|
# module pyparsing.py # # Copyright (c) 2003-2018 Paul T. McGuire # # Permission is hereby granted, free of charge, to any person obtaining # a copy of this software and associated documentation files (the # "Software"), to deal in the Software without restriction, including # without limitation the rights to use, copy, modify, merge, publish, # distribute, sublicense, and/or sell copies of the Software, and to # permit persons to whom the Software is furnished to do so, subject to # the following conditions: # # The above copyright notice and this permission notice shall be # included in all copies or substantial portions of the Software. # # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, # EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF # MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. # IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY # CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, # TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE # SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. #
""" pyparsing module - Classes and methods to define and execute parsing grammars =============================================================================
The pyparsing module is an alternative approach to creating and executing simple grammars, vs. the traditional lex/yacc approach, or the use of regular expressions. With pyparsing, you don't need to learn a new syntax for defining grammars or matching expressions - the parsing module provides a library of classes that you use to construct the grammar directly in Python.
Here is a program to parse "Hello, World!" (or any greeting of the form C{"<salutation>, <addressee>!"}), built up using L{Word}, L{Literal}, and L{And} elements (L{'+'<ParserElement.__add__>} operator gives L{And} expressions, strings are auto-converted to L{Literal} expressions)::
from pyparsing import Word, alphas
# define grammar of a greeting greet = Word(alphas) + "," + Word(alphas) + "!"
hello = "Hello, World!" print (hello, "->", greet.parseString(hello))
The program outputs the following::
Hello, World! -> ['Hello', ',', 'World', '!']
The Python representation of the grammar is quite readable, owing to the self-explanatory class names, and the use of '+', '|' and '^' operators.
The L{ParseResults} object returned from L{ParserElement.parseString<ParserElement.parseString>} can be accessed as a nested list, a dictionary, or an object with named attributes.
The pyparsing module handles some of the problems that are typically vexing when writing text parsers: - extra or missing whitespace (the above program will also handle "Hello,World!", "Hello , World !", etc.) - quoted strings - embedded comments
Getting Started - ----------------- Visit the classes L{ParserElement} and L{ParseResults} to see the base classes that most other pyparsing classes inherit from. Use the docstrings for examples of how to: - construct literal match expressions from L{Literal} and L{CaselessLiteral} classes - construct character word-group expressions using the L{Word} class - see how to create repetitive expressions using L{ZeroOrMore} and L{OneOrMore} classes - use L{'+'<And>}, L{'|'<MatchFirst>}, L{'^'<Or>}, and L{'&'<Each>} operators to combine simple expressions into more complex ones - associate names with your parsed results using L{ParserElement.setResultsName} - find some helpful expression short-cuts like L{delimitedList} and L{oneOf} - find more useful common expressions in the L{pyparsing_common} namespace class """
except ImportError: from threading import RLock
# Python 3 except ImportError: # Python 2.7 from collections import Iterable from collections import MutableMapping
except ImportError: try: from ordereddict import OrderedDict as _OrderedDict except ImportError: _OrderedDict = None
#~ sys.stderr.write( "testing pyparsing module, version %s, %s\n" % (__version__,__versionTime__ ) )
'And', 'CaselessKeyword', 'CaselessLiteral', 'CharsNotIn', 'Combine', 'Dict', 'Each', 'Empty', 'FollowedBy', 'Forward', 'GoToColumn', 'Group', 'Keyword', 'LineEnd', 'LineStart', 'Literal', 'MatchFirst', 'NoMatch', 'NotAny', 'OneOrMore', 'OnlyOnce', 'Optional', 'Or', 'ParseBaseException', 'ParseElementEnhance', 'ParseException', 'ParseExpression', 'ParseFatalException', 'ParseResults', 'ParseSyntaxException', 'ParserElement', 'QuotedString', 'RecursiveGrammarException', 'Regex', 'SkipTo', 'StringEnd', 'StringStart', 'Suppress', 'Token', 'TokenConverter', 'White', 'Word', 'WordEnd', 'WordStart', 'ZeroOrMore', 'alphanums', 'alphas', 'alphas8bit', 'anyCloseTag', 'anyOpenTag', 'cStyleComment', 'col', 'commaSeparatedList', 'commonHTMLEntity', 'countedArray', 'cppStyleComment', 'dblQuotedString', 'dblSlashComment', 'delimitedList', 'dictOf', 'downcaseTokens', 'empty', 'hexnums', 'htmlComment', 'javaStyleComment', 'line', 'lineEnd', 'lineStart', 'lineno', 'makeHTMLTags', 'makeXMLTags', 'matchOnlyAtCol', 'matchPreviousExpr', 'matchPreviousLiteral', 'nestedExpr', 'nullDebugAction', 'nums', 'oneOf', 'opAssoc', 'operatorPrecedence', 'printables', 'punc8bit', 'pythonStyleComment', 'quotedString', 'removeQuotes', 'replaceHTMLEntity', 'replaceWith', 'restOfLine', 'sglQuotedString', 'srange', 'stringEnd', 'stringStart', 'traceParseAction', 'unicodeString', 'upcaseTokens', 'withAttribute', 'indentedBlock', 'originalTextFor', 'ungroup', 'infixNotation','locatedExpr', 'withClass', 'CloseMatch', 'tokenMap', 'pyparsing_common', ]
# build list of single arg builtins, that can be used as parse actions
else: _MAX_INT = sys.maxint range = xrange
def _ustr(obj): """Drop-in replacement for str(obj) that tries to be Unicode friendly. It first tries str(obj). If that fails with a UnicodeEncodeError, then it tries unicode(obj). It then < returns the unicode object | encodes it with the default encoding | ... >. """ if isinstance(obj,unicode): return obj
try: # If this works, then _ustr(obj) has the same behaviour as str(obj), so # it won't break any existing code. return str(obj)
except UnicodeEncodeError: # Else encode it ret = unicode(obj).encode(sys.getdefaultencoding(), 'xmlcharrefreplace') xmlcharref = Regex(r'&#\d+;') xmlcharref.setParseAction(lambda t: '\\u' + hex(int(t[0][2:-1]))[2:]) return xmlcharref.transformString(ret)
# build list of single arg builtins, tolerant of Python version, that can be used as parse actions singleArgBuiltins = [] import __builtin__ for fname in "sum len sorted reversed list tuple set any all min max".split(): try: singleArgBuiltins.append(getattr(__builtin__,fname)) except AttributeError: continue
"""Escape &, <, >, ", ', etc. in a string of data."""
# ampersand must be replaced first from_symbols = '&><"\'' to_symbols = ('&'+s+';' for s in "amp gt lt quot apos".split()) for from_,to_ in zip(from_symbols, to_symbols): data = data.replace(from_, to_) return data
"""base exception class for all parsing runtime exceptions""" # Performance tuning: we construct a *lot* of these, so keep this # constructor as small and fast as possible self.msg = pstr self.pstr = "" else:
def _from_exception(cls, pe): """ internal factory method to simplify creating one type of ParseException from another - avoids having __init__ signature conflicts among subclasses """ return cls(pe.pstr, pe.loc, pe.msg, pe.parserElement)
"""supported attributes by name are: - lineno - returns the line number of the exception text - col - returns the column number of the exception text - line - returns the line containing the exception text """ if( aname == "lineno" ): return lineno( self.loc, self.pstr ) elif( aname in ("col", "column") ): return col( self.loc, self.pstr ) elif( aname == "line" ): return line( self.loc, self.pstr ) else: raise AttributeError(aname)
def __str__( self ): return "%s (at char %d), (line:%d, col:%d)" % \ ( self.msg, self.loc, self.lineno, self.column ) def __repr__( self ): return _ustr(self) """Extracts the exception line from the input string, and marks the location of the exception with a special symbol. """ line_str = self.line line_column = self.column - 1 if markerString: line_str = "".join((line_str[:line_column], markerString, line_str[line_column:])) return line_str.strip() return "lineno col line".split() + dir(type(self))
""" Exception thrown when parse expressions don't match class; supported attributes by name are: - lineno - returns the line number of the exception text - col - returns the column number of the exception text - line - returns the line containing the exception text
Example:: try: Word(nums).setName("integer").parseString("ABC") except ParseException as pe: print(pe) print("column: {}".format(pe.col))
prints:: Expected integer (at char 0), (line:1, col:1) column: 1 """
"""user-throwable exception thrown when inconsistent parse content is found; stops all parsing immediately"""
"""just like L{ParseFatalException}, but thrown internally when an L{ErrorStop<And._ErrorStop>} ('-' operator) indicates that parsing is to stop immediately because an unbacktrackable syntax error has been found"""
#~ class ReparseException(ParseBaseException): #~ """Experimental class - parse actions can raise this exception to cause #~ pyparsing to reparse the input string: #~ - with a modified input string, and/or #~ - with a modified start location #~ Set the values of the ReparseException in the constructor, and raise the #~ exception in a parse action to cause pyparsing to use the new string/location. #~ Setting the values as None causes no change to be made. #~ """ #~ def __init_( self, newstring, restartLoc ): #~ self.newParseText = newstring #~ self.reparseLoc = restartLoc
"""exception thrown by L{ParserElement.validate} if the grammar could be improperly recursive""" self.parseElementTrace = parseElementList
def __str__( self ): return "RecursiveGrammarException: %s" % self.parseElementTrace
def __repr__(self): return repr(self.tup[0]) self.tup = (self.tup[0],i)
""" Structured parse results, to provide multiple means of access to the parsed data: - as a list (C{len(results)}) - by list index (C{results[0], results[1]}, etc.) - by attribute (C{results.<resultsName>} - see L{ParserElement.setResultsName})
Example:: integer = Word(nums) date_str = (integer.setResultsName("year") + '/' + integer.setResultsName("month") + '/' + integer.setResultsName("day")) # equivalent form: # date_str = integer("year") + '/' + integer("month") + '/' + integer("day")
# parseString returns a ParseResults object result = date_str.parseString("1999/12/31")
def test(s, fn=repr): print("%s -> %s" % (s, fn(eval(s)))) test("list(result)") test("result[0]") test("result['month']") test("result.day") test("'month' in result") test("'minutes' in result") test("result.dump()", str) prints:: list(result) -> ['1999', '/', '12', '/', '31'] result[0] -> '1999' result['month'] -> '12' result.day -> '31' 'month' in result -> True 'minutes' in result -> False result.dump() -> ['1999', '/', '12', '/', '31'] - day: 31 - month: 12 - year: 1999 """
# Performance tuning: we construct a *lot* of these, so keep this # constructor as small and fast as possible toklist = [] self.__toklist = list(toklist) else:
self.__accumNames[name] = 0 name = _ustr(name) # will always return a str, but use _ustr for consistency toklist = [ toklist ] else: else:
else: else: return ParseResults([ v[0] for v in self.__tokdict[i] ])
self.__toklist[k] = v sub = v else:
# convert int to slice if i < 0: i += mylen i = slice(i, i+1) # get removed indices # fixup indices in token dictionary for j in removed: for k, (value, position) in enumerate(occurrences): occurrences[k] = _ParseResultsWithOffset(value, position - (position > j)) else: del self.__tokdict[i]
return k in self.__tokdict
if hasattr(self.__tokdict, "iterkeys"): return self.__tokdict.iterkeys() else: return iter(self.__tokdict)
return (self[k] for k in self._iterkeys())
return ((k, self[k]) for k in self._iterkeys())
"""Returns an iterator of all named result keys (Python 3.x only)."""
"""Returns an iterator of all named result values (Python 3.x only)."""
else: iterkeys = _iterkeys """Returns an iterator of all named result keys (Python 2.x only)."""
itervalues = _itervalues """Returns an iterator of all named result values (Python 2.x only)."""
iteritems = _iteritems """Returns an iterator of all named result key-value tuples (Python 2.x only)."""
def keys( self ): """Returns all named result keys (as a list in Python 2.x, as an iterator in Python 3.x).""" return list(self.iterkeys())
def values( self ): """Returns all named result values (as a list in Python 2.x, as an iterator in Python 3.x).""" return list(self.itervalues())
def items( self ): """Returns all named result key-values (as a list of tuples in Python 2.x, as an iterator in Python 3.x).""" return list(self.iteritems())
"""Since keys() returns an iterator, this method is helpful in bypassing code that looks for the existence of any defined results names."""
