GH-3987 introduce InternedIRI for use in Vocabularies to improve performance

[hmottestad]

GitHub issue resolved: #3987

Briefly describe the changes proposed in this PR:

• New (internalOnly) class InternedIRI which interns the stringValue and pre-calculates the hashCode
• Use InternedIRI in Vocabularies and in CoreDatatype

---

PR Author Checklist (see the contributor guidelines for more details):

• my pull request is self-contained
• I've added tests for the changes I made
• I've applied code formatting (you can use mvn process-resources to format from the command line)
• I've squashed my commits where necessary
• every commit message starts with the issue number (GH-xxxx) followed by a meaningful description of the change

[hmottestad]

I've changed both #3990 and this PR to draft for the time being. They both definitely improve performance, but that was only really obvious when combined with #3934 .

[JervenBolleman]

Hi just a quick thought. Will this really help at scale? I have had bad experiences with intern on string similar to https://shipilev.net/jvm/anatomy-quarks/10-string-intern/

Never mind I see you notes this in the comments 😹

[hmottestad]

Before

Benchmark                                                                                        (param)  Mode  Cnt   Score   Error  Units
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencer                            moreRdfs::12180  avgt    5  12.385 ± 0.043  ms/op
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencerMultipleTransactions        moreRdfs::12180  avgt    5   9.372 ± 0.072  ms/op
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencerMultipleTransactionsSchema  moreRdfs::12180  avgt    5   9.447 ± 0.019  ms/op
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencerSchema                      moreRdfs::12180  avgt    5   9.726 ± 0.029  ms/op
ReasoningBenchmark.noReasoning                                                           moreRdfs::12180  avgt    5   3.877 ± 0.016  ms/op
ReasoningBenchmark.noReasoningMultipleTransactions                                       moreRdfs::12180  avgt    5   3.858 ± 0.036  ms/op

After

Benchmark                                                                                        (param)  Mode  Cnt   Score   Error  Units
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencer                            moreRdfs::12180  avgt    5  11.068 ± 0.074  ms/op
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencerMultipleTransactions        moreRdfs::12180  avgt    5   8.388 ± 0.013  ms/op
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencerMultipleTransactionsSchema  moreRdfs::12180  avgt    5   8.540 ± 0.140  ms/op
ReasoningBenchmark.forwardChainingSchemaCachingRDFSInferencerSchema                      moreRdfs::12180  avgt    5   8.866 ± 0.029  ms/op
ReasoningBenchmark.noReasoning                                                           moreRdfs::12180  avgt    5   3.901 ± 0.012  ms/op
ReasoningBenchmark.noReasoningMultipleTransactions                                       moreRdfs::12180  avgt    5   3.718 ± 0.021  ms/op

[hmottestad]

This is showing a decent 10% performance improvement in the more complex RDFS reasoning benchmark.