We Keep Coding

Syntax error in ply(lex/yacc) when input text file is not in sequence with the grammar sequence

The following code works well when the text file are in sequence with code i.e, Introduction then Information but gives error if Information comes before Introduction. What would be the solution to handle this using lex/yacc? Thank in advance.
import ply.lex as lex
# List of token names. This is always required
tokens = [
'CheckupInformation',
'Introduction',
'Information',
'perfect',
'sick',
'LPAREN',
'RPAREN',
'CHAR',
'NUMBER'
]
def t_CheckupInformation(t) : 'CheckupInformation' ; return t
def t_Introduction(t) : 'Introduction' ; return t
def t_Information(t) : 'Information' ; return t
def t_perfect(t): 'perfect'; return t
def t_sick(t) : 'sick'; return t
t_LPAREN = r'\('
t_RPAREN = r'\)'
t_CHAR = r'[a-zA-Z_][a-zA-Z0-9_\-]*'
t_ignore = " \t"
# Define a rule so we can track line numbers
def t_NUMBER(t):
r'[+\-0-9_][0-9_]*'
t.lexer.lineno += len(t.value)
try:
t.value = int(t.value)
except ValueError:
print("Integer value too large %s" % t.value)
t.value = 0
return t
def t_SEMICOLON(t):
r'\;.*'
t.lexer.lineno += len(t.value)
def t_newline(t):
r'\n+'
t.lexer.lineno += len(t.value)
# Error handling rule
def t_error(t):
print("Illegal character '%s'" % t.value[0])
t.lexer.skip(1)
# Build the lexer
lexer = lex.lex()
# define upper level classes first
class stat:
def __init__(self):
self.statement = ""
self.intro = list()
self.body = list()
P=stat()
def p_stat(p):
'Stat : LPAREN CheckupInformation statIntro statBody RPAREN'
p[0]=(p[1],p[2],p[3],p[4],p[5])
def p_Intro(p) :
'''statIntro : LPAREN Introduction Name RPAREN
| statIntro LPAREN Introduction Name RPAREN
| empty'''
if len(p)==5:
p[0] = (p[3])
elif len(p)==6:
p[0] = (p[4])
else:
p[0]= None
P.intro.append(p[0])
def p_Name(p):
'Name : CHAR'
p[0]=p[1]
def p_Body(p):
'''statBody : LPAREN Information bodyinfo RPAREN
| statBody LPAREN Information bodyinfo RPAREN'''
if len(p)==5:
p[0] = (p[3])
elif len(p)==6:
p[0] = (p[4])
P.body.append(p[0])
def p_bodyinfo(p):
'''bodyinfo : LPAREN CHAR perfect RPAREN
| LPAREN CHAR sick RPAREN'''
p[0]=p[2],p[3]
def p_empty(p):
'empty : '
print("This function is called")
pass
def p_error(p):
print("Syntax error in input '%s'!" % p.value)
import ply.yacc as yacc
parser = yacc.yacc()
import sys
if len(sys.argv) < 2 :
sys.exit("Usage: %s <filename>" % sys.argv[0])
fp = open(sys.argv[1])
contents=fp.read()
result=parser.parse(contents)
print("(CheckupInformation")
if (P.intro) != None:
for x in range(len(P.intro)):
print(" (Introduction %s)" %(P.intro[x]))
for x in range(len(P.body)):
print(" (Information( %s %s))" %(P.body[x]))
print(")")
The code works well for file1 & cannot handle file2.
ERROR:
Syntax error in input '(Introduction'!
(CheckupInformation
(Introduction None)
(Information( Anonymous1 perfect))
)
File1:
(CheckupInformation
(Introduction John)
(Introduction Patt)
(Information(Anonymous1 perfect))
(Information(Anonymous2 sick))
)
File2:
(CheckupInformation
(Information(Anonymous1 perfect))
(Information(Anonymous2 sick))
(Introduction John)
(Introduction Patt)
)