""" Removes and returns item at specified index (default=C{last}). Supports both C{list} and C{dict} semantics for C{pop()}. If passed no argument or an integer argument, it will use C{list} semantics and pop tokens from the list of parsed tokens. If passed a non-integer argument (most likely a string), it will use C{dict} semantics and pop the corresponding value from any defined results names. A second default return value argument is supported, just as in C{dict.pop()}.
Example:: def remove_first(tokens): tokens.pop(0) print(OneOrMore(Word(nums)).parseString("0 123 321")) # -> ['0', '123', '321'] print(OneOrMore(Word(nums)).addParseAction(remove_first).parseString("0 123 321")) # -> ['123', '321']
label = Word(alphas) patt = label("LABEL") + OneOrMore(Word(nums)) print(patt.parseString("AAB 123 321").dump())
# Use pop() in a parse action to remove named result (note that corresponding value is not # removed from list form of results) def remove_LABEL(tokens): tokens.pop("LABEL") return tokens patt.addParseAction(remove_LABEL) print(patt.parseString("AAB 123 321").dump()) prints:: ['AAB', '123', '321'] - LABEL: AAB
['AAB', '123', '321'] """ if not args: args = [-1] for k,v in kwargs.items(): if k == 'default': args = (args[0], v) else: raise TypeError("pop() got an unexpected keyword argument '%s'" % k) if (isinstance(args[0], int) or len(args) == 1 or args[0] in self): index = args[0] ret = self[index] del self[index] return ret else: defaultvalue = args[1] return defaultvalue
""" Returns named result matching the given key, or if there is no such name, then returns the given C{defaultValue} or C{None} if no C{defaultValue} is specified.
Similar to C{dict.get()}.
Example:: integer = Word(nums) date_str = integer("year") + '/' + integer("month") + '/' + integer("day")
result = date_str.parseString("1999/12/31") print(result.get("year")) # -> '1999' print(result.get("hour", "not specified")) # -> 'not specified' print(result.get("hour")) # -> None """ if key in self: return self[key] else: return defaultValue
""" Inserts new element at location index in the list of parsed tokens.
Similar to C{list.insert()}.
Example:: print(OneOrMore(Word(nums)).parseString("0 123 321")) # -> ['0', '123', '321']
# use a parse action to insert the parse location in the front of the parsed results def insert_locn(locn, tokens): tokens.insert(0, locn) print(OneOrMore(Word(nums)).addParseAction(insert_locn).parseString("0 123 321")) # -> [0, '0', '123', '321'] """ self.__toklist.insert(index, insStr) # fixup indices in token dictionary for name,occurrences in self.__tokdict.items(): for k, (value, position) in enumerate(occurrences): occurrences[k] = _ParseResultsWithOffset(value, position + (position > index))
""" Add single element to end of ParseResults list of elements.
Example:: print(OneOrMore(Word(nums)).parseString("0 123 321")) # -> ['0', '123', '321']
# use a parse action to compute the sum of the parsed integers, and add it to the end def append_sum(tokens): tokens.append(sum(map(int, tokens))) print(OneOrMore(Word(nums)).addParseAction(append_sum).parseString("0 123 321")) # -> ['0', '123', '321', 444] """ self.__toklist.append(item)
""" Add sequence of elements to end of ParseResults list of elements.
Example:: patt = OneOrMore(Word(alphas))
# use a parse action to append the reverse of the matched strings, to make a palindrome def make_palindrome(tokens): tokens.extend(reversed([t[::-1] for t in tokens])) return ''.join(tokens) print(patt.addParseAction(make_palindrome).parseString("lskdj sdlkjf lksd")) # -> 'lskdjsdlkjflksddsklfjkldsjdksl' """ if isinstance(itemseq, ParseResults): self += itemseq else: self.__toklist.extend(itemseq)
""" Clear all elements and results names. """ del self.__toklist[:] self.__tokdict.clear()
if name in self.__tokdict: if name not in self.__accumNames: return self.__tokdict[name][-1][0] else: return ParseResults([ v[0] for v in self.__tokdict[name] ]) else: return ""
ret = self.copy() ret += other return ret
for (k,vlist) in otheritems for v in vlist]
if isinstance(other,int) and other == 0: # useful for merging many ParseResults using sum() builtin return self.copy() else: # this may raise a TypeError - so be it return other + self
def __repr__( self ): return "(%s, %s)" % ( repr( self.__toklist ), repr( self.__tokdict ) )
def __str__( self ): return '[' + ', '.join(_ustr(i) if isinstance(i, ParseResults) else repr(i) for i in self.__toklist) + ']'
out.append(sep) out += item._asStringList() else:
""" Returns the parse results as a nested list of matching tokens, all converted to strings.
Example:: patt = OneOrMore(Word(alphas)) result = patt.parseString("sldkj lsdkj sldkj") # even though the result prints in string-like form, it is actually a pyparsing ParseResults print(type(result), result) # -> <class 'pyparsing.ParseResults'> ['sldkj', 'lsdkj', 'sldkj']
# Use asList() to create an actual list result_list = result.asList() print(type(result_list), result_list) # -> <class 'list'> ['sldkj', 'lsdkj', 'sldkj'] """
""" Returns the named parse results as a nested dictionary.
Example:: integer = Word(nums) date_str = integer("year") + '/' + integer("month") + '/' + integer("day")
result = date_str.parseString('12/31/1999') print(type(result), repr(result)) # -> <class 'pyparsing.ParseResults'> (['12', '/', '31', '/', '1999'], {'day': [('1999', 4)], 'year': [('12', 0)], 'month': [('31', 2)]})
result_dict = result.asDict() print(type(result_dict), repr(result_dict)) # -> <class 'dict'> {'day': '1999', 'year': '12', 'month': '31'}
# even though a ParseResults supports dict-like access, sometime you just need to have a dict import json print(json.dumps(result)) # -> Exception: TypeError: ... is not JSON serializable print(json.dumps(result.asDict())) # -> {"month": "31", "day": "1999", "year": "12"} """ if PY_3: item_fn = self.items else: item_fn = self.iteritems
def toItem(obj): if isinstance(obj, ParseResults): if obj.haskeys(): return obj.asDict() else: return [toItem(v) for v in obj] else: return obj
return dict((k,toItem(v)) for k,v in item_fn())
""" Returns a new copy of a C{ParseResults} object. """
""" (Deprecated) Returns the parse results as XML. Tags are created for tokens and lists that have defined results names. """ nl = "\n" out = [] namedItems = dict((v[1],k) for (k,vlist) in self.__tokdict.items() for v in vlist) nextLevelIndent = indent + " "
# collapse out indents if formatting is not desired if not formatted: indent = "" nextLevelIndent = "" nl = ""
selfTag = None if doctag is not None: selfTag = doctag else: if self.__name: selfTag = self.__name
if not selfTag: if namedItemsOnly: return "" else: selfTag = "ITEM"
out += [ nl, indent, "<", selfTag, ">" ]
for i,res in enumerate(self.__toklist): if isinstance(res,ParseResults): if i in namedItems: out += [ res.asXML(namedItems[i], namedItemsOnly and doctag is None, nextLevelIndent, formatted)] else: out += [ res.asXML(None, namedItemsOnly and doctag is None, nextLevelIndent, formatted)] else: # individual token, see if there is a name for it resTag = None if i in namedItems: resTag = namedItems[i] if not resTag: if namedItemsOnly: continue else: resTag = "ITEM" xmlBodyText = _xml_escape(_ustr(res)) out += [ nl, nextLevelIndent, "<", resTag, ">", xmlBodyText, "</", resTag, ">" ]
out += [ nl, indent, "</", selfTag, ">" ] return "".join(out)
for k,vlist in self.__tokdict.items(): for v,loc in vlist: if sub is v: return k return None
r""" Returns the results name for this token expression. Useful when several different expressions might match at a particular location.
Example:: integer = Word(nums) ssn_expr = Regex(r"\d\d\d-\d\d-\d\d\d\d") house_number_expr = Suppress('#') + Word(nums, alphanums) user_data = (Group(house_number_expr)("house_number") | Group(ssn_expr)("ssn") | Group(integer)("age")) user_info = OneOrMore(user_data)
result = user_info.parseString("22 111-22-3333 #221B") for item in result: print(item.getName(), ':', item[0]) prints:: age : 22 ssn : 111-22-3333 house_number : 221B """ if self.__name: return self.__name elif self.__parent: par = self.__parent() if par: return par.__lookup(self) else: return None elif (len(self) == 1 and len(self.__tokdict) == 1 and next(iter(self.__tokdict.values()))[0][1] in (0,-1)): return next(iter(self.__tokdict.keys())) else: return None
""" Diagnostic method for listing out the contents of a C{ParseResults}. Accepts an optional C{indent} argument so that this string can be embedded in a nested display of other data.
Example:: integer = Word(nums) date_str = integer("year") + '/' + integer("month") + '/' + integer("day")
result = date_str.parseString('12/31/1999') print(result.dump()) prints:: ['12', '/', '31', '/', '1999'] - day: 1999 - month: 31 - year: 12 """ out = [] NL = '\n' out.append( indent+_ustr(self.asList()) ) if full: if self.haskeys(): items = sorted((str(k), v) for k,v in self.items()) for k,v in items: if out: out.append(NL) out.append( "%s%s- %s: " % (indent,(' '*depth), k) ) if isinstance(v,ParseResults): if v: out.append( v.dump(indent,depth+1) ) else: out.append(_ustr(v)) else: out.append(repr(v)) elif any(isinstance(vv,ParseResults) for vv in self): v = self for i,vv in enumerate(v): if isinstance(vv,ParseResults): out.append("\n%s%s[%d]:\n%s%s%s" % (indent,(' '*(depth)),i,indent,(' '*(depth+1)),vv.dump(indent,depth+1) )) else: out.append("\n%s%s[%d]:\n%s%s%s" % (indent,(' '*(depth)),i,indent,(' '*(depth+1)),_ustr(vv)))
return "".join(out)
""" Pretty-printer for parsed results as a list, using the C{pprint} module. Accepts additional positional or keyword args as defined for the C{pprint.pprint} method. (U{http://docs.python.org/3/library/pprint.html#pprint.pprint})
Example:: ident = Word(alphas, alphanums) num = Word(nums) func = Forward() term = ident | num | Group('(' + func + ')') func <<= ident + Group(Optional(delimitedList(term))) result = func.parseString("fna a,b,(fnb c,d,200),100") result.pprint(width=40) prints:: ['fna', ['a', 'b', ['(', 'fnb', ['c', 'd', '200'], ')'], '100']] """ pprint.pprint(self.asList(), *args, **kwargs)
# add support for pickle protocol return ( self.__toklist, ( self.__tokdict.copy(), self.__parent is not None and self.__parent() or None, self.__accumNames, self.__name ) )
self.__toklist = state[0] (self.__tokdict, par, inAccumNames, self.__name) = state[1] self.__accumNames = {} self.__accumNames.update(inAccumNames) if par is not None: self.__parent = wkref(par) else: self.__parent = None
return self.__toklist, self.__name, self.__asList, self.__modal
return (dir(type(self)) + list(self.keys()))
"""Returns current column within a string, counting newlines as line separators. The first column is number 1.
Note: the default parsing behavior is to expand tabs in the input string before starting the parsing process. See L{I{ParserElement.parseString}<ParserElement.parseString>} for more information on parsing strings containing C{<TAB>}s, and suggested methods to maintain a consistent view of the parsed string, the parse location, and line and column positions within the parsed string. """ s = strg return 1 if 0<loc<len(s) and s[loc-1] == '\n' else loc - s.rfind("\n", 0, loc)
"""Returns current line number within a string, counting newlines as line separators. The first line is number 1.
