ANTLR4. Accessing context objects with actions - antlr

Question

Parsing input files with size bigger than computer memory requires three steps:
The use of unbuffered character and token streams
Copy text out of sliding buffer and store in tokens
Ask the parser not to create parse trees
The objective is a simplified grammar to parse a file composed of many lines, each of one composed of many words.
With this requirement in mind, the following code uses #members action and sub classing the parser generated by ANTLR
The method printPagesAndWords receives a List of LineContext objects.
It prints the total number of lines (with the start method provided by LineContext), but it cannot access to the number of WORDs per line, whith the method WORD() provided by the LineContext object.
This is the output obtained:
Number of lines in page: 3
Line start token: word1
Number of words in line: 0
Line start token: 2323
Number of words in line: 0
Line start token: 554545
Number of words in line: 0
Furthermore, If I try to get the WORDs in a line, for example changing the line
System.out.println("Number of words in line:\t"+row.WORD().size()+"\n");
by the line
System.out.println("Number of words in line:\t"+row.WORD(1).getText()+"\n");
The following exception is thrown:
Exception in thread "main" java.lang.NullPointerException
at TestContext.printPagesAndWords(TestContext.java:11)
at ContextParser.read(ContextParser.java:130)
at Main.main(Main.java:11)
The following is the complete set of files:
ContextLexer.g4
lexer grammar ContextLexer;
NL
: '\r'? '\n'
;
WORD
: ~[ \t\n\r]+
;
SPACE
: [ \t] -> skip
;
ContextParser.g4
parser grammar ContextParser;
options {
tokenVocab=ContextLexer;
}
#members{
void printPagesAndWords(List<LineContext> rows){};
}
read
: dataLine+=line* {printPagesAndWords($dataLine);}
;
line
: WORD* NL
;
TestContext which extends ContextParser
import org.antlr.v4.runtime.TokenStream;
import java.util.List;
public class TestContext extends ContextParser {
public TestContext(TokenStream input) {
super(input);
}
void printPagesAndWords(List<LineContext> rows){
System.out.println("Number of lines in page:\t" + rows.size()+"\n");
for(LineContext row: rows){
System.out.println("Line start token:\t\t\t"+row.start.getText());
System.out.println("Number of words in line:\t"+row.WORD().size()+"\n");
}
};
}
The Main Class:
import org.antlr.v4.runtime.*;
import java.io.IOException;
public class Main {
public static void main(String[] args) throws IOException {
String source = "word1 word2 number anothernumber \n 2323 55r ere\n554545 lll 545\n";
ContextLexer lexer = new ContextLexer(CharStreams.fromString(source));
lexer.setTokenFactory(new CommonTokenFactory(true));
TokenStream tokens = new UnbufferedTokenStream<CommonToken>(lexer);
TestContext parser = new TestContext(tokens);
parser.setBuildParseTree(false);
parser.read();
}
}