This might not be the answer you wanted, but I found myself unable to just change one or two lines in your code. The following is still far from perfect, but I think it is approaching a reasonable approach to your problem. I tried to annotate it with useful comments. Please read through it carefully and try to understand why I did what I did, referring to the Ply manual as necessary (some references are in the code comments, but there's lots of useful background information in the document which I didn't reference specifically).
Good luck.
import ply.lex as lex
# Keyword handling copied from the Ply manual, https://www.dabeaz.com/ply/ply.html#ply_nn6
reserved = {
'CheckupInformation': 'TK_CheckupInformation',
'Introduction': 'TK_Introduction',
'Information': 'TK_Information',
'perfect': 'TK_perfect',
'sick': 'TK_sick',
}
# I changed CHAR to WORD because CHAR sounds like a character
tokens = ['NUMBER','WORD'] + list(reserved.values())
def t_WORD(t):
r'[a-zA-Z_][a-zA-Z0-9_-]*'
t.type = reserved.get(t.value,'WORD') # Check for reserved words
return t
# See the Ply manual: https://www.dabeaz.com/ply/ply.html#ply_nn11
literals = '()'
# See the Ply manual: https://www.dabeaz.com/ply/ply.html#ply_nn8
t_ignore = ' \t\n'
t_ignore_COMMENT = r'\;.*'
# Fixed the regex. You can't have a sign in the middle of a number.
def t_NUMBER(t):
r'[+-]?[0-9_]+'
try:
t.value = int(t.value)
except ValueError:
print("Integer value too large %s" % t.value)
t.value = 0
return t
# See below for the definition of lineno_for_token
# Error handling rule
def t_error(t):
print("Illegal character '%s' at line %d'" % (
t.value[0], t.lexer.lineno_for_token(t)))
t.lexer.skip(1)
# Build the lexer
lexer = lex.lex()
# Ply tracks the character index automatically as lexer.lexpos, and every
# token it produces has a lexpos attribute. So there is no need to misuse
# the lineno attribute for that purpose. It should be the line number of
# the token, as its name indicates.
# You don't seem to use lineno (or lexpos) anywhere, but it is handy for
# error messages. But since it is used rarely, it's easier to compute it
# on demand by counting newlines to the lex position.
# Perhaps this should just be added to the lexer we just built.
lex.Lexer.lineno_for_token = lambda self, t: 1 + self.lexdata.count('\n', 0, t.lexpos)
# Fixed this to use an upper-class name and to derive from object.
# Object to hold a top-level description
class Stat(object):
# Attributes used for components
components = {'intro', 'body'}
def __init__(self, component_dict):
self.statement = "" # I don't know what this is used for
# Copy the components dictionary as attributes, using
# an empty list as default
for k in self.components:
setattr(self, k, component_dict.get(k, ()))
# Verify that we used every key in the dict.
for k in component_dict.keys():
if k not in self.components:
print("Warning! Ignoring " + k
+ " because it is not in Stat.components")
# Arrange for the object to print as expected
def __repr__(self):
return '(CheckupInformation %r %r)' % (self.intro, self.body)
# Instead of having a global "P" object (whose name is not very useful),
# we return a Stat object
def p_stat(p):
""" stat : '(' TK_CheckupInformation components ')' """
p[0] = Stat(p[3])
# We allow all components to be optional and order independent here. We
# also allow them all to be repeated. But that could be made more precise.
# components is a dictionary whose values are lists
def p_components_empty(p):
""" components : """
p[0] = { }
def p_components_append(p):
""" components : components component """
p[0] = p[1]
# The component is a two-element tuple
key, value = p[2]
if key in p[0]:
p[0][key].append(value)
else:
p[0][key] = [value]
# Syntax for each component type (just one element, not a list)
# component is a tuple of (key, value)
# All of the productions just copy the value from some specific syntax.
def p_component(p):
""" component : statIntro
| statBody
"""
p[0] = p[1]
def p_statIntro(p):
"""statIntro : '(' TK_Introduction WORD ')' """
p[0] = ('intro', p[3])
def p_statBody(p):
"""statBody : '(' TK_Information bodyinfo ')' """
p[0] = ('body', p[3])
# bodyinfo is a tuple of (identifier, status)
def p_bodyinfo(p):
"""bodyinfo : '(' WORD TK_perfect ')'
| '(' WORD TK_sick ')'
"""
p[0] = (p[2],p[3])
def p_error(p):
print("Syntax error in input '%s'! at line %d" % (
p.value, p.lexer.lineno_for_token(p)))
import ply.yacc as yacc
parser = yacc.yacc()
# Only do this if we're called from the command line
if __name__ == "__main__":
import sys
if len(sys.argv) < 2 :
sys.exit("Usage: %s <filename>" % sys.argv[0])
with open(sys.argv[1]) as fp:
stat = parser.parse(fp.read())
if stat is not None:
print("(CheckupInformation")
for x in range(len(stat.intro)):
print(" (Introduction %s)" %(stat.intro[x]))
for x in range(len(stat.body)):
print(" (Information( %s %s))" %(stat.body[x]))
print(")")