Note: the default parsing behavior is to expand tabs in the input string before starting the parsing process. See L{I{ParserElement.parseString}<ParserElement.parseString>} for more information on parsing strings containing C{<TAB>}s, and suggested methods to maintain a consistent view of the parsed string, the parse location, and line and column positions within the parsed string. """ return strg.count("\n",0,loc) + 1
"""Returns the line of text containing loc within a string, counting newlines as line separators. """ lastCR = strg.rfind("\n", 0, loc) nextCR = strg.find("\n", loc) if nextCR >= 0: return strg[lastCR+1:nextCR] else: return strg[lastCR+1:]
print (("Match " + _ustr(expr) + " at loc " + _ustr(loc) + "(%d,%d)" % ( lineno(loc,instring), col(loc,instring) )))
print ("Matched " + _ustr(expr) + " -> " + str(toks.asList()))
print ("Exception raised:" + _ustr(exc))
"""'Do-nothing' debug action, to suppress debugging output during parsing.""" pass
# Only works on Python 3.x - nonlocal is toxic to Python 2 installs #~ 'decorator to trim function calls to match the arity of the target' #~ def _trim_arity(func, maxargs=3): #~ if func in singleArgBuiltins: #~ return lambda s,l,t: func(t) #~ limit = 0 #~ foundArity = False #~ def wrapper(*args): #~ nonlocal limit,foundArity #~ while 1: #~ try: #~ ret = func(*args[limit:]) #~ foundArity = True #~ return ret #~ except TypeError: #~ if limit == maxargs or foundArity: #~ raise #~ limit += 1 #~ continue #~ return wrapper
# this version is Python 2.x-3.x cross-compatible 'decorator to trim function calls to match the arity of the target'
# traceback return data structure changed in Py3.5 - normalize back to plain tuples # special handling for Python 3.5.0 - extra deep call stack by 1 frames = traceback.extract_tb(tb, limit=limit) frame_summary = frames[-1] return [frame_summary[:2]] else: extract_stack = traceback.extract_stack extract_tb = traceback.extract_tb
# synthesize what would be returned by traceback.extract_stack at the call to # user's parse action 'func', so that we don't incur call penalty at parse time
# IF ANY CODE CHANGES, EVEN JUST COMMENTS OR BLANK LINES, BETWEEN THE NEXT LINE AND # THE CALL TO FUNC INSIDE WRAPPER, LINE_DIFF MUST BE MODIFIED!!!!
except TypeError: # re-raise TypeErrors if they did not come from our arity testing if foundArity[0]: raise else: try: tb = sys.exc_info()[-1] if not extract_tb(tb, limit=2)[-1][:2] == pa_call_line_synth: raise finally: del tb
if limit[0] <= maxargs: limit[0] += 1 continue raise
# copy func name to wrapper for sensible debug output getattr(func, '__class__').__name__) except Exception: func_name = str(func)
"""Abstract base level parser element class."""
def setDefaultWhitespaceChars( chars ): r""" Overrides the default whitespace chars
Example:: # default whitespace chars are space, <TAB> and newline OneOrMore(Word(alphas)).parseString("abc def\nghi jkl") # -> ['abc', 'def', 'ghi', 'jkl']
# change to just treat newline as significant ParserElement.setDefaultWhitespaceChars(" \t") OneOrMore(Word(alphas)).parseString("abc def\nghi jkl") # -> ['abc', 'def'] """ ParserElement.DEFAULT_WHITE_CHARS = chars
def inlineLiteralsUsing(cls): """ Set class to be used for inclusion of string literals into a parser.
Example:: # default literal class used is Literal integer = Word(nums) date_str = integer("year") + '/' + integer("month") + '/' + integer("day")
date_str.parseString("1999/12/31") # -> ['1999', '/', '12', '/', '31']
# change to Suppress ParserElement.inlineLiteralsUsing(Suppress) date_str = integer("year") + '/' + integer("month") + '/' + integer("day")
date_str.parseString("1999/12/31") # -> ['1999', '12', '31'] """ ParserElement._literalStringClass = cls
#~ self.name = "<unknown>" # don't define self.name, let subclasses try/except upcall
""" Make a copy of this C{ParserElement}. Useful for defining different parse actions for the same parsing pattern, using copies of the original parse element.
Example:: integer = Word(nums).setParseAction(lambda toks: int(toks[0])) integerK = integer.copy().addParseAction(lambda toks: toks[0]*1024) + Suppress("K") integerM = integer.copy().addParseAction(lambda toks: toks[0]*1024*1024) + Suppress("M")
print(OneOrMore(integerK | integerM | integer).parseString("5K 100 640K 256M")) prints:: [5120, 100, 655360, 268435456] Equivalent form of C{expr.copy()} is just C{expr()}:: integerM = integer().addParseAction(lambda toks: toks[0]*1024*1024) + Suppress("M") """
""" Define name for this expression, makes debugging and exception messages clearer.
Example:: Word(nums).parseString("ABC") # -> Exception: Expected W:(0123...) (at char 0), (line:1, col:1) Word(nums).setName("integer").parseString("ABC") # -> Exception: Expected integer (at char 0), (line:1, col:1) """ self.exception.msg = self.errmsg
""" Define name for referencing matching tokens as a nested attribute of the returned parse results. NOTE: this returns a *copy* of the original C{ParserElement} object; this is so that the client can define a basic element, such as an integer, and reference it in multiple places with different names.
You can also set results names using the abbreviated syntax, C{expr("name")} in place of C{expr.setResultsName("name")} - see L{I{__call__}<__call__>}.
Example:: date_str = (integer.setResultsName("year") + '/' + integer.setResultsName("month") + '/' + integer.setResultsName("day"))
# equivalent form: date_str = integer("year") + '/' + integer("month") + '/' + integer("day") """ name = name[:-1] listAllMatches=True
"""Method to invoke the Python pdb debugger when this element is about to be parsed. Set C{breakFlag} to True to enable, False to disable. """ if breakFlag: _parseMethod = self._parse def breaker(instring, loc, doActions=True, callPreParse=True): import pdb pdb.set_trace() return _parseMethod( instring, loc, doActions, callPreParse ) breaker._originalParseMethod = _parseMethod self._parse = breaker else: if hasattr(self._parse,"_originalParseMethod"): self._parse = self._parse._originalParseMethod return self
""" Define one or more actions to perform when successfully matching parse element definition. Parse action fn is a callable method with 0-3 arguments, called as C{fn(s,loc,toks)}, C{fn(loc,toks)}, C{fn(toks)}, or just C{fn()}, where: - s = the original string being parsed (see note below) - loc = the location of the matching substring - toks = a list of the matched tokens, packaged as a C{L{ParseResults}} object If the functions in fns modify the tokens, they can return them as the return value from fn, and the modified list of tokens will replace the original. Otherwise, fn does not need to return any value.
Optional keyword arguments: - callDuringTry = (default=C{False}) indicate if parse action should be run during lookaheads and alternate testing
Note: the default parsing behavior is to expand tabs in the input string before starting the parsing process. See L{I{parseString}<parseString>} for more information on parsing strings containing C{<TAB>}s, and suggested methods to maintain a consistent view of the parsed string, the parse location, and line and column positions within the parsed string.
Example:: integer = Word(nums) date_str = integer + '/' + integer + '/' + integer
date_str.parseString("1999/12/31") # -> ['1999', '/', '12', '/', '31']
# use parse action to convert to ints at parse time integer = Word(nums).setParseAction(lambda toks: int(toks[0])) date_str = integer + '/' + integer + '/' + integer
# note that integer fields are now ints, not strings date_str.parseString("1999/12/31") # -> [1999, '/', 12, '/', 31] """
""" Add one or more parse actions to expression's list of parse actions. See L{I{setParseAction}<setParseAction>}.
See examples in L{I{copy}<copy>}. """
"""Add a boolean predicate function to expression's list of parse actions. See L{I{setParseAction}<setParseAction>} for function call signatures. Unlike C{setParseAction}, functions passed to C{addCondition} need to return boolean success/fail of the condition.
Optional keyword arguments: - message = define a custom message to be used in the raised exception - fatal = if True, will raise ParseFatalException to stop parsing immediately; otherwise will raise ParseException
Example:: integer = Word(nums).setParseAction(lambda toks: int(toks[0])) year_int = integer.copy() year_int.addCondition(lambda toks: toks[0] >= 2000, message="Only support years 2000 and later") date_str = year_int + '/' + integer + '/' + integer
result = date_str.parseString("1999/12/31") # -> Exception: Only support years 2000 and later (at char 0), (line:1, col:1) """ if not bool(_trim_arity(fn)(s,l,t)): raise exc_type(s,l,msg)
"""Define action to perform if parsing fails at this expression. Fail acton fn is a callable function that takes the arguments C{fn(s,loc,expr,err)} where: - s = string being parsed - loc = location where expression match was attempted and failed - expr = the parse expression that failed - err = the exception thrown The function returns no value. It may throw C{L{ParseFatalException}} if it is desired to stop parsing immediately.""" self.failAction = fn return self
exprsFound = True while exprsFound: exprsFound = False for e in self.ignoreExprs: try: while 1: loc,dummy = e._parse( instring, loc ) exprsFound = True except ParseException: pass return loc
loc = self._skipIgnorables( instring, loc )
loc += 1
#~ @profile
#~ print ("Match",self,"at loc",loc,"(%d,%d)" % ( lineno(loc,instring), col(loc,instring) )) if (self.debugActions[0] ): self.debugActions[0]( instring, loc, self ) if callPreParse and self.callPreparse: preloc = self.preParse( instring, loc ) else: preloc = loc tokensStart = preloc try: try: loc,tokens = self.parseImpl( instring, preloc, doActions ) except IndexError: raise ParseException( instring, len(instring), self.errmsg, self ) except ParseBaseException as err: #~ print ("Exception raised:", err) if self.debugActions[2]: self.debugActions[2]( instring, tokensStart, self, err ) if self.failAction: self.failAction( instring, tokensStart, self, err ) raise else: else: raise ParseException( instring, len(instring), self.errmsg, self ) else:
try: for fn in self.parseAction: tokens = fn( instring, tokensStart, retTokens ) if tokens is not None: retTokens = ParseResults( tokens, self.resultsName, asList=self.saveAsList and isinstance(tokens,(ParseResults,list)), modal=self.modalResults ) except ParseBaseException as err: #~ print "Exception raised in user parse action:", err if (self.debugActions[2] ): self.debugActions[2]( instring, tokensStart, self, err ) raise else: self.resultsName, asList=self.saveAsList and isinstance(tokens,(ParseResults,list)), modal=self.modalResults ) #~ print ("Matched",self,"->",retTokens.asList()) if (self.debugActions[1] ): self.debugActions[1]( instring, tokensStart, loc, self, retTokens )
raise ParseException( instring, loc, self.errmsg, self)
try: self.tryParse(instring, loc) except (ParseException, IndexError): return False else: return True
cache = {} self.not_in_cache = not_in_cache = object()
def get(self, key): return cache.get(key, not_in_cache)
def set(self, key, value): cache[key] = value
def clear(self): cache.clear()
def cache_len(self): return len(cache)
self.get = types.MethodType(get, self) self.set = types.MethodType(set, self) self.clear = types.MethodType(clear, self) self.__len__ = types.MethodType(cache_len, self)
self.not_in_cache = not_in_cache = object()
cache = _OrderedDict()
def get(self, key): return cache.get(key, not_in_cache)
def set(self, key, value): cache[key] = value while len(cache) > size: try: cache.popitem(False) except KeyError: pass
def clear(self): cache.clear()
def cache_len(self): return len(cache)
self.get = types.MethodType(get, self) self.set = types.MethodType(set, self) self.clear = types.MethodType(clear, self) self.__len__ = types.MethodType(cache_len, self)
else: class _FifoCache(object): def __init__(self, size): self.not_in_cache = not_in_cache = object()
cache = {} key_fifo = collections.deque([], size)
def get(self, key): return cache.get(key, not_in_cache)
def set(self, key, value): cache[key] = value while len(key_fifo) > size: cache.pop(key_fifo.popleft(), None) key_fifo.append(key)
def clear(self): cache.clear() key_fifo.clear()
def cache_len(self): return len(cache)
self.get = types.MethodType(get, self) self.set = types.MethodType(set, self) self.clear = types.MethodType(clear, self) self.__len__ = types.MethodType(cache_len, self)
# argument cache for optimizing repeated calls when backtracking through recursive expressions
# this method gets repeatedly called during backtracking with the same arguments - # we can cache these arguments and save ourselves the trouble of re-parsing the contained expression HIT, MISS = 0, 1 lookup = (self, instring, loc, callPreParse, doActions) with ParserElement.packrat_cache_lock: cache = ParserElement.packrat_cache value = cache.get(lookup) if value is cache.not_in_cache: ParserElement.packrat_cache_stats[MISS] += 1 try: value = self._parseNoCache(instring, loc, doActions, callPreParse) except ParseBaseException as pe: # cache a copy of the exception, without the traceback cache.set(lookup, pe.__class__(*pe.args)) raise else: cache.set(lookup, (value[0], value[1].copy())) return value else: ParserElement.packrat_cache_stats[HIT] += 1 if isinstance(value, Exception): raise value return (value[0], value[1].copy())
def resetCache():
"""Enables "packrat" parsing, which adds memoizing to the parsing logic. Repeated parse attempts at the same string location (which happens often in many complex grammars) can immediately return a cached value, instead of re-executing parsing/validating code. Memoizing is done of both valid results and parsing exceptions.
