkedIden...
These zero checks are speculative; but if you take a look at what the map does even when it's empty, I bet this will prove itself if the methods are called often enough.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -0,0 +1,77 @@ +package ai.timefold.solver.core.impl.util; + +import java.util.AbstractSet; +import java.util.IdentityHashMap; +import java.util.Iterator; + +import org.jspecify.annotations.NullMarked; + +/** + * This set does not support null keys. + * + * @param <V> + */ +@NullMarked +public final class LinkedIden...
contains+remove is worse than just remove, if you don't need to support null keys.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -0,0 +1,77 @@ +package ai.timefold.solver.core.impl.util; + +import java.util.AbstractSet; +import java.util.IdentityHashMap; +import java.util.Iterator; + +import org.jspecify.annotations.NullMarked; + +/** + * This set does not support null keys. + * + * @param <V> + */ +@NullMarked +public final class LinkedIden...
computeIfAbsent does the same thing, but it also forces the JVM to create a lambda instance. This does not.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -155,9 +155,18 @@ public void forEach(Consumer<? super T> tupleConsumer) { */ @Override public Iterator<T> iterator() { + if (size == 0) { + return Collections.emptyIterator();
Reduces allocations in some pathological cases.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -25,38 +28,51 @@ import ai.timefold.solver.core.impl.domain.variable.supply.Supply; import ai.timefold.solver.core.impl.domain.variable.supply.SupplyManager; import ai.timefold.solver.core.impl.score.director.InnerScoreDirector; +import ai.timefold.solver.core.impl.util.LinkedIdentityHashSet; + +import org.jspeci...
Don't create what doesn't need creating.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -215,14 +233,17 @@ public void beforeVariableChanged(VariableDescriptor<Solution_> variableDescript } public void afterElementUnassigned(ListVariableDescriptor<Solution_> variableDescriptor, Object element) { - Collection<ListVariableListenerNotifiable<Solution_>> notifiables = notifiableRegistry....
Don't use what doesn't need using.
timefold-solver
github_2023
java
1,480
TimefoldAI
triceo
@@ -240,61 +261,83 @@ public void beforeListVariableChanged(ListVariableDescriptor<Solution_> variable public void afterListVariableChanged(ListVariableDescriptor<Solution_> variableDescriptor, Object entity, int fromIndex, int toIndex) { var notifiables = notifiableRegistry.get(variableDescr...
No need to create the event if we already have the notification. Help the GC do less work, help yourself.
timefold-solver
github_2023
java
1,479
TimefoldAI
triceo
@@ -15,7 +17,14 @@ FeatureBuildItem feature() { } @BuildStep - void registerTimefoldJacksonModule(BuildProducer<AdditionalBeanBuildItem> additionalBeans) { + ClassPathJacksonModuleBuildItem registerTimefoldJacksonModule() { + // Make timefold-solver-jackson discoverable by quarkus-rest + ...
I'm good with removing this.
timefold-quickstarts
github_2023
python
703
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,87 @@ +from timefold.solver.score import * +from datetime import time +from typing import Final
Unused import ```suggestion ```
timefold-quickstarts
github_2023
python
703
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,255 @@ +import json +from itertools import count +from random import Random +from datetime import datetime, time, timedelta +from typing import List, Callable, TypeVar + +from .domain import * + +random = Random(0) +T = TypeVar('T') +L = TypeVar('L') + +# Constants for employee skills +ATTENDANT_SKILL = "Fli...
I personally find this hard to read/reason about. I recommend using a for loop instead: ```suggestion for start_date in dates: apply_random_value(count_dates, flights, lambda f: f.departure_utc_date_time is None, lambda f: flight_consumer(f, start_date)) ```
timefold-quickstarts
github_2023
python
703
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,255 @@ +import json +from itertools import count +from random import Random +from datetime import datetime, time, timedelta +from typing import List, Callable, TypeVar + +from .domain import * + +random = Random(0) +T = TypeVar('T') +L = TypeVar('L') + +# Constants for employee skills +ATTENDANT_SKILL = "Fli...
The walrus `x := y` operator can be confusing and not usually encountered. I recommend importing [itertools.count](https://docs.python.org/3/library/itertools.html#itertools.count) and using it as an id generator ```suggestion ids = map(str, itertools.count()) for airport in flight_airports: for _ ...
timefold-quickstarts
github_2023
python
703
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,255 @@ +import json +from itertools import count +from random import Random +from datetime import datetime, time, timedelta +from typing import List, Callable, TypeVar + +from .domain import * + +random = Random(0) +T = TypeVar('T') +L = TypeVar('L') + +# Constants for employee skills +ATTENDANT_SKILL = "Fli...
