feat (interactive): Migrating Gremlin compilation into the `Calsite`-based IR layer

[longbinlai]

The GIE IR layer is re-factored based on Calsite, which has now been used to compile the Cypher language into multiple runtime. However, the compilation of Gremlin is still in the pre-IR route, making the dual-language support of GIE in an inconsistent state, and the Gremlin path can not benefit from the on-going developing features of IR, including new optimization rules.

It's now ready to initiate the migration.

[shirly121]

The following operators have currently been supported for conversion from syntax to Graph Relational: graph：

• V()
• E()
• outE
• inE()
• bothE()
• out() (include path_expand)
• in()
• both()
• inV()
• outV()
• otherV()
• endV()

filter：

• hasLabel
• hasId
• hasNot
• has supported gremlin predicates:
  • eq
  • neq
  • gt
  • gte
  • lt
  • lte
  • and
  • or
  • within
  • without
  • containing
  • notContaining
  • startingWith
  • notStartingWith
  • endingWith
  • notEndingWith
  • not
  • inside
  • outside

Project：

• values
• valueMap
• elementMap
• select (exclude apply)

[shirly121]

Todo List: Support for other Syntax to Graph Relational conversion:

• [x] order
• [x] limit
• [x] dedup
• [x] groupBy
• [x] union
• [x] match
• [x] expr
• [x] subtask
• [ ] subgraph

在原先gremlin链路新支持的功能：

• [ ] until and more path_opt in path_expand
• [ ] sample
• [ ] unfold

Support for Graph Relational to IR Core conversion:

• [ ] support nested map : select('a', 'b').by(valueMap(..))

• [x] index predicate Allowing within operator in index predicate to support queries like g.V().hasId(1,2,3)

• [x] MAP_VALUE_CONSTRUCTOR
  Further expand the implementation of varMap operator

• [x] POSIX_REGEX_CASE_SENSITIVE To support startingWith/endingWith/containing/regex

• [x] SEARCH SEARCH and Sarg in Calcite are used to represent a value within a continuous or discrete interval, which need to be further unified with the current ir core structure. For example:

    has('age', within(17, 20)) -> SEARCH(DEFAULT.age, Sarg[17, 20])
    has('age', without(17, 20)) -> SEARCH(DEFAULT.age, Sarg[(-∞..17), (17..20), (20..+∞)])
  
    has('age', inside(17, 20)) -> SEARCH(DEFAULT.age, Sarg[[18..19]])
    has('age', outside(17, 20)) -> SEARCH(DEFAULT.age, Sarg[(-∞..18), (19..+∞)])