Parameters: - cache_size_limit - (default=C{128}) - if an integer value is provided will limit the size of the packrat cache; if None is passed, then the cache size will be unbounded; if 0 is passed, the cache will be effectively disabled.
This speedup may break existing programs that use parse actions that have side-effects. For this reason, packrat parsing is disabled when you first import pyparsing. To activate the packrat feature, your program must call the class method C{ParserElement.enablePackrat()}. If your program uses C{psyco} to "compile as you go", you must call C{enablePackrat} before calling C{psyco.full()}. If you do not do this, Python will crash. For best results, call C{enablePackrat()} immediately after importing pyparsing.
Example:: import pyparsing pyparsing.ParserElement.enablePackrat() """ if not ParserElement._packratEnabled: ParserElement._packratEnabled = True if cache_size_limit is None: ParserElement.packrat_cache = ParserElement._UnboundedCache() else: ParserElement.packrat_cache = ParserElement._FifoCache(cache_size_limit) ParserElement._parse = ParserElement._parseCache
""" Execute the parse expression with the given string. This is the main interface to the client code, once the complete expression has been built.
If you want the grammar to require that the entire input string be successfully parsed, then set C{parseAll} to True (equivalent to ending the grammar with C{L{StringEnd()}}).
Note: C{parseString} implicitly calls C{expandtabs()} on the input string, in order to report proper column numbers in parse actions. If the input string contains tabs and the grammar uses parse actions that use the C{loc} argument to index into the string being parsed, you can ensure you have a consistent view of the input string by: - calling C{parseWithTabs} on your grammar before calling C{parseString} (see L{I{parseWithTabs}<parseWithTabs>}) - define your parse action using the full C{(s,loc,toks)} signature, and reference the input string using the parse action's C{s} argument - explictly expand the tabs in your input string before calling C{parseString}
Example:: Word('a').parseString('aaaaabaaa') # -> ['aaaaa'] Word('a').parseString('aaaaabaaa', parseAll=True) # -> Exception: Expected end of text """ #~ self.saveAsList = True e.streamline() loc = self.preParse( instring, loc ) se = Empty() + StringEnd() se._parse( instring, loc ) except ParseBaseException as exc: if ParserElement.verbose_stacktrace: raise else: # catch and re-raise exception from here, clears out pyparsing internal stack trace raise exc else:
""" Scan the input string for expression matches. Each match will return the matching tokens, start location, and end location. May be called with optional C{maxMatches} argument, to clip scanning after 'n' matches are found. If C{overlap} is specified, then overlapping matches will be reported.
Note that the start and end locations are reported relative to the string being parsed. See L{I{parseString}<parseString>} for more information on parsing strings with embedded tabs.
Example:: source = "sldjf123lsdjjkf345sldkjf879lkjsfd987" print(source) for tokens,start,end in Word(alphas).scanString(source): print(' '*start + '^'*(end-start)) print(' '*start + tokens[0])
prints::
sldjf123lsdjjkf345sldkjf879lkjsfd987 ^^^^^ sldjf ^^^^^^^ lsdjjkf ^^^^^^ sldkjf ^^^^^^ lkjsfd """ if not self.streamlined: self.streamline() for e in self.ignoreExprs: e.streamline()
if not self.keepTabs: instring = _ustr(instring).expandtabs() instrlen = len(instring) loc = 0 preparseFn = self.preParse parseFn = self._parse ParserElement.resetCache() matches = 0 try: while loc <= instrlen and matches < maxMatches: try: preloc = preparseFn( instring, loc ) nextLoc,tokens = parseFn( instring, preloc, callPreParse=False ) except ParseException: loc = preloc+1 else: if nextLoc > loc: matches += 1 yield tokens, preloc, nextLoc if overlap: nextloc = preparseFn( instring, loc ) if nextloc > loc: loc = nextLoc else: loc += 1 else: loc = nextLoc else: loc = preloc+1 except ParseBaseException as exc: if ParserElement.verbose_stacktrace: raise else: # catch and re-raise exception from here, clears out pyparsing internal stack trace raise exc
""" Extension to C{L{scanString}}, to modify matching text with modified tokens that may be returned from a parse action. To use C{transformString}, define a grammar and attach a parse action to it that modifies the returned token list. Invoking C{transformString()} on a target string will then scan for matches, and replace the matched text patterns according to the logic in the parse action. C{transformString()} returns the resulting transformed string.
Example:: wd = Word(alphas) wd.setParseAction(lambda toks: toks[0].title())
print(wd.transformString("now is the winter of our discontent made glorious summer by this sun of york.")) Prints:: Now Is The Winter Of Our Discontent Made Glorious Summer By This Sun Of York. """ out = [] lastE = 0 # force preservation of <TAB>s, to minimize unwanted transformation of string, and to # keep string locs straight between transformString and scanString self.keepTabs = True try: for t,s,e in self.scanString( instring ): out.append( instring[lastE:s] ) if t: if isinstance(t,ParseResults): out += t.asList() elif isinstance(t,list): out += t else: out.append(t) lastE = e out.append(instring[lastE:]) out = [o for o in out if o] return "".join(map(_ustr,_flatten(out))) except ParseBaseException as exc: if ParserElement.verbose_stacktrace: raise else: # catch and re-raise exception from here, clears out pyparsing internal stack trace raise exc
""" Another extension to C{L{scanString}}, simplifying the access to the tokens found to match the given parse expression. May be called with optional C{maxMatches} argument, to clip searching after 'n' matches are found.
Example:: # a capitalized word starts with an uppercase letter, followed by zero or more lowercase letters cap_word = Word(alphas.upper(), alphas.lower())
print(cap_word.searchString("More than Iron, more than Lead, more than Gold I need Electricity"))
# the sum() builtin can be used to merge results into a single ParseResults object print(sum(cap_word.searchString("More than Iron, more than Lead, more than Gold I need Electricity"))) prints:: [['More'], ['Iron'], ['Lead'], ['Gold'], ['I'], ['Electricity']] ['More', 'Iron', 'Lead', 'Gold', 'I', 'Electricity'] """ try: return ParseResults([ t for t,s,e in self.scanString( instring, maxMatches ) ]) except ParseBaseException as exc: if ParserElement.verbose_stacktrace: raise else: # catch and re-raise exception from here, clears out pyparsing internal stack trace raise exc
""" Generator method to split a string using the given expression as a separator. May be called with optional C{maxsplit} argument, to limit the number of splits; and the optional C{includeSeparators} argument (default=C{False}), if the separating matching text should be included in the split results.
Example:: punc = oneOf(list(".,;:/-!?")) print(list(punc.split("This, this?, this sentence, is badly punctuated!"))) prints:: ['This', ' this', '', ' this sentence', ' is badly punctuated', ''] """ splits = 0 last = 0 for t,s,e in self.scanString(instring, maxMatches=maxsplit): yield instring[last:s] if includeSeparators: yield t[0] last = e yield instring[last:]
""" Implementation of + operator - returns C{L{And}}. Adding strings to a ParserElement converts them to L{Literal}s by default.
Example:: greet = Word(alphas) + "," + Word(alphas) + "!" hello = "Hello, World!" print (hello, "->", greet.parseString(hello)) Prints:: Hello, World! -> ['Hello', ',', 'World', '!'] """ warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None
""" Implementation of + operator when left operand is not a C{L{ParserElement}} """ warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None
""" Implementation of - operator, returns C{L{And}} with error stop """ if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) if not isinstance( other, ParserElement ): warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None return self + And._ErrorStop() + other
""" Implementation of - operator when left operand is not a C{L{ParserElement}} """ if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) if not isinstance( other, ParserElement ): warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None return other - self
""" Implementation of * operator, allows use of C{expr * 3} in place of C{expr + expr + expr}. Expressions may also me multiplied by a 2-integer tuple, similar to C{{min,max}} multipliers in regular expressions. Tuples may also include C{None} as in: - C{expr*(n,None)} or C{expr*(n,)} is equivalent to C{expr*n + L{ZeroOrMore}(expr)} (read as "at least n instances of C{expr}") - C{expr*(None,n)} is equivalent to C{expr*(0,n)} (read as "0 to n instances of C{expr}") - C{expr*(None,None)} is equivalent to C{L{ZeroOrMore}(expr)} - C{expr*(1,None)} is equivalent to C{L{OneOrMore}(expr)}
Note that C{expr*(None,n)} does not raise an exception if more than n exprs exist in the input stream; that is, C{expr*(None,n)} does not enforce a maximum number of expr occurrences. If this behavior is desired, then write C{expr*(None,n) + ~expr} """ other = (0, other[1]) if other[0] == 0: return ZeroOrMore(self) if other[0] == 1: return OneOrMore(self) else: return self*other[0] + ZeroOrMore(self) else: raise TypeError("cannot multiply 'ParserElement' and ('%s','%s') objects", type(other[0]),type(other[1])) else: raise TypeError("cannot multiply 'ParserElement' and '%s' objects", type(other))
raise ValueError("cannot multiply ParserElement by negative value") raise ValueError("second tuple value must be greater or equal to first tuple value") raise ValueError("cannot multiply ParserElement by 0 or (0,0)")
else: if minElements == 1: ret = self + makeOptionalList(optElements) else: ret = And([self]*minElements) + makeOptionalList(optElements) else: else: ret = self else:
return self.__mul__(other)
""" Implementation of | operator - returns C{L{MatchFirst}} """ other = ParserElement._literalStringClass( other ) warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None
""" Implementation of | operator when left operand is not a C{L{ParserElement}} """ if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) if not isinstance( other, ParserElement ): warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None return other | self
""" Implementation of ^ operator - returns C{L{Or}} """ other = ParserElement._literalStringClass( other ) warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None
""" Implementation of ^ operator when left operand is not a C{L{ParserElement}} """ if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) if not isinstance( other, ParserElement ): warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None return other ^ self
""" Implementation of & operator - returns C{L{Each}} """ if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) if not isinstance( other, ParserElement ): warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None return Each( [ self, other ] )
""" Implementation of & operator when left operand is not a C{L{ParserElement}} """ if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) if not isinstance( other, ParserElement ): warnings.warn("Cannot combine element of type %s with ParserElement" % type(other), SyntaxWarning, stacklevel=2) return None return other & self
""" Implementation of ~ operator - returns C{L{NotAny}} """
""" Shortcut for C{L{setResultsName}}, with C{listAllMatches=False}.
If C{name} is given with a trailing C{'*'} character, then C{listAllMatches} will be passed as C{True}.
If C{name} is omitted, same as calling C{L{copy}}.
Example:: # these are equivalent userdata = Word(alphas).setResultsName("name") + Word(nums+"-").setResultsName("socsecno") userdata = Word(alphas)("name") + Word(nums+"-")("socsecno") """ else:
""" Suppresses the output of this C{ParserElement}; useful to keep punctuation from cluttering up returned output. """
""" Disables the skipping of whitespace before matching the characters in the C{ParserElement}'s defined pattern. This is normally only used internally by the pyparsing module, but may be needed in some whitespace-sensitive grammars. """
""" Overrides the default whitespace chars """
""" Overrides default behavior to expand C{<TAB>}s to spaces before parsing the input string. Must be called before C{parseString} when the input grammar contains elements that match C{<TAB>} characters. """ self.keepTabs = True return self
""" Define expression to be ignored (e.g., comments) while doing pattern matching; may be called repeatedly, to define multiple comment or other ignorable patterns.
Example:: patt = OneOrMore(Word(alphas)) patt.parseString('ablaj /* comment */ lskjd') # -> ['ablaj']
patt.ignore(cStyleComment) patt.parseString('ablaj /* comment */ lskjd') # -> ['ablaj', 'lskjd'] """ if isinstance(other, basestring): other = Suppress(other)
if isinstance( other, Suppress ): if other not in self.ignoreExprs: self.ignoreExprs.append(other) else: self.ignoreExprs.append( Suppress( other.copy() ) ) return self
""" Enable display of debugging messages while doing pattern matching. """ self.debugActions = (startAction or _defaultStartDebugAction, successAction or _defaultSuccessDebugAction, exceptionAction or _defaultExceptionDebugAction) self.debug = True return self
""" Enable display of debugging messages while doing pattern matching. Set C{flag} to True to enable, False to disable.
