如何在 Python3 的 ANTLR4 中捕获无关输入? [英] How to catch extraneous input in ANTLR4 in Python3?

问题描述

是的，这几乎是 如何在 ANTLR4 中捕获无关输入的重复? - 但在 Java 的情况下，我需要一个 Python 解决方案，而移植 Java 解决方案对我不起作用！

Yes, this is nearly a duplicate of How to catch extraneous input in ANTLR4? - but that is in the case of Java, I need a solution for Python, and porting the Java solution is just not working for me!

ANTLR4 正在写入以下错误":到控制台.我希望这导致硬停止，但解析器继续并最终重新同步:

ANTLR4 is writing the following "error" to the console. I want this to result in a hard stop, but the parser continues and eventually re-syncs:

line 12:28 extraneous input 'FREDDY' expecting {'AS', ...

我希望这是一个致命的错误.

I want this to be a fatal error.

因此，我定义了自己的错误侦听器，并覆盖了所有 4 个方法:

So, I've defined my own Error Listener, and overriden all 4 methods:

class MyErrorListener( ErrorListener ):

    __slots__ = [ 'file_name' ]
    
    def __init__(self, file_name="<stdin>"):
        self.file_name = file_name

    def syntaxError(self, recognizer, offendingSymbol, line, column, msg, e):
        print(f"{self.file_name} {line}:{column} {msg[0:min(80,len(msg)-1)]}", file=sys.stderr)
        sys.exit(1)
        
    def reportAmbiguity(self, recognizer, dfa, startIndex, stopIndex, exact, ambigAlts, configs):
        print(f"{self.file_name} {startIndex}-{stopIndex} reportAmbiguity {ambigAlts}", file=sys.stderr)
        sys.exit(1)

    def reportAttemptingFullContext(self, recognizer, dfa, startIndex, stopIndex, conflictingAlts, configs):
        print(f"{self.file_name} {startIndex}-{stopIndex} conflictingAlts {conflictingAlts}", file=sys.stderr)
        sys.exit(1)

    def reportContextSensitivity(self, recognizer, dfa, startIndex, stopIndex, prediction, configs):
        print(f"{self.file_name} {startIndex}-{stopIndex} ContextSensitivity {prediction}", file=sys.stderr)
        sys.exit(1)

我创建了我的词法分析器，删除了默认的控制台错误侦听器，并将我的添加到:

I create my lexer, remove the default console error listener, and add mine in:

        st = InputStream(upper_case)
        lexer:PlSqlLexer = PlSqlLexer(st)
        # Add your error listener to the lexer if required
        lexer.removeErrorListeners()
        lexer.addErrorListener( MyErrorListener(f) )
        stream = CommonTokenStream(lexer)
        parser = PlSqlParser(stream)
        tree = parser.sql_script()
        printer = PlSqlPrintListener(txt)
        walker = ParseTreeWalker()
        walker.walk(printer, tree)

而这是捕捉一些错误，例如:

And this is catching some errors, such as:

sql_parser.py 1:0 token recognition error at: '"""\n

(那是我将 Python 程序输入 SQL 解析器 - 当然它不应该工作！)

(That's me feeding a Python program into an SQL Parser - of course it shouldn't work!)

但我的错误列表器没有捕捉到我在顶部报告的错误.

But my Error Listner isn't catching the error that I reported at the top.

有什么建议吗?

• ANTLR 解析器生成器版本 4.9.2
• Python 3.7.3
• antlr4-python3-runtime v4.9.2

推荐答案

解决方案；词法分析器和解析器都是识别器".识别器是听众收听的地方.

The solution; both the Lexer and the Parser are "Recognizers". The Recognizer is where the listener listens.

因此，我们有两种类型的错误，词法错误和解析错误.我只是将我的听众附加到 Lexer，我还需要将它附加到 Parser.修改后的代码:

In consequence, we have two types of errors, Lexing errors and Parsing errors. I was only attaching my listener to the Lexer, I needed to also attach it to the Parser. Revised code:

        ...
        lexer:PlSqlLexer = PlSqlLexer(st)
        # Add your error listener to the lexer if required
        lexer.removeErrorListeners()
        lexer.addErrorListener( my_listener )
        
        stream = CommonTokenStream(lexer)
        
        parser = PlSqlParser(stream)
        # Add your error listener to the _parser_ if required
        parser.removeErrorListeners()
        parser.addErrorListener( my_listener )
        ...

这篇关于如何在 Python3 的 ANTLR4 中捕获无关输入?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！