My understanding is that Salesforce haven't currently published such a document. Instead they want you to use the Tooling API to get the tokens etc...
See Grammar for creating an Apex parser
Depending on your scenario you might find the Apex Language Server a useful tool. It does the Apex parsing etc... behind the scenes for the current VSCode tooling.
Incidentally, the Salesforce StackExchange site linked above is a great place to ask Salesforce specific questions.