Can be removed if my suggestion above is applied. ```suggestion ```
timefold-quickstarts
github_2023
python
703
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,177 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider...
Not needed; if `__gt__` is not defined, `__lt__` will be used. ```suggestion ```
timefold-quickstarts
github_2023
python
703
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,177 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider...
Tuple comparison logic can simplify this: ```suggestion def _comparison_key(self): return (departure_utc_date_time, departure_airport, arrival_utc_date_time, arrival_airport, flight_number) def __lt__(self, other: "Flight") -> bool: return self._comparison_key() < other._comparison_...
timefold-quickstarts
github_2023
python
694
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,61 @@ +from timefold.solver.score import * +from datetime import time +from typing import Final + +from .domain import * +from .score_analysis import LoadBalanceJustification + + +@constraint_provider +def define_constraints(constraint_factory: ConstraintFactory): + return [ + one_assignment_per_da...
Extra line ```suggestion ```
timefold-quickstarts
github_2023
python
694
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,97 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider,...
I would make this default to `False` (in Java, all `boolean` values are initialized to `false` even if they don't appear in a constructor and don't have a field initializer. ```suggestion pinned: Annotated[bool, PlanningPin, Field(default=False)] ```
timefold-quickstarts
github_2023
python
694
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,68 @@ +from timefold.solver.test import ConstraintVerifier +from tournament_scheduling.domain import * +from tournament_scheduling.constraints import * + +constraint_verifier = ConstraintVerifier.build(define_constraints, TournamentSchedule, TeamAssignment) +# Sample teams and days (equivalent to static vari...
If you make `pinned` default to `False`, you won't need to specify it on all these objects (applies to the whole file).
timefold-quickstarts
github_2023
python
694
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,68 @@ +from timefold.solver.test import ConstraintVerifier +from tournament_scheduling.domain import * +from tournament_scheduling.constraints import * + +constraint_verifier = ConstraintVerifier.build(define_constraints, TournamentSchedule, TeamAssignment) +# Sample teams and days (equivalent to static vari...
I don't think the latter part of the comment is helpful. It would only confuse a Python developer looking at it. I would remove this comment, since it is obvious this is sample data (and there no such comment in Java). ```suggestion ```
timefold-quickstarts
github_2023
others
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,159 @@ +<!DOCTYPE> +<html lang="en"> +<head> + <meta content="text/html; charset=UTF-8" http-equiv="Content-Type"> + <meta content="width=device-width, initial-scale=1" name="viewport"> + <title>Sports League Scheduling - Timefold Solver on Python</title>
```suggestion <title>Sports League Scheduling - Timefold Solver for Python</title> ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,155 @@ +from timefold.solver.score import * +from datetime import time +from typing import Final + +from .domain import * + + +MAX_CONSECUTIVE_MATCHES: Final[int] = 4 + + +@constraint_provider +def define_constraints(constraint_factory: ConstraintFactory): + return [ + matches_on_same_day(constrain...
The `if isinstance(match.home_team, Team) and isinstance(match.away_team, Team) else 0` should be unnecessary; all matches home and away teams should be set. It might be best to add a method to Team `get_distance(self, team: Team)` to perform the lookup logic.
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,155 @@ +from timefold.solver.score import * +from datetime import time +from typing import Final + +from .domain import * + + +MAX_CONSECUTIVE_MATCHES: Final[int] = 4 + + +@constraint_provider +def define_constraints(constraint_factory: ConstraintFactory): + return [ + matches_on_same_day(constrain...
I would inline `is_invalid_classic_match`
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,126 @@ +import json +import random
To prevent accidentally using a shared random, only import the `Random` class from the package. ```suggestion from random import Random ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,126 @@ +import json +import random +from datetime import datetime, time, timedelta +from typing import List, Callable + +from .domain import * + +
You should create a seeded random so the generated demo data is consistent. ```suggestion random = Random(0) ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,126 @@ +import json +import random +from datetime import datetime, time, timedelta +from typing import List, Callable + +from .domain import * + + +DISTANCE_IN_KM = [ + [0, 2163, 2163, 2160, 2156, 2156, 2163, 340, 1342, 512, 3038, 1526, 2054, 2054], + [2163, 0, 11, 50, 813, 813, 11, 1967, 842, 1661, 11...
I would rename this to just `generate_demo_data` ```suggestion def generate_demo_data() -> LeagueSchedule: ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,113 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider...
