Grammar for lookahead matching backwards when parsing a play

[TakodaS]

from arpeggio.cleanpeg import ParserPEG

grammar = r"""
    grammar = line_end* scene+ EOF
    scene = scene_start action+ scene_end
    scene_start = "SCENE" number+ "BEGIN" line_end
    scene_end= "END SCENE" line_end*
    action = ( background/disappears/appears/dialogue ) line_end+
    background = "BG:" caps_sentence
    appears =  caps_sentence &"APPEARS:"
    disappears =  caps_sentence &"DISAPPEARS:"
    dialogue = caps_sentence ":" text+ 
    caps_sentence = caps_word+
    caps_word   =  caps/number/punctuation 
    text       = punctuation/lower/caps/number
    punctuation = r"[!?.,'’…^]" 
    number      = r"[0-9]"
    caps        = r"[A-Z]"
    lower = r"[a-z]"
    line_end = "\n"
    """

if __name__ == "__main__":
    kwargs = {"ws": "\t\r ", "debug": True}
    input_text = """
    SCENE 01 BEGIN
    BG: SOME BACKGROUND
    CHARACTER APPEARS:
    CHARACTER: Dramatic dialogue!
    CHARACTER DISAPPEARS:
    END SCENE
    """.strip()
    ParserPEG(grammar, "grammar", **kwargs).parse(input_text)

I am trying to parse a play using Arpeggio and can't seem to get the lookahead working satisfactorily. Currently I get the error
arpeggio.NoMatch: Expected punctuation or lower or caps or number at position (3, 23) => 'R APPEARS:* CHARA'.

Which occurs because the appears rule is not being matched since caps_sentence swallows up APPEARS and DISAPPEARS and instead Arpeggio tries to match dialogue which fails. Is there any way to first match APPEARS and DISAPPEARS at the end of the sentence with a lookahead? Many thanks!

[igordejanovic]

Syntactic predicates (&|!) are non-consuming. You need to consume those APPEARS and DISAPPEARS keywords like:

    appears =  caps_sentence "APPEARS:"
    disappears =  caps_sentence "DISAPPEARS:"

But, that brings another problem. caps_sentence will eat those keywords. So to prevent it from eating APPEARS and DISAPPEARS we use negative lookahead. The full working example is (notice the use of !keywords in caps:

from arpeggio.cleanpeg import ParserPEG

grammar = r"""
    grammar = line_end* scene+ EOF
    scene = scene_start action+ scene_end
    scene_start = "SCENE" number+ "BEGIN" line_end
    scene_end= "END SCENE" line_end*
    action = ( background/disappears/appears/dialogue ) line_end+
    background = "BG:" caps_sentence
    appears =  caps_sentence "APPEARS:"
    disappears =  caps_sentence "DISAPPEARS:"
    dialogue = caps_sentence ":" text+
    caps_sentence = caps_word+
    caps_word   =  caps/number/punctuation
    text       = punctuation/lower/caps/number
    punctuation = r"[!?.,'’…^]"
    number      = r"[0-9]"
    keywords = "APPEARS" / "DISAPPEARS"
    caps        = !keywords r"[A-Z]"
    lower = r"[a-z]"
    line_end = "\n"
    """

if __name__ == "__main__":
    kwargs = {"ws": "\t\r ", "debug": True}
    input_text = """
    SCENE 01 BEGIN
    BG: SOME BACKGROUND
    CHARACTER APPEARS:
    CHARACTER: Dramatic dialogue!
    CHARACTER DISAPPEARS:
    END SCENE
    """.strip()
    ParserPEG(grammar, "grammar", **kwargs).parse(input_text)

[TakodaS]

Fantastic thank you!