API, IR, etc design improvements

[0xalpharush]

Jotting down a list of changes that could improve future versions of Slither and its framework

• Get rid of HighLevelCall <-> state variable/function footgun (missing instance check for Function and accessing ir.function causing AttributeError ) fix(cache-array-length): handle when HighLevelCall is a StateVariable #2019
• Get rid of InternalCall <-> SolidityCall sharing

  slither/slither/core/cfg/node.py

  Line 899 in 2c792b2

  # TODO: consider removing dependancy of solidity_call to internal_call

• Get rid of late lookup for SolidityImportPlaceholder
• Get rid of late custom error signature lookup
• Get rid of ternary conversion or rework it altogether
• Rework file/detector filtering from scratch (Filter paths before parsing and lowering AST when possible #2421, Filtering #2428)
• Rename SlitherCompilationUnitSolc to SolidityCompilationUnit
• Investigate using iterator/generator over for loop or other API's for detectors like refactor AbstractDetector to use the visitor pattern  #1751
• Need to have a way for detector to show call path e.g. call in loop detector ([Bug]: calls-loop does not report the context of the loop #1468)
• Need to show origin of taint and flow ([False-Positive]: Block timestamp and Dangerous strict equalities #2425 (comment))
• Retain data flow as an actual graph e.g. we don't have a straightforward way to get the operations that created a variable like but maybe we can come up with an intuitive way to get this info and add it to the public API.