Improve general auto-complete / intellisense (Not perfect but better)

[Ho11ow1]

Still not perfect but it really can't be however it is better than it was.

Improvements

• All immediate namespaces are shown in full persistent, preferences, gui, config, bubble, etc...
• Persistent now shows all variables and the 2 specific persistent methods
• Auto-complete now shows a full list of accessible variables and methods under a customStore
• Variables created via default x = customStore.Class() now also have intellisense for their methods and fields
• All immediately available variables, methods, classes and stores now show up in intellisense

Limitations

• It unfortunately doesn't do everything i would like it too however it is better than what it was e.g. custom.var. does not show it's appropriate methods
• From my testing this is due to const range = document.getWordRangeAtPosition(prefixPosition); forcing us to only receive the last non empty component of the entire line (rentale.all_items. => kwPrefix = all_items) because of the default regex that vscode uses for getWordRangeAtPosition. However if we use the regex: /[\w.]+$/ to capture the entire expression, we then break call, call screen and jump auto-complete

Thoughts

• Maybe an entire re-write would be better. On vscode-init firstly we go through navigation.json, then parse each .rpy and _ren.py file so we have the context of all custom stores, custom store variables, methods etc... From there we could implement full call and jump auto-complete so it would be aware of all the available labels and screens, However that is outside the scope of this PR and is really just an additional thought
• With that said i'll look at the parser branch