SemanticTokensService does not corretly implement LSP specification

[jnt0r]

The SemanticTokensService implementation does not correctly implement the LSP specification in two points.

1. It only supports the three TokenTypes: KEYWORD_ID, DEFAULT_ID, NUMBER_ID. In the specification there are way more TokenTypes.
2. When converting the TokenType from String to index there is always an offset of 1 added which is not correct.

private int getTokenType(final String[] ids) {
    for (String id : ids) {
	Integer index = tokenTypes.indexOf(id);
	if (index != -1) {
		return index + 1;
	}
    }
    return 0;
}
...
Integer positionTokenType = getTokenType(lightweightPosition.getIds());
Integer positionTokenModifiers = getTokenModifiers(lightweightPosition.getIds());

if (positionTokenType != 0 || positionTokenModifiers != 0) {
    int deltaOffset = lightweightPosition.getOffset() - lastOffset;
    ...

The specification says:
Token types are looked up by index, so a tokenType value of 1 means tokenTypes[1].
https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#textDocument_semanticTokens

I will open a PR to fix these issues, but I wanted to address them as a problem first.