Benchmark feature for multiple solvers

[simontiffert]

To support different types of planning (additional algorithm configurations like acceptedCountLimit) I wanted to use the multiple solver feature in Quarkus:

quarkus.timefold.solver."fast".solver-config-xml=solverConfigFast.xml
quarkus.timefold.solver."deep".solver-config-xml=solverConfigDeep.xml

But using it I got the error:
When defining multiple solvers, the benchmark feature is not enabled.

Is there a build in solution for Quarkus to still use the benchmark?
It would be great to configure the benchmark per solver in a similar way:

quarkus.timefold.benchmark."fast".solver-benchmark-config-xml=solverBenchmarkConfig.xml

[triceo]

Hello Simon! As far as I know, there is no way to do this out of the box.
You can inject the individual solvers, and build the benchmark configs for them manually, but that's a poor substitute.

[simontiffert]

I have changed the import from timefold-solver-quarkus-benchmark to timefold-solver-benchmark.
The benchmark is now called directly without the Quarkus @Inject for the benchmarkFactory (and I needed to specify the benchmarkDirectory in the XML).

Now it runs as:

    @Test
    public void benchmark() throws IOException, URISyntaxException {
        PlannerBenchmarkFactory benchmarkFactory = PlannerBenchmarkFactory.createFromXmlResource("solverBenchmarkConfig.xml");
        Schedule schedule = objectMapper.readValue(new URI("file:src/test/resources/data.json").toURL(), Schedule.class);
        benchmarkFactory.buildPlannerBenchmark(schedule).benchmarkAndShowReportInBrowser();
    }

But how do I access the SolutionManager in a multi-solver environment. Is there a working @Inject or should I initialize it manually. The @Named was not working:

SolutionManager<Schedule, HardMediumSoftScore> solutionManager = SolutionManager.create(solverManager);

[triceo]

What you're trying to do is not possible - @Named has the unfortunate property that it can only ever reference one thing. So if you have two different solver configs, you can't really use it. The reference would be ambiguous - not much we can do about it, AFAIK that's CDI spec.

The single solver environment is the default, and we make it easy. Advanced use cases require advanced configuration.