Provide SemanticTokensLegend#encode method to perform encoding according to standard

[laeubi]

I'm currently try to use getting some basic syntax highlights with SemanticTokensLegend / org.eclipse.lsp4j.services.TextDocumentService.semanticTokensFull(SemanticTokensParams) due to the specification this requires a non trivial transformation of types/modifiers/locations into a stream of ints.

To ease implementation (and add validation) I think it would be usefull to have a method in SemanticTokensLegend that helps in encoding things for example returning an IntStream and taking the keyword, the modifier and a location as input.

This method can the for example fail if invalid token types are send and so on.

[pisv]

Sounds like a good idea to me. However, from the API point of view, I would rather suggest adding the following methods to the SemanticTokensLegend:

	int encodeTokenType(String tokenType)
	int encodeTokenModifiers(Set<String> tokenModifiers)
	String decodeTokenType(int tokenType)
	Set<String> decodeTokenModifiers(int tokenModifiers)

@laeubi Would you like to submit a PR for that?