Improve dictionary#indexOf routine with sparse index

[hqx871]

Use sparse index to improve GenericIndexed#indexOf performance, especially for large dictionary.

Description

The sorted GenericalIndexed#indexOf use binary search to get index of one object value. This routine will call ObjectStrategy#fromByteBuffer to decode value from the byteBuffer, which is expensive. I suggest build a sparse index for large dictionary when load it. this can get 40+% performance improvement. I made a benchmark here.

Benchmark	(cardinality)	(indexGranularity)	(searchNum)	(sparseType)	(strLen)	Mode	Cnt	Score	Error	Units
DictionaryWrapperBenchmark.indexOf	1000000	0	10000	array	16	avgt	10	23.760	±2.310	ms/op
DictionaryWrapperBenchmark.indexOf	1000000	512	10000	array	16	avgt	10	17.015	±4.904	ms/op
DictionaryWrapperBenchmark.indexOf	1000000	1024	10000	array	16	avgt	10	14.533	±3.370	ms/op
DictionaryWrapperBenchmark.indexOf	1000000	2048	10000	array	16	avgt	10	13.390	±0.748	ms/op
DictionaryWrapperBenchmark.indexOf	1000000	4096	10000	array	16	avgt	10	14.672	±1.223	ms/op
DictionaryWrapperBenchmark.indexOf	1000000	8192	10000	array	16	avgt	10	15.090	±0.721	ms/op

---

Key changed/added classes in this PR

• SparseArrayIndexed
• GenericIndexed
• DictionaryWrapStrategy

---

This PR has:

• been self-reviewed.
  • using the concurrency checklist (Remove this item if the PR doesn't have any relation to concurrency.)
• added documentation for new or modified features or behaviors.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.