从ANTLR语法创建树

Question

I have written an ANTLR Grammar file and now i need to generate an adjacency matrix telling me which rule in grammar is associated with which one. for ex :- start : ('show' | 'give' | 'get') 'me' ('all')? phrase | 'I' 'would' 'like' ('all')? phrase | phrase;

phrase : constructPhrase (('and')? constructPhrase)* | constructPhrase 'and' ('its' | 'their') constructPhrase | constructPhrase functionPhrase | functionPhrase ;

Here we have I would like associated with each other..so i need to read the grammar file and generate an adjacency matrix.

Answer 1

I have created a project on github https://github.com/julianthome/inmemantlr that takes the g4 grammar definition and a text file that contains the actual content to parse. It will automatically generate an AST for you where the nodes do contain the parsing rules and the token associated with that rule. It might be useful for your purpose.

It is written in Java -- you could plug in any g4 file you like ... the parser generation and AST generation is done automatically. Code examples can be found in the test cases.

Best wishes and kind regards Julian

Answer 2

Parsing an ANTLR grammar is not difficult. I have done this as part of my code completion implementation in MySQL Workbench .

You need the ANTLR meta grammar and generate a parser from it. Then use that to load you own grammar into a structure you can use to generate the matrix from.

As a head start you could use the ANTLR grammar parser I have created , but that is for a C/C++ target, not Java. So, you have to get the ANTLR3.g file from the ANTLR homepage and create your own parser from it.