Example:: wd = Word(alphas).setName("alphaword") integer = Word(nums).setName("numword") term = wd | integer
# turn on debugging for wd wd.setDebug()
OneOrMore(term).parseString("abc 123 xyz 890")
prints:: Match alphaword at loc 0(1,1) Matched alphaword -> ['abc'] Match alphaword at loc 3(1,4) Exception raised:Expected alphaword (at char 4), (line:1, col:5) Match alphaword at loc 7(1,8) Matched alphaword -> ['xyz'] Match alphaword at loc 11(1,12) Exception raised:Expected alphaword (at char 12), (line:1, col:13) Match alphaword at loc 15(1,16) Exception raised:Expected alphaword (at char 15), (line:1, col:16)
The output shown is that produced by the default debug actions - custom debug actions can be specified using L{setDebugActions}. Prior to attempting to match the C{wd} expression, the debugging message C{"Match <exprname> at loc <n>(<line>,<col>)"} is shown. Then if the parse succeeds, a C{"Matched"} message is shown, or an C{"Exception raised"} message is shown. Also note the use of L{setName} to assign a human-readable name to the expression, which makes debugging and exception messages easier to understand - for instance, the default name created for the C{Word} expression without calling C{setName} is C{"W:(ABCD...)"}. """ if flag: self.setDebugActions( _defaultStartDebugAction, _defaultSuccessDebugAction, _defaultExceptionDebugAction ) else: self.debug = False return self
def __str__( self ): return self.name
def __repr__( self ): return _ustr(self)
pass
""" Check defined expressions for valid structure, check for infinite recursive definitions. """ self.checkRecursion( [] )
""" Execute the parse expression on the given file or filename. If a filename is specified (instead of a file object), the entire file is opened, read, and closed before parsing. """ try: file_contents = file_or_filename.read() except AttributeError: with open(file_or_filename, "r") as f: file_contents = f.read() try: return self.parseString(file_contents, parseAll) except ParseBaseException as exc: if ParserElement.verbose_stacktrace: raise else: # catch and re-raise exception from here, clears out pyparsing internal stack trace raise exc
if isinstance(other, ParserElement): return self is other or vars(self) == vars(other) elif isinstance(other, basestring): return self.matches(other) else: return super(ParserElement,self)==other
return not (self == other)
return hash(id(self))
return self == other
return not (self == other)
""" Method for quick testing of a parser against a test string. Good for simple inline microtests of sub expressions while building up larger parser.
Parameters: - testString - to test against this expression for a match - parseAll - (default=C{True}) - flag to pass to C{L{parseString}} when running tests
Example:: expr = Word(nums) assert expr.matches("100") """ try: self.parseString(_ustr(testString), parseAll=parseAll) return True except ParseBaseException: return False
""" Execute the parse expression on a series of test strings, showing each test, the parsed results or where the parse failed. Quick and easy way to run a parse expression against a list of sample strings.
Parameters: - tests - a list of separate test strings, or a multiline string of test strings - parseAll - (default=C{True}) - flag to pass to C{L{parseString}} when running tests - comment - (default=C{'#'}) - expression for indicating embedded comments in the test string; pass None to disable comment filtering - fullDump - (default=C{True}) - dump results as list followed by results names in nested outline; if False, only dump nested list - printResults - (default=C{True}) prints test output to stdout - failureTests - (default=C{False}) indicates if these tests are expected to fail parsing
Returns: a (success, results) tuple, where success indicates that all tests succeeded (or failed if C{failureTests} is True), and the results contain a list of lines of each test's output
Example:: number_expr = pyparsing_common.number.copy()
result = number_expr.runTests(''' # unsigned integer 100 # negative integer -100 # float with scientific notation 6.02e23 # integer with scientific notation 1e-12 ''') print("Success" if result[0] else "Failed!")
result = number_expr.runTests(''' # stray character 100Z # missing leading digit before '.' -.100 # too many '.' 3.14.159 ''', failureTests=True) print("Success" if result[0] else "Failed!") prints:: # unsigned integer 100 [100]
# negative integer -100 [-100]
# float with scientific notation 6.02e23 [6.02e+23]
# integer with scientific notation 1e-12 [1e-12]
Success
# stray character 100Z ^ FAIL: Expected end of text (at char 3), (line:1, col:4)
# missing leading digit before '.' -.100 ^ FAIL: Expected {real number with scientific notation | real number | signed integer} (at char 0), (line:1, col:1)
# too many '.' 3.14.159 ^ FAIL: Expected end of text (at char 4), (line:1, col:5)
Success
Each test string must be on a single line. If you want to test a string that spans multiple lines, create a test like this::
expr.runTest(r"this is a test\\n of strings that spans \\n 3 lines")
(Note that this is a raw string literal, you must include the leading 'r'.) """ if isinstance(tests, basestring): tests = list(map(str.strip, tests.rstrip().splitlines())) if isinstance(comment, basestring): comment = Literal(comment) allResults = [] comments = [] success = True for t in tests: if comment is not None and comment.matches(t, False) or comments and not t: comments.append(t) continue if not t: continue out = ['\n'.join(comments), t] comments = [] try: t = t.replace(r'\n','\n') result = self.parseString(t, parseAll=parseAll) out.append(result.dump(full=fullDump)) success = success and not failureTests except ParseBaseException as pe: fatal = "(FATAL)" if isinstance(pe, ParseFatalException) else "" if '\n' in t: out.append(line(pe.loc, t)) out.append(' '*(col(pe.loc,t)-1) + '^' + fatal) else: out.append(' '*pe.loc + '^' + fatal) out.append("FAIL: " + str(pe)) success = success and failureTests result = pe except Exception as exc: out.append("FAIL-EXCEPTION: " + str(exc)) success = success and failureTests result = exc
if printResults: if fullDump: out.append('') print('\n'.join(out))
allResults.append((t, result))
return success, allResults
""" Abstract C{ParserElement} subclass, for defining atomic matching patterns. """
""" An empty token, will always match. """
""" A token that will never match. """ super(NoMatch,self).__init__() self.name = "NoMatch" self.mayReturnEmpty = True self.mayIndexError = False self.errmsg = "Unmatchable token"
raise ParseException(instring, loc, self.errmsg, self)
""" Token to exactly match a specified string.
Example:: Literal('blah').parseString('blah') # -> ['blah'] Literal('blah').parseString('blahfooblah') # -> ['blah'] Literal('blah').parseString('bla') # -> Exception: Expected "blah"
For case-insensitive matching, use L{CaselessLiteral}.
For keyword matching (force word break before and after the matched string), use L{Keyword} or L{CaselessKeyword}. """ except IndexError: warnings.warn("null string passed to Literal; use Empty() instead", SyntaxWarning, stacklevel=2) self.__class__ = Empty
# Performance tuning: this routine gets called a *lot* # if this is a single character match string and the first character matches, # short-circuit as quickly as possible, and avoid calling startswith #~ @profile (self.matchLen==1 or instring.startswith(self.match,loc)) ): raise ParseException(instring, loc, self.errmsg, self)
""" Token to exactly match a specified string as a keyword, that is, it must be immediately followed by a non-keyword character. Compare with C{L{Literal}}: - C{Literal("if")} will match the leading C{'if'} in C{'ifAndOnlyIf'}. - C{Keyword("if")} will not; it will only match the leading C{'if'} in C{'if x=1'}, or C{'if(y==2)'} Accepts two optional constructor arguments in addition to the keyword string: - C{identChars} is a string of characters that would be valid identifier characters, defaulting to all alphanumerics + "_" and "$" - C{caseless} allows case-insensitive matching, default is C{False}.
Example:: Keyword("start").parseString("start") # -> ['start'] Keyword("start").parseString("starting") # -> Exception
For case-insensitive matching, use L{CaselessKeyword}. """
super(Keyword,self).__init__() if identChars is None: identChars = Keyword.DEFAULT_KEYWORD_CHARS self.match = matchString self.matchLen = len(matchString) try: self.firstMatchChar = matchString[0] except IndexError: warnings.warn("null string passed to Keyword; use Empty() instead", SyntaxWarning, stacklevel=2) self.name = '"%s"' % self.match self.errmsg = "Expected " + self.name self.mayReturnEmpty = False self.mayIndexError = False self.caseless = caseless if caseless: self.caselessmatch = matchString.upper() identChars = identChars.upper() self.identChars = set(identChars)
if self.caseless: if ( (instring[ loc:loc+self.matchLen ].upper() == self.caselessmatch) and (loc >= len(instring)-self.matchLen or instring[loc+self.matchLen].upper() not in self.identChars) and (loc == 0 or instring[loc-1].upper() not in self.identChars) ): return loc+self.matchLen, self.match else: if (instring[loc] == self.firstMatchChar and (self.matchLen==1 or instring.startswith(self.match,loc)) and (loc >= len(instring)-self.matchLen or instring[loc+self.matchLen] not in self.identChars) and (loc == 0 or instring[loc-1] not in self.identChars) ): return loc+self.matchLen, self.match raise ParseException(instring, loc, self.errmsg, self)
c = super(Keyword,self).copy() c.identChars = Keyword.DEFAULT_KEYWORD_CHARS return c
def setDefaultKeywordChars( chars ): """Overrides the default Keyword chars """ Keyword.DEFAULT_KEYWORD_CHARS = chars
""" Token to match a specified string, ignoring case of letters. Note: the matched results will always be in the case of the given match string, NOT the case of the input text.
Example:: OneOrMore(CaselessLiteral("CMD")).parseString("cmd CMD Cmd10") # -> ['CMD', 'CMD', 'CMD']
(Contrast with example for L{CaselessKeyword}.) """ super(CaselessLiteral,self).__init__( matchString.upper() ) # Preserve the defining literal. self.returnString = matchString self.name = "'%s'" % self.returnString self.errmsg = "Expected " + self.name
if instring[ loc:loc+self.matchLen ].upper() == self.match: return loc+self.matchLen, self.returnString raise ParseException(instring, loc, self.errmsg, self)
""" Caseless version of L{Keyword}.
Example:: OneOrMore(CaselessKeyword("CMD")).parseString("cmd CMD Cmd10") # -> ['CMD', 'CMD']
(Contrast with example for L{CaselessLiteral}.) """ super(CaselessKeyword,self).__init__( matchString, identChars, caseless=True )
if ( (instring[ loc:loc+self.matchLen ].upper() == self.caselessmatch) and (loc >= len(instring)-self.matchLen or instring[loc+self.matchLen].upper() not in self.identChars) ): return loc+self.matchLen, self.match raise ParseException(instring, loc, self.errmsg, self)
""" A variation on L{Literal} which matches "close" matches, that is, strings with at most 'n' mismatching characters. C{CloseMatch} takes parameters: - C{match_string} - string to be matched - C{maxMismatches} - (C{default=1}) maximum number of mismatches allowed to count as a match
The results from a successful parse will contain the matched text from the input string and the following named results: - C{mismatches} - a list of the positions within the match_string where mismatches were found - C{original} - the original match_string used to compare against the input string
If C{mismatches} is an empty list, then the match was an exact match.
Example:: patt = CloseMatch("ATCATCGAATGGA") patt.parseString("ATCATCGAAXGGA") # -> (['ATCATCGAAXGGA'], {'mismatches': [[9]], 'original': ['ATCATCGAATGGA']}) patt.parseString("ATCAXCGAAXGGA") # -> Exception: Expected 'ATCATCGAATGGA' (with up to 1 mismatches) (at char 0), (line:1, col:1)
# exact match patt.parseString("ATCATCGAATGGA") # -> (['ATCATCGAATGGA'], {'mismatches': [[]], 'original': ['ATCATCGAATGGA']})
# close match allowing up to 2 mismatches patt = CloseMatch("ATCATCGAATGGA", maxMismatches=2) patt.parseString("ATCAXCGAAXGGA") # -> (['ATCAXCGAAXGGA'], {'mismatches': [[4, 9]], 'original': ['ATCATCGAATGGA']}) """ super(CloseMatch,self).__init__() self.name = match_string self.match_string = match_string self.maxMismatches = maxMismatches self.errmsg = "Expected %r (with up to %d mismatches)" % (self.match_string, self.maxMismatches) self.mayIndexError = False self.mayReturnEmpty = False
start = loc instrlen = len(instring) maxloc = start + len(self.match_string)
if maxloc <= instrlen: match_string = self.match_string match_stringloc = 0 mismatches = [] maxMismatches = self.maxMismatches
for match_stringloc,s_m in enumerate(zip(instring[loc:maxloc], self.match_string)): src,mat = s_m if src != mat: mismatches.append(match_stringloc) if len(mismatches) > maxMismatches: break else: loc = match_stringloc + 1 results = ParseResults([instring[start:loc]]) results['original'] = self.match_string results['mismatches'] = mismatches return loc, results
raise ParseException(instring, loc, self.errmsg, self)
""" Token for matching words composed of allowed character sets. Defined with string containing all allowed initial characters, an optional string containing allowed body characters (if omitted, defaults to the initial character set), and an optional minimum, maximum, and/or exact length. The default value for C{min} is 1 (a minimum value < 1 is not valid); the default values for C{max} and C{exact} are 0, meaning no maximum or exact length restriction. An optional C{excludeChars} parameter can list characters that might be found in the input C{bodyChars} string; useful to define a word of all printables except for one or two characters, for instance.