Generalizing integer calculator example from PLY to floating number

I am reading the first example from
https://github.com/dabeaz/ply
It is a basic calculator allowing for only expression involving '(',')','+','-','*','/', integers and assignement (for instance x=3) and throwing the evaluation of the expression (even it it's result is not an integer, for instance '3/4').
I would like to allow for floating numbers, so that I basically modified the code from the example as follows but it doesn't work :
# -----------------------------------------------------------------------------
# calc.py
#
# A simple calculator with variables.
# -----------------------------------------------------------------------------
tokens = (
'NAME','INTEGER', 'FLOAT',
'PLUS','MINUS','TIMES','DIVIDE','EQUALS',
'LPAREN','RPAREN',
)
# Tokens
t_PLUS = r'\+'
t_MINUS = r'-'
t_TIMES = r'\*'
t_DIVIDE = r'/'
t_EQUALS = r'='
t_LPAREN = r'\('
t_RPAREN = r'\)'
t_NAME = r'[a-zA-Z_][a-zA-Z0-9_]*'
def t_INTEGER(t):
r'\d+'
t.value = int(t.value)
return t
def t_FLOAT(t):
r'/^(?!0\d)\d*(\.\d+)?$/mg'
t.value = float(t.value)
return t
# Ignored characters
t_ignore = " \t"
def t_newline(t):
r'\n+'
t.lexer.lineno += t.value.count("\n")
def t_error(t):
print("Illegal character '%s'" % t.value[0])
t.lexer.skip(1)
# Build the lexer
import ply.lex as lex
lex.lex()
# Precedence rules for the arithmetic operators
precedence = (
('left','PLUS','MINUS'),
('left','TIMES','DIVIDE'),
('right','UMINUS'),
)
# dictionary of names (for storing variables)
names = { }
def p_statement_assign(p):
'statement : NAME EQUALS expression'
names[p[1]] = p[3]
def p_statement_expr(p):
'statement : expression'
print(p[1])
def p_expression_binop(p):
'''expression : expression PLUS expression
| expression MINUS expression
| expression TIMES expression
| expression DIVIDE expression'''
if p[2] == '+' : p[0] = p[1] + p[3]
elif p[2] == '-': p[0] = p[1] - p[3]
elif p[2] == '*': p[0] = p[1] * p[3]
elif p[2] == '/': p[0] = p[1] / p[3]
def p_expression_uminus(p):
'expression : MINUS expression %prec UMINUS'
p[0] = -p[2]
def p_expression_group(p):
'expression : LPAREN expression RPAREN'
p[0] = p[2]
def p_expression_integer(p):
'expression : INTEGER'
p[0] = p[1]
def p_expression_float(p):
'expression : FLOAT'
p[0] = p[1]
def p_expression_name(p):
'expression : NAME'
try:
p[0] = names[p[1]]
except LookupError:
print("Undefined name '%s'" % p[1])
p[0] = 0
def p_error(p):
print("Syntax error at '%s'" % p.value)
import ply.yacc as yacc
yacc.yacc()
while True:
try:
s = input('calc > ')
except EOFError:
break
yacc.parse(s)
I have error :
calc > 3.14+1
Illegal character '.'
Syntax error at '14'

ply parses the T_xxx members in the order of declaration (using reflection on your module). What happens here is that T_INTEGER matches before T_FLOAT. So the integer part of your float is parsed, then ply chokes on the dot.
That would directly work if your regex for floats wasn't off (completely missed that point in my first answer, blinded by the obvious wrong order).
I've simplified it to \d+\.\d+ (which doesn't match 1. or .9 so not the best choice), but you can borrow a better one taken from a similar issue: PLY lexer for numbers always returns double
You have to get T_FLOAT parsed before T_INTEGER. Just swap both declarations to do so:
def t_FLOAT(t):
r'\d+\.\d+'
# a better regex taking exponents into account:
'[-+]?[0-9]+(\.([0-9]+)?([eE][-+]?[0-9]+)?|[eE][-+]?[0-9]+)'
t.value = float(t.value)
return t
def t_INTEGER(t):
r'\d+'
t.value = int(t.value)
return t
As a general rule for ply, do this for all patterns that are longer/more specific than others to avoid conflicts.