The home and away teams should always be set. Also remove the `Field(default=None)`. ```suggestion home_team: Annotated[Team, ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,113 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider...
```suggestion away_team: Annotated[Team, ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,113 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider...
I would add a `get_distance(self, team: Team)` method here.
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,113 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider...
Slightly unaligned. ```suggestion PlanningEntityCollectionProperty] ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,113 @@ +from timefold.solver import SolverStatus +from timefold.solver.domain import (planning_entity, planning_solution, PlanningId, PlanningVariable, + PlanningEntityCollectionProperty, + ProblemFactCollectionProperty, ValueRangeProvider...
I would annotate this attribute with `Field(default=False)` so you wouldn't need to specify it every time in the test. ```suggestion weekend_or_holiday: Annotated[bool, Field(default=False)] ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,165 @@ +from timefold.solver.test import ConstraintVerifier + +from sports_league_scheduling.domain import * +from sports_league_scheduling.constraints import * + +constraint_verifier = ConstraintVerifier.build(define_constraints, LeagueSchedule, Match) + + +def test_matches_same_day(): + # Two matches fo...
`round=None` and `classic_match=False` are the defaults for `Match` and do not need to be specified (applies to the whole file wherever `classic_match=False` or `round=None`). ```suggestion match3 = Match(id="3", home_team=home_team, away_team=rival_team) ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -0,0 +1,165 @@ +from timefold.solver.test import ConstraintVerifier + +from sports_league_scheduling.domain import * +from sports_league_scheduling.constraints import * + +constraint_verifier = ConstraintVerifier.build(define_constraints, LeagueSchedule, Match) + + +def test_matches_same_day(): + # Two matches fo...
Don't need to specify `distance_to_team` if you are not populating it (it will create an empty dict using the default factory) (applies to the whole file wherever `distance_to_team={}` is used). ```suggestion home_team = Team(id="1", name="TeamA") ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -81,7 +81,7 @@ def start_to_away_hop(constraint_factory: ConstraintFactory) -> Constraint: .for_each(Match) .if_not_exists(Round, Joiners.equal(lambda match: match.round.index - 1, lambda match_round: match_round.index)) - .penaliz...
Parenthesis uneccessary ```suggestion .penalize(HardSoftScore.ONE_SOFT, lambda match: match.away_team.get_distance(match.home_team)) ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -91,7 +91,7 @@ def home_to_away_hop(constraint_factory: ConstraintFactory) -> Constraint: .join(Match, Joiners.equal(lambda match: match.home_team, lambda match: match.away_team), Joiners.equal(lambda match: match.round.index + 1, lambda match: match.round.index)) -...
```suggestion .penalize(HardSoftScore.ONE_SOFT, lambda match, other_match: match.home_team.get_distance(other_match.home_team)) ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -101,7 +101,7 @@ def away_to_away_hop(constraint_factory: ConstraintFactory) -> Constraint: .join(Match, Joiners.equal(lambda match: match.away_team, lambda match: match.away_team), Joiners.equal(lambda match: match.round.index + 1, lambda match: match.round.index))...
```suggestion .penalize(HardSoftScore.ONE_SOFT, lambda match, other_match: match.home_team.get_distance(other_match.home_team)) ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -111,45 +111,22 @@ def away_to_home_hop(constraint_factory: ConstraintFactory) -> Constraint: .join(Match, Joiners.equal(lambda match: match.away_team, lambda match: match.home_team), Joiners.equal(lambda match: match.round.index + 1, lambda match: match.round.index...
```suggestion .penalize(HardSoftScore.ONE_SOFT, lambda match, other_match: match.home_team.get_distance(match.away_team)) ```
timefold-quickstarts
github_2023
python
685
TimefoldAI
Christopher-Chianelli
@@ -111,45 +111,22 @@ def away_to_home_hop(constraint_factory: ConstraintFactory) -> Constraint: .join(Match, Joiners.equal(lambda match: match.away_team, lambda match: match.home_team), Joiners.equal(lambda match: match.round.index + 1, lambda match: match.round.index...
```suggestion .penalize(HardSoftScore.ONE_SOFT, lambda match: match.home_team.get_distance(match.away_team)) ```
timefold-quickstarts
github_2023
others
601
TimefoldAI
triceo
@@ -121,4 +120,28 @@ jobs: export NEW_PYTHON_VERSION="999-dev0" .github/scripts/change_versions.sh git commit -am "build: move back to version $NEW_VERSION" - git push origin $RELEASE_BRANCH_NAME \ No newline at end of file + git push origin $RELEASE_BRANCH_NAME + + ...
