Separate verdant view and lilgit into separate packages about verdant HOT 4 CLOSED

Ok, I've decoupled the pieces in Typescript, but for performance, a bunch of analysis operations from lilgit Typescript side really need to be offloaded to python.

from verdant.

Current APIs operations:

• Utils
  • findNodeAtRange
  • inRange
  • reduceASTDict (could be done at parsing stage?)
  • dictToCodeCellNodey (why is this needed?)
• Match [Py]
  • recieve_newVersion
  • finalizeMatch
  • grabUnmatchedParents
  • matchParentNodes
  • scoreMatch
  • matchLeaves
  • findMatchOptions
  • buildStarNode
  • matchLiterals
  • forceReplace
• Resolve
  • repairMarkdown
  • repairCellAST
  • shiftAllAfter
  • shiftAllChildren
  • dictToNodeyList
  • nodeyToLeaves
• coldStart [Ts] is for when we start a notebook with no prior history
• hot start [Ts] is for when we start a notebook with prior history and want to match on init
• freshNodey & createCellNodey (can probably be combined?)
  • generateCodeNodey is different how?
• repairCell (why are we repairing only at the cell level???)
  • repairCodeCell
• parseRequest [Ts]
• failSafeParse [Py]
• cleanCodeString [Ts]

from verdant.

Seeing some formatting on the Ts end that should be handled in the original parsing:

if ("type" in dict === false) {
      dict.type = "Module";
    }

Also, need to return safe defaults from parsing if an error occurs.

from verdant.

Next step: isolate everything out that should be moved to Py, and make sure app works with no Py language support ... ie, how the app will work with Julia or R, make sure that safe default behavior works well, and that will ensure the language support API is cleanly separated

from verdant.