You have two problems in your lex file. First is the token order as explained by Jean-François: the longer tokens must be defined first in lex (ref from ply doc.):
When building the master regular expression, rules are added in the following order:
All tokens defined by functions are added in the same order as they appear in the lexer file.
Tokens defined by strings are added next by sorting them in order of decreasing regular expression length (longer expressions are added first).
But the string defining the token shall be a re compatible string. Your FLOAT definition is awfully broken here. If we define a float as composed of exactly one dot, and optional digits before or after the dot and not a dot alone, an acceptable definition could be:
r'(\d*\.\d+)|(\d+\.\d*)'
In particular, the shlashes / shall not be included in the string...

Python PLY Yacc "syntax error"

Okay, so I'm trying to build a parser of my mini-language (obviously), and setting variables seems to be properly working. But as soon as Yacc comes across a function definition, it just gives me a syntax error, and a couple of EOF errors (which I know are from when Yacc has no remaining rules to set) and nothing else happens... Where did I go wrong?
Here's an example of the syntax I'm parsing:
$name = "John Doe"
$age = 72
$waterInOceans = 95.4
!testFunction {
}
Where the !testFunction { } section is defining a function (based off of the exclamation point). I don't know if that's going to be useful in debugging.
# The Lexer
import ply.lex as lex
tokens = ["MINUS", "SEPARATOR", "MODIFIER", "FUNCTION_NAME", "UNDEF_BLOCK", "VARIABLE_NAME", "EQUALS", "STRING", "FLOAT", "INNER_CONTENT", "ARGUMENTS", "INTEGER", "PLUS"]
def t_ARGUMENTS(t): # Finds arguments in calls and function definitions
r'\(.*\)'
t.value = t.value[1:-1] # strip parenthesis
t.value = t.value.split(" && ")
return t
def t_STRING(t): # finds strings
r'"\w.+"'
t.value = t.value[1:-1] # strips the quotation marks of the string
return t
def t_FLOAT(t): # finds floats
r'\d+.\d+'
t.value = float(t.value)
return t
def t_INTEGER(t):
r'\d+'
t.value = int(t.value)
return t
def t_VARIABLE_NAME(t):
r'\$\w*\b'
t.value = t.value[1:]
return t
def t_INNER_CONTENT(t):
r'\{\n.*\n\}|\{.*\}'
t.value = t.value[1:-1]
return t
def t_FUNCTION_NAME(t):
r'!\w+'
t.value = t.value[1:]
return t
t_ignore = r"\n|\t|\r"
t_EQUALS = r"\="
t_PLUS = r"\+"
t_MINUS = r"-"
t_MODIFIER = r"\."
t_SEPARATOR = r"\,"
t_UNDEF_BLOCK = r"\w+" # Any block of text that is left over and isn't assigned by the end (used by functions)
def t_error(t):
t.lexer.skip(1)
lex.lex()
#opened = open("example.zeq", "r")
#content = opened.read()
#opened.close()
#lex.input(content)
And then the Yacc half:
# The Yacc parser
import ply.yacc as yacc
import compiler # Get the compiler (tokenizer; compiler.py) which generates tokens
import sys
from os import system
##############
### IGNORE ###
tokens = compiler.tokens
#system("clear")
print("Executing "+sys.argv[1]+" |\n"+("-"*(len(sys.argv[1])+12)))
### IGNORE ###
##############
VARIABLES = {}
FUNCTIONS = {}
def p_assign(p): # Set new variable
'''assignment : VARIABLE_NAME EQUALS compound
| VARIABLE_NAME EQUALS STRING
| VARIABLE_NAME EQUALS INTEGER
| VARIABLE_NAME EQUALS FLOAT'''
#print("Setting '{}' to '{}'...".format(str(p[1]), str(p[3])))
VARIABLES[p[1]] = p[3]
def p_number(p): # Combines floats and integers into a blanket non-terminal for simplicity sakes
'''number : FLOAT
| INTEGER'''
p[0] = p[1]
def p_compound(p): # Complete the value *before* the variable is assigned!
'''compound : number PLUS number
| number MINUS number'''
type1 = type(p[1])
type2 = type(p[3])
operator = p[2]
if operator == "+":
p[0] = p[1] + p[3]
elif operator == "-":
p[0] = p[1] - p[3]
def p_undefined(p):
'''undefined : UNDEF_BLOCK'''
print("Undefined block")
def p_function(p):
'''function : FUNCTION_NAME INNER_CONTENT'''
print("Creating a function")
name = p[1]
content = p[2]
FUNCTIONS[name] = content
def p_empty(p):
'''empty : '''
#~ def p_error(p):
#~ if p:
#~ print("Syntax error: "+p.type)
#~ else:
#~ pass
parser = yacc.yacc()
opened = open(sys.argv[1], "r")
content = opened.read()
opened.close()
for line in content.splitlines():
parser.parse(line)
print(VARIABLES)
print(FUNCTIONS)
I'm waiting for it to be a simple overlooked detail...