Why are we doing this if none of workflows in the other repos had this? Can't we delete the branch if it exists, like we do everywhere else? For consistency.
timefold-quickstarts
github_2023
others
601
TimefoldAI
triceo
@@ -1,4 +1,4 @@ -name: Release +'name: Release
?
timefold-quickstarts
github_2023
others
553
TimefoldAI
triceo
@@ -1,11 +1,9 @@ At this point, the release of _Timefold Quickstarts_ is ready to be published. -- [x] Release branch has been created. -- [x] Git tag has been published. -- [ ] Delete the branch `dry_run`. -- [ ] Create the version branch (e.g., `1.13.x`) from the `stable` one. +- Release branch has been created. +...
Arguably, this can be removed entirely. It doesn't provide any extra information.
timefold-quickstarts
github_2023
others
536
TimefoldAI
triceo
@@ -28,6 +28,13 @@ jobs: java-version: [ 17 ] # Only the first supported LTS; already too many jobs here. timeout-minutes: 120 steps: + # Clone timefold-quickstarts + # No need to check for stale repo, as Github merges the main repo into the fork automatically.
Please remove this line (and others like it), we no longer check for stale repo anywhere.
timefold-quickstarts
github_2023
java
536
TimefoldAI
triceo
@@ -45,7 +45,7 @@ public class Job { /** * Start is after cleanup. */ - @CascadingUpdateShadowVariable(targetMethodName = "updateStartCleaningDateTime", sourceVariableNames = {"line", "previousJob"}) + @CascadingUpdateShadowVariable(targetMethodName = "updateStartCleaningDateTime") private L...
Didn't we discuss that _nothing_ is allowed to depend on cascading shadows? How do piggybacks work, if the shadows are not part of the var listener chain anymore, but instead are processed independently in their own loop?
timefold-quickstarts
github_2023
java
527
TimefoldAI
triceo
@@ -40,15 +41,19 @@ public class Job { @NextElementShadowVariable(sourceVariableName = "jobs") private Job nextJob; - /** Start is after cleanup. */ - @ShadowVariable(variableListenerClass = StartDateTimeUpdatingVariableListener.class, sourceVariableName = "line") - @ShadowVariable(variableListener...
Observation: that's a lot of boilerplate. The annotation - 6 times. The name of the method - 6 times. The name of each variable - three times.
timefold-quickstarts
github_2023
java
527
TimefoldAI
triceo
@@ -1,102 +0,0 @@ -package org.acme.foodpackaging.domain.solver; - -import java.time.LocalDateTime; -import java.util.Objects; - -import org.acme.foodpackaging.domain.Job; -import org.acme.foodpackaging.domain.Line; -import org.acme.foodpackaging.domain.PackagingSchedule; -import ai.timefold.solver.core.api.domain.vari...
Beautiful!
timefold-quickstarts
github_2023
python
528
TimefoldAI
triceo
@@ -41,3 +42,14 @@ def __post_init__(self): self.description = (f"Vehicle '{self.vehicle_name}' total travel time is " f"{self.total_driving_time_seconds // (60 * 60)} hours " f"{round(self.total_driving_time_seconds / 60)} minutes.") + + +@dataclass +c...
We've agreed a while back that we won't be adding justifications to the quickstarts. Please remove; if the Java quickstart still has them, that's an accident and please remove those too.
timefold-quickstarts
github_2023
javascript
528
TimefoldAI
triceo
@@ -20,6 +21,37 @@ const visitGroup = L.layerGroup().addTo(map); const homeLocationGroup = L.layerGroup().addTo(map); const routeGroup = L.layerGroup().addTo(map); +/************************************ Time line constants and variable definitions ************************************/ + +const byVehiclePanel = docu...
Please make sure this file is identical to the Java frontend. (I haven't checked whether it is or not.)
timefold-quickstarts
github_2023
javascript
496
TimefoldAI
zepfred
@@ -335,8 +335,8 @@ function analyze() { function getScoreComponents(score) { let components = {hard: 0, medium: 0, soft: 0}; - $.each([...score.matchAll(/(-?[0-9]+)(hard|medium|soft)/g)], (i, parts) => { - components[parts[2]] = parseInt(parts[1], 10); + $.each([...score.matchAll(/(-?\d*(\.\d+)?)(...
Agreed. Let's create a papercut story to ensure consistency between the quickstart modules.