L{srange} is useful for defining custom character set strings for defining C{Word} expressions, using range notation from regular expression character sets.
A common mistake is to use C{Word} to match a specific literal string, as in C{Word("Address")}. Remember that C{Word} uses the string argument to define I{sets} of matchable characters. This expression would match "Add", "AAA", "dAred", or any other word made up of the characters 'A', 'd', 'r', 'e', and 's'. To match an exact literal string, use L{Literal} or L{Keyword}.
pyparsing includes helper strings for building Words: - L{alphas} - L{nums} - L{alphanums} - L{hexnums} - L{alphas8bit} (alphabetic characters in ASCII range 128-255 - accented, tilded, umlauted, etc.) - L{punc8bit} (non-alphabetic characters in ASCII range 128-255 - currency, symbols, superscripts, diacriticals, etc.) - L{printables} (any non-whitespace character)
Example:: # a word composed of digits integer = Word(nums) # equivalent to Word("0123456789") or Word(srange("0-9"))
# a word with a leading capital, and zero or more lowercase capital_word = Word(alphas.upper(), alphas.lower())
# hostnames are alphanumeric, with leading alpha, and '-' hostname = Word(alphas, alphanums+'-')
# roman numeral (not a strict parser, accepts invalid mix of characters) roman = Word("IVXLCDM")
# any string of non-whitespace characters, except for ',' csv_value = Word(printables, excludeChars=",") """ bodyChars = ''.join(c for c in bodyChars if c not in excludeChars) else:
raise ValueError("cannot specify a minimum length < 1; use Optional(Word()) if zero-length word is permitted")
self.maxLen = max else:
self.reString = "%s[%s]*" % \ (re.escape(self.initCharsOrig), _escapeRegexRangeChars(self.bodyCharsOrig),) else: (_escapeRegexRangeChars(self.initCharsOrig), _escapeRegexRangeChars(self.bodyCharsOrig),) self.reString = r"\b"+self.reString+r"\b" except Exception: self.re = None
raise ParseException(instring, loc, self.errmsg, self)
raise ParseException(instring, loc, self.errmsg, self)
loc += 1
throwException = True if (start>0 and instring[start-1] in bodychars) or (loc<instrlen and instring[loc] in bodychars): throwException = True
raise ParseException(instring, loc, self.errmsg, self)
return loc, instring[start:loc]
def __str__( self ): try: return super(Word,self).__str__() except Exception: pass
if self.strRepr is None:
def charsAsStr(s): if len(s)>4: return s[:4]+"..." else: return s
if ( self.initCharsOrig != self.bodyCharsOrig ): self.strRepr = "W:(%s,%s)" % ( charsAsStr(self.initCharsOrig), charsAsStr(self.bodyCharsOrig) ) else: self.strRepr = "W:(%s)" % charsAsStr(self.initCharsOrig)
return self.strRepr
r""" Token for matching strings that match a given regular expression. Defined with string specifying the regular expression in a form recognized by the inbuilt Python re module. If the given regex contains named groups (defined using C{(?P<name>...)}), these will be preserved as named parse results.
Example:: realnum = Regex(r"[+-]?\d+\.\d*") date = Regex(r'(?P<year>\d{4})-(?P<month>\d\d?)-(?P<day>\d\d?)') # ref: http://stackoverflow.com/questions/267399/how-do-you-match-only-valid-roman-numerals-with-a-regular-expression roman = Regex(r"M{0,4}(CM|CD|D?C{0,3})(XC|XL|L?X{0,3})(IX|IV|V?I{0,3})") """ """The parameters C{pattern} and C{flags} are passed to the C{re.compile()} function as-is. See the Python C{re} module for an explanation of the acceptable patterns and flags."""
warnings.warn("null string passed to Regex; use Empty() instead", SyntaxWarning, stacklevel=2)
except sre_constants.error: warnings.warn("invalid pattern (%s) passed to Regex" % pattern, SyntaxWarning, stacklevel=2) raise
elif isinstance(pattern, Regex.compiledREtype): self.re = pattern self.pattern = \ self.reString = str(pattern) self.flags = flags
else: raise ValueError("Regex may only be constructed with a string or a compiled RE object")
raise ParseException(instring, loc, self.errmsg, self)
for k in d: ret[k] = d[k]
def __str__( self ): try: return super(Regex,self).__str__() except Exception: pass
if self.strRepr is None: self.strRepr = "Re:(%s)" % repr(self.pattern)
return self.strRepr
r""" Token for matching strings that are delimited by quoting characters.
Defined with the following parameters: - quoteChar - string of one or more characters defining the quote delimiting string - escChar - character to escape quotes, typically backslash (default=C{None}) - escQuote - special quote sequence to escape an embedded quote string (such as SQL's "" to escape an embedded ") (default=C{None}) - multiline - boolean indicating whether quotes can span multiple lines (default=C{False}) - unquoteResults - boolean indicating whether the matched text should be unquoted (default=C{True}) - endQuoteChar - string of one or more characters defining the end of the quote delimited string (default=C{None} => same as quoteChar) - convertWhitespaceEscapes - convert escaped whitespace (C{'\t'}, C{'\n'}, etc.) to actual whitespace (default=C{True})
Example:: qs = QuotedString('"') print(qs.searchString('lsjdf "This is the quote" sldjf')) complex_qs = QuotedString('{{', endQuoteChar='}}') print(complex_qs.searchString('lsjdf {{This is the "quote"}} sldjf')) sql_qs = QuotedString('"', escQuote='""') print(sql_qs.searchString('lsjdf "This is the quote with ""embedded"" quotes" sldjf')) prints:: [['This is the quote']] [['This is the "quote"']] [['This is the quote with "embedded" quotes']] """
# remove white space from quote chars - wont work anyway warnings.warn("quoteChar cannot be the empty string",SyntaxWarning,stacklevel=2) raise SyntaxError()
else: endQuoteChar = endQuoteChar.strip() if not endQuoteChar: warnings.warn("endQuoteChar cannot be the empty string",SyntaxWarning,stacklevel=2) raise SyntaxError()
self.flags = re.MULTILINE | re.DOTALL self.pattern = r'%s(?:[^%s%s]' % \ ( re.escape(self.quoteChar), _escapeRegexRangeChars(self.endQuoteChar[0]), (escChar is not None and _escapeRegexRangeChars(escChar) or '') ) else: ( re.escape(self.quoteChar), _escapeRegexRangeChars(self.endQuoteChar[0]), (escChar is not None and _escapeRegexRangeChars(escChar) or '') ) self.pattern += ( '|(?:' + ')|(?:'.join("%s[^%s]" % (re.escape(self.endQuoteChar[:i]), _escapeRegexRangeChars(self.endQuoteChar[i])) for i in range(len(self.endQuoteChar)-1,0,-1)) + ')' ) self.pattern += (r'|(?:%s)' % re.escape(escQuote)) self.pattern += (r'|(?:%s.)' % re.escape(escChar)) self.escCharReplacePattern = re.escape(self.escChar)+"(.)"
except sre_constants.error: warnings.warn("invalid pattern (%s) passed to Regex" % self.pattern, SyntaxWarning, stacklevel=2) raise
result = instring[loc] == self.firstQuoteChar and self.re.match(instring,loc) or None if not result: raise ParseException(instring, loc, self.errmsg, self)
loc = result.end() ret = result.group()
if self.unquoteResults:
# strip off quotes ret = ret[self.quoteCharLen:-self.endQuoteCharLen]
if isinstance(ret,basestring): # replace escaped whitespace if '\\' in ret and self.convertWhitespaceEscapes: ws_map = { r'\t' : '\t', r'\n' : '\n', r'\f' : '\f', r'\r' : '\r', } for wslit,wschar in ws_map.items(): ret = ret.replace(wslit, wschar)
# replace escaped characters if self.escChar: ret = re.sub(self.escCharReplacePattern, r"\g<1>", ret)
# replace escaped quotes if self.escQuote: ret = ret.replace(self.escQuote, self.endQuoteChar)
return loc, ret
def __str__( self ): try: return super(QuotedString,self).__str__() except Exception: pass
if self.strRepr is None: self.strRepr = "quoted string, starting with %s ending with %s" % (self.quoteChar, self.endQuoteChar)
return self.strRepr
""" Token for matching words composed of characters I{not} in a given set (will include whitespace in matched characters if not listed in the provided exclusion set - see example). Defined with string containing all disallowed characters, and an optional minimum, maximum, and/or exact length. The default value for C{min} is 1 (a minimum value < 1 is not valid); the default values for C{max} and C{exact} are 0, meaning no maximum or exact length restriction.
Example:: # define a comma-separated-value as anything that is not a ',' csv_value = CharsNotIn(',') print(delimitedList(csv_value).parseString("dkls,lsdkjf,s12 34,@!#,213")) prints:: ['dkls', 'lsdkjf', 's12 34', '@!#', '213'] """
raise ValueError("cannot specify a minimum length < 1; use Optional(CharsNotIn()) if zero-length char group is permitted")
self.maxLen = max else:
raise ParseException(instring, loc, self.errmsg, self)
start = loc loc += 1 notchars = self.notChars maxlen = min( start+self.maxLen, len(instring) ) while loc < maxlen and \ (instring[loc] not in notchars): loc += 1
if loc - start < self.minLen: raise ParseException(instring, loc, self.errmsg, self)
return loc, instring[start:loc]
def __str__( self ): try: return super(CharsNotIn, self).__str__() except Exception: pass
if self.strRepr is None: if len(self.notChars) > 4: self.strRepr = "!W:(%s...)" % self.notChars[:4] else: self.strRepr = "!W:(%s)" % self.notChars
return self.strRepr
""" Special matching class for matching whitespace. Normally, whitespace is ignored by pyparsing grammars. This class is included when some whitespace structures are significant. Define with a string containing the whitespace characters to be matched; default is C{" \\t\\r\\n"}. Also takes optional C{min}, C{max}, and C{exact} arguments, as defined for the C{L{Word}} class. """ " " : "<SPC>", "\t": "<TAB>", "\n": "<LF>", "\r": "<CR>", "\f": "<FF>", } #~ self.leaveWhitespace()
self.maxLen = max else:
self.maxLen = exact self.minLen = exact
if not(instring[ loc ] in self.matchWhite): raise ParseException(instring, loc, self.errmsg, self) start = loc loc += 1 maxloc = start + self.maxLen maxloc = min( maxloc, len(instring) ) while loc < maxloc and instring[loc] in self.matchWhite: loc += 1
if loc - start < self.minLen: raise ParseException(instring, loc, self.errmsg, self)
return loc, instring[start:loc]
""" Token to advance to a specific column of input text; useful for tabular report scraping. """ super(GoToColumn,self).__init__() self.col = colno
if col(loc,instring) != self.col: instrlen = len(instring) if self.ignoreExprs: loc = self._skipIgnorables( instring, loc ) while loc < instrlen and instring[loc].isspace() and col( loc, instring ) != self.col : loc += 1 return loc
thiscol = col( loc, instring ) if thiscol > self.col: raise ParseException( instring, loc, "Text not in expected column", self ) newloc = loc + self.col - thiscol ret = instring[ loc: newloc ] return newloc, ret
""" Matches if current position is at the beginning of a line within the parse string
Example::
test = '''\ AAA this line AAA and this line AAA but not this one B AAA and definitely not this one '''
for t in (LineStart() + 'AAA' + restOfLine).searchString(test): print(t)
Prints:: ['AAA', ' this line'] ['AAA', ' and this line']
"""
if col(loc, instring) == 1: return loc, [] raise ParseException(instring, loc, self.errmsg, self)
""" Matches if current position is at the end of a line within the parse string """
if loc<len(instring): if instring[loc] == "\n": return loc+1, "\n" else: raise ParseException(instring, loc, self.errmsg, self) elif loc == len(instring): return loc+1, [] else: raise ParseException(instring, loc, self.errmsg, self)
""" Matches if current position is at the beginning of the parse string """
# see if entire string up to here is just whitespace and ignoreables if loc != self.preParse( instring, 0 ): raise ParseException(instring, loc, self.errmsg, self)
""" Matches if current position is at the end of the parse string """
raise ParseException(instring, loc, self.errmsg, self) elif loc > len(instring): return loc, [] else: raise ParseException(instring, loc, self.errmsg, self)
""" Matches if the current position is at the beginning of a Word, and is not preceded by any character in a given set of C{wordChars} (default=C{printables}). To emulate the C{\b} behavior of regular expressions, use C{WordStart(alphanums)}. C{WordStart} will also match at the beginning of the string being parsed, or at the beginning of a line. """ super(WordStart,self).__init__() self.wordChars = set(wordChars) self.errmsg = "Not at the start of a word"
if loc != 0: if (instring[loc-1] in self.wordChars or instring[loc] not in self.wordChars): raise ParseException(instring, loc, self.errmsg, self) return loc, []
""" Matches if the current position is at the end of a Word, and is not followed by any character in a given set of C{wordChars} (default=C{printables}). To emulate the C{\b} behavior of regular expressions, use C{WordEnd(alphanums)}. C{WordEnd} will also match at the end of the string being parsed, or at the end of a line. """ super(WordEnd,self).__init__() self.wordChars = set(wordChars) self.skipWhitespace = False self.errmsg = "Not at the end of a word"
instrlen = len(instring) if instrlen>0 and loc<instrlen: if (instring[loc] in self.wordChars or instring[loc-1] not in self.wordChars): raise ParseException(instring, loc, self.errmsg, self) return loc, []
""" Abstract subclass of ParserElement, for combining and post-processing parsed tokens. """ exprs = list(exprs)
self.exprs = [ ParserElement._literalStringClass( exprs ) ] # if sequence of strings provided, wrap with Literal exprs = map(ParserElement._literalStringClass, exprs) else: try: self.exprs = list( exprs ) except TypeError: self.exprs = [ exprs ]
return self.exprs[i]
self.exprs.append( other ) self.strRepr = None return self
"""Extends C{leaveWhitespace} defined in base class, and also invokes C{leaveWhitespace} on all contained expressions."""