When you ask Ply (or yacc, for that matter) to parse an input, it attempts to recognize a single instance of the top-level non-terminal (or "starting symbol"). This will usually a grammatical description of the entire input, so it will often have a name like program, although there are use cases in which it is useful to parse just a part of the input.
Ply (and yacc) assume that the first grammar production is for the starting symbol. In your case, the first production is assignment, and so that is what it will try to parse (and nothing else). assignment cannot derive a function definition or any other statement type, so those cause syntax errors.
If you want to explicitly tell Ply what the top-level symbol is, you can do so. See the manual section on starting symbols.

Parsing templates with Python-like if-blocks using pyparsing

I'm trying to parse a simple template language with pyparsing which looks kind of like this:
if <expr>: <statements>
elif <expr>:
if <expr>:
<statements>
else:
<statements>
<statements>
In which statements are either if-blocks with being valid Python code or any text otherwise. If a string starts with an "if" keyword, that means it's an if-block right away. In fact, lines of text can also contain interpolations of the type {variable}, but that's about it.
A full valid example of the this syntax:
if a == b: do some {interpolation} stuff
elif a.x > b.y:
if True:
some {more} interpolation
and some more text
else: foo {bar}
arbitrary {text} here
lorem {ipsum}
To detect Python expressions, I've subclassed pyparsing.Token and it sort of seems to work:
import ast
from pyparsing import *
class PythonExpression(Token):
name = 'PythonExpression'
def parseImpl(self, s, loc, doActions=True):
max_loc = s.find('\n') if '\n' in s else len(s)
best_loc = None
for n in range(loc + 1, max_loc + 1):
try:
tree = ast.parse(s[loc:n])
except:
continue
if isinstance(tree, ast.Module):
if len(tree.body) is 1:
if isinstance(tree.body[0], ast.Expr):
best_loc = n
if best_loc is not None:
return best_loc, s[loc:best_loc]
raise ParseException(s, loc, 'invalid Python expression')
expr = 'if foo[1 : "bar:baz"] == 1 : passqwe'
print (Keyword('if') + PythonExpression()).parseString(expr).asList()
results in
['if', 'foo[1 : "bar:baz"] == 1 ']
However, I'm a bit lost with the use of pyparsing.indentedBlock, cannot seem to make it parse the entire grammar no matter what. My last attempt is this (note that it only contains the if statement implementation; there are also optional elif and else blocks):
ParserElement.setDefaultWhitespaceChars(' \t')
colon = Literal(':').suppress()
if_clause = PythonExpression() + colon
if_statement = Group(Keyword('if') + if_clause)
non_white = Regex(r'\S+')
anything = Combine(non_white + restOfLine)
statement = Forward()
indent_stack = [1]
if_block = Group(if_statement + ((anything) | indentedBlock(statement, indent_stack)))
other = ~Keyword('if') + anything
statement << (if_block | other)
parser = OneOrMore(indentedBlock(statement, indent_stack, False))
data = """\
if foo[1:2]:
bar
baz
if foo[3]:
bar : baz
if foo[4]: bar
baz
foo
"""
pprint.pprint(parser.parseString(data).asList())
Which starts parsing correctly but then stops:
[[[[['if', 'foo[1:2]'], [['bar'], ['baz']]]]]]
I've also tried explicitly adding + lineEnd.suppress() to anything but that didn't seem to help. I'm sure I'm doing something stupid here, probably related to newlines, but can't really figure it out.
On a side note, how do I detect interpolation patterns in anything in the example above (if the example above was to work), so that foo {bar} baz gets parsed to a ['foo', Var('bar'), 'baz']? It's easy to detect {var}, but what's the correct expression for plain text then that's not greedy enough to consume everything thrown at it and that doesn't mess up the if/elif/else logic (I tried using SkipTo but that became quite cumbersome)?
EDIT: adding the separate grammar for parsing the interpolation with an example:
class Substitution(object):
def __init__(self, s, l, t):
self.name = t[0]
def __repr__(self):
return 'Substitution(%r)' % self.name
ParserElement.setDefaultWhitespaceChars(' \t')
lbrace = Literal('{').suppress()
rbrace = Literal('}').suppress()
name = Word(alphas, alphanums + '_')
substitution = Combine(lbrace + name + rbrace).setParseAction(Substitution)
text = SkipTo(substitution | lineEnd.suppress(), include=True).leaveWhitespace()
parser = OneOrMore(text | substitution)
parser.parseString('hello \n {world} {invalid.sub} \n foo {bar} baz ').asList()
outputs
['hello ',
[' ', Substitution('world')],
' {invalid.sub} ',
[' foo ', Substitution('bar')],
' baz ']
As you can see, in this case the parser is not being quite correct at grouping lines together (it doesn't allow text to follow substitution somewhy), but it shows the point. These Substitution objects are then to be later processed at runtime.

