[Python] PLY Error: Symbol 'NOT' used, but not defined as a token or a rule

I'm encountering an issue with PLY (Python Lex-Yacc) where it's failing to recognize a token in the parser that I've defined. This error is preventing the parser from being built.

• I'm using PLY 3.11 (the latest available on PyPI)
• I'm using Python 3.13

When trying to build the parser, I receive the following error:

ERROR: C:\Users\###\Desktop\Scratch\scrybe\src\scrybe\scriptparser\parser.py:236: Symbol 'NOT' used, but not defined as a token or a rule
WARNING: Token 'COMMENT' defined, but not used
WARNING: Token 'NEWLINE' defined, but not used
WARNING: Token 'UMINUS' defined, but not used
WARNING: There are 3 unused tokens
Traceback (most recent call last):
...
parser = yacc.yacc()
File "C:\Program Files\Python313\Lib\site-packages\ply\yacc.py", line 3432, in yacc
raise YaccError('Unable to build parser')
ply.yacc.YaccError: Unable to build parser


Here is the relevant code for my lexer:

reserved = {
...
"this": "THIS",
"not": "NOT",
"and": "AND",
...
}

tokens.extend(reserved.values())

...

def t_SYMBOL(token):
r"\b[a-zA-Z_]\w*\b"
token.type = reserved.get(token.value, "VARIABLE")
return token


Here is my precedence table in parser.py:

precedence = (
...
("left", "AND"),
("right", "NOT"),
("left", "IN"),
...
)


And here is the code from my parser which generates the error:

def p_logical_operation(prod):
"""logical_operation : expression LESSTHAN expression
| expression GREATERTHAN expression
| expression LESSTHANEQUAL expression
| expression GREATERTHANEQUAL expression
| expression EQUALTO expression
| expression NOTEQUALTO expression
| expression AND expression
| expression OR expression
| expression IN expression
| NOT expression"""
prod[0] = {
"lexpos": prod.lexpos(1) if len(prod) == 3 else prod[1]["lexpos"] if isinstance(prod[1], dict) else prod.lexpos(2),
"type": "logical operation",
"condition": prod[2] if len(prod) == 4 else prod[1],
}
if len(prod) == 4:
prod[0].update({
"comparand 1": prod[1],
"comparand 2": prod[3]
})
else:
prod[0].update({"comparand": prod[2]})


I've confirmed that the lexer is producing the expected tokens when run separately on a piece of test input.

Question: Why is PLY unable to recognize the 'NOT' token despite it being defined in the reserved dictionary, especially given that I can replace it with any other reserved keyword and have it work fine? How can I resolve this issue and successfully build the parser? Any help or insights would be greatly appreciated.

Continue reading...