if isinstance( other, Suppress ): if other not in self.ignoreExprs: super( ParseExpression, self).ignore( other ) for e in self.exprs: e.ignore( self.ignoreExprs[-1] ) else: super( ParseExpression, self).ignore( other ) for e in self.exprs: e.ignore( self.ignoreExprs[-1] ) return self
def __str__( self ): try: return super(ParseExpression,self).__str__() except Exception: pass
if self.strRepr is None: self.strRepr = "%s:(%s)" % ( self.__class__.__name__, _ustr(self.exprs) ) return self.strRepr
# collapse nested And's of the form And( And( And( a,b), c), d) to And( a,b,c,d ) # but only if there are no parse actions or resultsNames on the nested And's # (likewise for Or's and MatchFirst's) not(other.parseAction) and other.resultsName is None and not other.debug ):
not(other.parseAction) and other.resultsName is None and not other.debug ):
tmp = validateTrace[:]+[self] for e in self.exprs: e.validate(tmp) self.checkRecursion( [] )
""" Requires all given C{ParseExpression}s to be found in the given order. Expressions may be separated by whitespace. May be constructed using the C{'+'} operator. May also be constructed using the C{'-'} operator, which will suppress backtracking.
Example:: integer = Word(nums) name_expr = OneOrMore(Word(alphas))
expr = And([integer("id"),name_expr("name"),integer("age")]) # more easily written as: expr = integer("id") + name_expr("name") + integer("age") """
super(And._ErrorStop,self).__init__(*args, **kwargs) self.name = '-' self.leaveWhitespace()
# pass False as last arg to _parse for first element, since we already # pre-parsed the string as part of our And pre-parsing errorStop = True continue try: loc, exprtokens = e._parse( instring, loc, doActions ) except ParseSyntaxException: raise except ParseBaseException as pe: pe.__traceback__ = None raise ParseSyntaxException._from_exception(pe) except IndexError: raise ParseSyntaxException(instring, len(instring), self.errmsg, self) else:
if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) return self.append( other ) #And( [ self, other ] )
subRecCheckList = parseElementList[:] + [ self ] for e in self.exprs: e.checkRecursion( subRecCheckList ) if not e.mayReturnEmpty: break
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "{" + " ".join(_ustr(e) for e in self.exprs) + "}"
return self.strRepr
""" Requires that at least one C{ParseExpression} is found. If two expressions match, the expression that matches the longest string will be used. May be constructed using the C{'^'} operator.
Example:: # construct Or using '^' operator
number = Word(nums) ^ Combine(Word(nums) + '.' + Word(nums)) print(number.searchString("123 3.1416 789")) prints:: [['123'], ['3.1416'], ['789']] """ else: self.mayReturnEmpty = True
except IndexError: if len(instring) > maxExcLoc: maxException = ParseException(instring,len(instring),e.errmsg,self) maxExcLoc = len(instring) else: # save match among all matches, to retry longest to shortest matches.append((loc2, e))
matches.sort(key=lambda x: -x[0]) for _,e in matches: try: return e._parse( instring, loc, doActions ) except ParseException as err: err.__traceback__ = None if err.loc > maxExcLoc: maxException = err maxExcLoc = err.loc
raise maxException else: raise ParseException(instring, loc, "no defined alternatives to match", self)
if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) return self.append( other ) #Or( [ self, other ] )
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "{" + " ^ ".join(_ustr(e) for e in self.exprs) + "}"
return self.strRepr
subRecCheckList = parseElementList[:] + [ self ] for e in self.exprs: e.checkRecursion( subRecCheckList )
""" Requires that at least one C{ParseExpression} is found. If two expressions match, the first one listed is the one that will match. May be constructed using the C{'|'} operator.
Example:: # construct MatchFirst using '|' operator
# watch the order of expressions to match number = Word(nums) | Combine(Word(nums) + '.' + Word(nums)) print(number.searchString("123 3.1416 789")) # Fail! -> [['123'], ['3'], ['1416'], ['789']]
# put more selective expression first number = Combine(Word(nums) + '.' + Word(nums)) | Word(nums) print(number.searchString("123 3.1416 789")) # Better -> [['123'], ['3.1416'], ['789']] """ else: self.mayReturnEmpty = True
except IndexError: if len(instring) > maxExcLoc: maxException = ParseException(instring,len(instring),e.errmsg,self) maxExcLoc = len(instring)
# only got here if no expression matched, raise exception for match that made it the furthest else: raise maxException else: raise ParseException(instring, loc, "no defined alternatives to match", self)
if isinstance( other, basestring ): other = ParserElement._literalStringClass( other ) return self.append( other ) #MatchFirst( [ self, other ] )
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "{" + " | ".join(_ustr(e) for e in self.exprs) + "}"
return self.strRepr
subRecCheckList = parseElementList[:] + [ self ] for e in self.exprs: e.checkRecursion( subRecCheckList )
""" Requires all given C{ParseExpression}s to be found, but in any order. Expressions may be separated by whitespace. May be constructed using the C{'&'} operator.
Example:: color = oneOf("RED ORANGE YELLOW GREEN BLUE PURPLE BLACK WHITE BROWN") shape_type = oneOf("SQUARE CIRCLE TRIANGLE STAR HEXAGON OCTAGON") integer = Word(nums) shape_attr = "shape:" + shape_type("shape") posn_attr = "posn:" + Group(integer("x") + ',' + integer("y"))("posn") color_attr = "color:" + color("color") size_attr = "size:" + integer("size")
# use Each (using operator '&') to accept attributes in any order # (shape and posn are required, color and size are optional) shape_spec = shape_attr & posn_attr & Optional(color_attr) & Optional(size_attr)
shape_spec.runTests(''' shape: SQUARE color: BLACK posn: 100, 120 shape: CIRCLE size: 50 color: BLUE posn: 50,80 color:GREEN size:20 shape:TRIANGLE posn:20,40 ''' ) prints:: shape: SQUARE color: BLACK posn: 100, 120 ['shape:', 'SQUARE', 'color:', 'BLACK', 'posn:', ['100', ',', '120']] - color: BLACK - posn: ['100', ',', '120'] - x: 100 - y: 120 - shape: SQUARE
shape: CIRCLE size: 50 color: BLUE posn: 50,80 ['shape:', 'CIRCLE', 'size:', '50', 'color:', 'BLUE', 'posn:', ['50', ',', '80']] - color: BLUE - posn: ['50', ',', '80'] - x: 50 - y: 80 - shape: CIRCLE - size: 50
color: GREEN size: 20 shape: TRIANGLE posn: 20,40 ['color:', 'GREEN', 'size:', '20', 'shape:', 'TRIANGLE', 'posn:', ['20', ',', '40']] - color: GREEN - posn: ['20', ',', '40'] - x: 20 - y: 40 - shape: TRIANGLE - size: 20 """ super(Each,self).__init__(exprs, savelist) self.mayReturnEmpty = all(e.mayReturnEmpty for e in self.exprs) self.skipWhitespace = True self.initExprGroups = True
if self.initExprGroups: self.opt1map = dict((id(e.expr),e) for e in self.exprs if isinstance(e,Optional)) opt1 = [ e.expr for e in self.exprs if isinstance(e,Optional) ] opt2 = [ e for e in self.exprs if e.mayReturnEmpty and not isinstance(e,Optional)] self.optionals = opt1 + opt2 self.multioptionals = [ e.expr for e in self.exprs if isinstance(e,ZeroOrMore) ] self.multirequired = [ e.expr for e in self.exprs if isinstance(e,OneOrMore) ] self.required = [ e for e in self.exprs if not isinstance(e,(Optional,ZeroOrMore,OneOrMore)) ] self.required += self.multirequired self.initExprGroups = False tmpLoc = loc tmpReqd = self.required[:] tmpOpt = self.optionals[:] matchOrder = []
keepMatching = True while keepMatching: tmpExprs = tmpReqd + tmpOpt + self.multioptionals + self.multirequired failed = [] for e in tmpExprs: try: tmpLoc = e.tryParse( instring, tmpLoc ) except ParseException: failed.append(e) else: matchOrder.append(self.opt1map.get(id(e),e)) if e in tmpReqd: tmpReqd.remove(e) elif e in tmpOpt: tmpOpt.remove(e) if len(failed) == len(tmpExprs): keepMatching = False
if tmpReqd: missing = ", ".join(_ustr(e) for e in tmpReqd) raise ParseException(instring,loc,"Missing one or more required elements (%s)" % missing )
# add any unmatched Optionals, in case they have default values defined matchOrder += [e for e in self.exprs if isinstance(e,Optional) and e.expr in tmpOpt]
resultlist = [] for e in matchOrder: loc,results = e._parse(instring,loc,doActions) resultlist.append(results)
finalResults = sum(resultlist, ParseResults([])) return loc, finalResults
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "{" + " & ".join(_ustr(e) for e in self.exprs) + "}"
return self.strRepr
subRecCheckList = parseElementList[:] + [ self ] for e in self.exprs: e.checkRecursion( subRecCheckList )
""" Abstract subclass of C{ParserElement}, for combining and post-processing parsed tokens. """ else: expr = ParserElement._literalStringClass(Literal(expr))
else: raise ParseException("",loc,self.errmsg,self)
if isinstance( other, Suppress ): if other not in self.ignoreExprs: super( ParseElementEnhance, self).ignore( other ) if self.expr is not None: self.expr.ignore( self.ignoreExprs[-1] ) else: super( ParseElementEnhance, self).ignore( other ) if self.expr is not None: self.expr.ignore( self.ignoreExprs[-1] ) return self
if self in parseElementList: raise RecursiveGrammarException( parseElementList+[self] ) subRecCheckList = parseElementList[:] + [ self ] if self.expr is not None: self.expr.checkRecursion( subRecCheckList )
tmp = validateTrace[:]+[self] if self.expr is not None: self.expr.validate(tmp) self.checkRecursion( [] )
def __str__( self ): try: return super(ParseElementEnhance,self).__str__() except Exception: pass
if self.strRepr is None and self.expr is not None: self.strRepr = "%s:(%s)" % ( self.__class__.__name__, _ustr(self.expr) ) return self.strRepr
""" Lookahead matching of the given parse expression. C{FollowedBy} does I{not} advance the parsing position within the input string, it only verifies that the specified parse expression matches at the current position. C{FollowedBy} always returns a null token list.