Recursive parsing for lisp like syntax

im trying to parse lines in the form:
(OP something something (OP something something ) ) ( OP something something )
Where OP is a symbol for a logical gate (AND, OR, NOT) and something is the thing i want to evaluate.
The output im looking for is something like:
{ 'OPERATOR': [condition1, condition2, .. , conditionN] }
Where a condition itself can be a dict/list pair itself (nested conditions). So far i tried something like:
tree = dict()
cond = list()
tree[OP] = cond
for string in conditions:
self.counter += 1
if string.startswith('('):
try:
OP = string[1]
except IndexError:
OP = 'AND'
finally:
if OP == '?':
OP = 'OR'
elif OP == '!':
OP = 'N'
# Recurse
cond.append(self.parse_conditions(conditions[self.counter:], OP))
break
elif not string.endswith(")"):
cond.append(string)
else:
return tree
return tree
I tried other ways aswell but i just can't wrap my head around this whole recursion thing so im wondering if i could get some pointers here, i looked around the web and i found some stuff about recursive descent parsing but the tutorials were all trying to do something more complicated than i needed.
PS: i realize i could do this with existing python libraries but what would i learn by doing that eh?

I'm posting this without further comments, for learning purposes (in the real life please do use a library). Note that there's no error checking (a homework for you!)
Feel free to ask if there's something you don't understand.
# PART 1. The Lexer
symbols = None
def read(input):
global symbols
import re
symbols = re.findall(r'\w+|[()]', input)
def getsym():
global symbols
return symbols[0] if symbols else None
def popsym():
global symbols
return symbols.pop(0)
# PART 2. The Parser
# Built upon the following grammar:
#
# program = expr*
# expr = '(' func args ')'
# func = AND|OR|NOT
# args = arg*
# arg = string|expr
# string = [a..z]
def program():
r = []
while getsym():
r.append(expr())
return r
def expr():
popsym() # (
f = func()
a = args()
popsym() # )
return {f: a}
def func():
return popsym()
def args():
r = []
while getsym() != ')':
r.append(arg())
return r
def arg():
if getsym() == '(':
return expr()
return string()
def string():
return popsym()
# TEST = Lexer + Parser
def parse(input):
read(input)
return program()
print parse('(AND a b (OR c d)) (NOT foo) (AND (OR x y))')
# [{'AND': ['a', 'b', {'OR': ['c', 'd']}]}, {'NOT': ['foo']}, {'AND': [{'OR': ['x', 'y']}]}]