[core] Add bloom filter for file index

[leaves12138]

Related to #3115

• API
  MAYBE something like:

CREATE TABLE <PAIMON_TABLE> (<COLUMN> <COLUMN_TYPE> , ...) WITH
(
"file.index.columns" = "c1:bloom:{items=200, fpp=0.1}; c2:bloom:{items=1000000, fpp=0.03}"
)

• CODE

1. Use bloom filter in paimon.
2. Add options for file index. Refer to https://docs.databricks.com/en/optimizations/bloom-filters.html

• NOTE

This pull request only commit the bloom filter class. The next will commit read and write code.

xx hash has better performace:
https://www.synnada.ai/glossary/xxhash
https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html

[JingsongLi]

items and fpp options should for a field, not table options.

[leaves12138]

orc: 64 bit hash
parquet: 64 bit hash
hadoop-common: 32 bit hash, hash function is old
kudu: 64 bit hash

[leaves12138]

orc: use number hash and bytes hash independently
parquet: put all in bytebuffer, use xx hash

[JingsongLi]

+1