Example:: # use FollowedBy to match a label only if it is followed by a ':' data_word = Word(alphas) label = data_word + FollowedBy(':') attr_expr = Group(label + Suppress(':') + OneOrMore(data_word, stopOn=label).setParseAction(' '.join))
OneOrMore(attr_expr).parseString("shape: SQUARE color: BLACK posn: upper left").pprint() prints:: [['shape', 'SQUARE'], ['color', 'BLACK'], ['posn', 'upper left']] """ super(FollowedBy,self).__init__(expr) self.mayReturnEmpty = True
self.expr.tryParse( instring, loc ) return loc, []
""" Lookahead to disallow matching with the given parse expression. C{NotAny} does I{not} advance the parsing position within the input string, it only verifies that the specified parse expression does I{not} match at the current position. Also, C{NotAny} does I{not} skip over leading whitespace. C{NotAny} always returns a null token list. May be constructed using the '~' operator.
Example::
""" #~ self.leaveWhitespace()
if self.expr.canParseNext(instring, loc): raise ParseException(instring, loc, self.errmsg, self) return loc, []
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "~{" + _ustr(self.expr) + "}"
return self.strRepr
ender = ParserElement._literalStringClass(ender)
try_not_ender = self.not_ender.tryParse
# must be at least one (but first see if we are the stopOn sentinel; # if so, fail) try_not_ender(instring, loc) try_not_ender(instring, loc) preloc = self_skip_ignorables( instring, loc ) else:
""" Repetition of one or more of the given expression.
Parameters: - expr - expression that must match one or more times - stopOn - (default=C{None}) - expression for a terminating sentinel (only required if the sentinel would ordinarily match the repetition expression)
Example:: data_word = Word(alphas) label = data_word + FollowedBy(':') attr_expr = Group(label + Suppress(':') + OneOrMore(data_word).setParseAction(' '.join))
text = "shape: SQUARE posn: upper left color: BLACK" OneOrMore(attr_expr).parseString(text).pprint() # Fail! read 'color' as data instead of next label -> [['shape', 'SQUARE color']]
# use stopOn attribute for OneOrMore to avoid reading label string as part of the data attr_expr = Group(label + Suppress(':') + OneOrMore(data_word, stopOn=label).setParseAction(' '.join)) OneOrMore(attr_expr).parseString(text).pprint() # Better -> [['shape', 'SQUARE'], ['posn', 'upper left'], ['color', 'BLACK']]
# could also be written as (attr_expr * (1,)).parseString(text).pprint() """
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "{" + _ustr(self.expr) + "}..."
return self.strRepr
""" Optional repetition of zero or more of the given expression.
Parameters: - expr - expression that must match zero or more times - stopOn - (default=C{None}) - expression for a terminating sentinel (only required if the sentinel would ordinarily match the repetition expression)
Example: similar to L{OneOrMore} """
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "[" + _ustr(self.expr) + "]..."
return self.strRepr
return False def __str__(self): return ""
""" Optional matching of the given expression.
Parameters: - expr - expression that must match zero or more times - default (optional) - value to be returned if the optional expression is not found.
Example:: # US postal code can be a 5-digit zip, plus optional 4-digit qualifier zip = Combine(Word(nums, exact=5) + Optional('-' + Word(nums, exact=4))) zip.runTests(''' # traditional ZIP code 12345
# ZIP+4 form 12101-0001
# invalid ZIP 98765- ''') prints:: # traditional ZIP code 12345 ['12345']
# ZIP+4 form 12101-0001 ['12101-0001']
# invalid ZIP 98765- ^ FAIL: Expected end of text (at char 5), (line:1, col:6) """
if self.expr.resultsName: tokens = ParseResults([ self.defaultValue ]) tokens[self.expr.resultsName] = self.defaultValue else: tokens = [ self.defaultValue ] else:
def __str__( self ): if hasattr(self,"name"): return self.name
if self.strRepr is None: self.strRepr = "[" + _ustr(self.expr) + "]"
return self.strRepr
""" Token for skipping over all undefined text until the matched expression is found.
Parameters: - expr - target expression marking the end of the data to be skipped - include - (default=C{False}) if True, the target expression is also parsed (the skipped text and target expression are returned as a 2-element list). - ignore - (default=C{None}) used to define grammars (typically quoted strings and comments) that might contain false matches to the target expression - failOn - (default=C{None}) define expressions that are not allowed to be included in the skipped test; if found before the target expression is found, the SkipTo is not a match
Example:: report = ''' Outstanding Issues Report - 1 Jan 2000
# | Severity | Description | Days Open -----+----------+-------------------------------------------+----------- 101 | Critical | Intermittent system crash | 6 94 | Cosmetic | Spelling error on Login ('log|n') | 14 79 | Minor | System slow when running too many reports | 47 ''' integer = Word(nums) SEP = Suppress('|') # use SkipTo to simply match everything up until the next SEP # - ignore quoted strings, so that a '|' character inside a quoted string does not match # - parse action will call token.strip() for each matched token, i.e., the description body string_data = SkipTo(SEP, ignore=quotedString) string_data.setParseAction(tokenMap(str.strip)) ticket_expr = (integer("issue_num") + SEP + string_data("sev") + SEP + string_data("desc") + SEP + integer("days_open"))
for tkt in ticket_expr.searchString(report): print tkt.dump() prints:: ['101', 'Critical', 'Intermittent system crash', '6'] - days_open: 6 - desc: Intermittent system crash - issue_num: 101 - sev: Critical ['94', 'Cosmetic', "Spelling error on Login ('log|n')", '14'] - days_open: 14 - desc: Spelling error on Login ('log|n') - issue_num: 94 - sev: Cosmetic ['79', 'Minor', 'System slow when running too many reports', '47'] - days_open: 47 - desc: System slow when running too many reports - issue_num: 79 - sev: Minor """ super( SkipTo, self ).__init__( other ) self.ignoreExpr = ignore self.mayReturnEmpty = True self.mayIndexError = False self.includeMatch = include self.asList = False if isinstance(failOn, basestring): self.failOn = ParserElement._literalStringClass(failOn) else: self.failOn = failOn self.errmsg = "No match found for "+_ustr(self.expr)
startloc = loc instrlen = len(instring) expr = self.expr expr_parse = self.expr._parse self_failOn_canParseNext = self.failOn.canParseNext if self.failOn is not None else None self_ignoreExpr_tryParse = self.ignoreExpr.tryParse if self.ignoreExpr is not None else None
tmploc = loc while tmploc <= instrlen: if self_failOn_canParseNext is not None: # break if failOn expression matches if self_failOn_canParseNext(instring, tmploc): break
if self_ignoreExpr_tryParse is not None: # advance past ignore expressions while 1: try: tmploc = self_ignoreExpr_tryParse(instring, tmploc) except ParseBaseException: break
try: expr_parse(instring, tmploc, doActions=False, callPreParse=False) except (ParseException, IndexError): # no match, advance loc in string tmploc += 1 else: # matched skipto expr, done break
else: # ran off the end of the input string without matching skipto expr, fail raise ParseException(instring, loc, self.errmsg, self)
# build up return values loc = tmploc skiptext = instring[startloc:loc] skipresult = ParseResults(skiptext)
if self.includeMatch: loc, mat = expr_parse(instring,loc,doActions,callPreParse=False) skipresult += mat
return loc, skipresult
""" Forward declaration of an expression to be defined later - used for recursive grammars, such as algebraic infix notation. When the expression is known, it is assigned to the C{Forward} variable using the '<<' operator.
Note: take care when assigning to C{Forward} not to overlook precedence of operators. Specifically, '|' has a lower precedence than '<<', so that:: fwdExpr << a | b | c will actually be evaluated as:: (fwdExpr << a) | b | c thereby leaving b and c out as parseable alternatives. It is recommended that you explicitly group the values inserted into the C{Forward}:: fwdExpr << (a | b | c) Converting to use the '<<=' operator instead will avoid this problem.
See L{ParseResults.pprint} for an example of a recursive parser created using C{Forward}. """
other = ParserElement._literalStringClass(other)
return self << other
self.skipWhitespace = False return self
if self not in validateTrace: tmp = validateTrace[:]+[self] if self.expr is not None: self.expr.validate(tmp) self.checkRecursion([])
def __str__( self ): if hasattr(self,"name"): return self.name return self.__class__.__name__ + ": ..."
# stubbed out for now - creates awful memory and perf issues self._revertClass = self.__class__ self.__class__ = _ForwardNoRecurse try: if self.expr is not None: retString = _ustr(self.expr) else: retString = "None" finally: self.__class__ = self._revertClass return self.__class__.__name__ + ": " + retString
else: ret = Forward() ret <<= self return ret
def __str__( self ): return "..."
""" Abstract subclass of C{ParseExpression}, for converting parsed results. """
""" Converter to concatenate all matching tokens to a single string. By default, the matching patterns must also be contiguous in the input string; this can be disabled by specifying C{'adjacent=False'} in the constructor.
Example:: real = Word(nums) + '.' + Word(nums) print(real.parseString('3.1416')) # -> ['3', '.', '1416'] # will also erroneously match the following print(real.parseString('3. 1416')) # -> ['3', '.', '1416']
real = Combine(Word(nums) + '.' + Word(nums)) print(real.parseString('3.1416')) # -> ['3.1416'] # no match when there are internal spaces print(real.parseString('3. 1416')) # -> Exception: Expected W:(0123...) """ # suppress whitespace-stripping in contained parse expressions, but re-enable it on the Combine itself
if self.adjacent: ParserElement.ignore(self, other) else: super( Combine, self).ignore( other ) return self
return [ retToks ] else:
""" Converter to return the matched tokens as a list - useful for returning tokens of C{L{ZeroOrMore}} and C{L{OneOrMore}} expressions.
Example:: ident = Word(alphas) num = Word(nums) term = ident | num func = ident + Optional(delimitedList(term)) print(func.parseString("fn a,b,100")) # -> ['fn', 'a', 'b', '100']
func = ident + Group(Optional(delimitedList(term))) print(func.parseString("fn a,b,100")) # -> ['fn', ['a', 'b', '100']] """
""" Converter to return a repetitive expression as a list, but also as a dictionary. Each element can also be referenced using the first token in the expression as its key. Useful for tabular report scraping when the first column can be used as a item key.
Example:: data_word = Word(alphas) label = data_word + FollowedBy(':') attr_expr = Group(label + Suppress(':') + OneOrMore(data_word).setParseAction(' '.join))
text = "shape: SQUARE posn: upper left color: light blue texture: burlap" attr_expr = (label + Suppress(':') + OneOrMore(data_word, stopOn=label).setParseAction(' '.join))
# print attributes as plain groups print(OneOrMore(attr_expr).parseString(text).dump())
# instead of OneOrMore(expr), parse using Dict(OneOrMore(Group(expr))) - Dict will auto-assign names result = Dict(OneOrMore(Group(attr_expr))).parseString(text) print(result.dump())
# access named fields as dict entries, or output as dict print(result['shape']) print(result.asDict()) prints:: ['shape', 'SQUARE', 'posn', 'upper left', 'color', 'light blue', 'texture', 'burlap']
[['shape', 'SQUARE'], ['posn', 'upper left'], ['color', 'light blue'], ['texture', 'burlap']] - color: light blue - posn: upper left - shape: SQUARE - texture: burlap SQUARE {'color': 'light blue', 'posn': 'upper left', 'texture': 'burlap', 'shape': 'SQUARE'} See more examples at L{ParseResults} of accessing fields by results name. """
for i,tok in enumerate(tokenlist): if len(tok) == 0: continue ikey = tok[0] if isinstance(ikey,int): ikey = _ustr(tok[0]).strip() if len(tok)==1: tokenlist[ikey] = _ParseResultsWithOffset("",i) elif len(tok)==2 and not isinstance(tok[1],ParseResults): tokenlist[ikey] = _ParseResultsWithOffset(tok[1],i) else: dictvalue = tok.copy() #ParseResults(i) del dictvalue[0] if len(dictvalue)!= 1 or (isinstance(dictvalue,ParseResults) and dictvalue.haskeys()): tokenlist[ikey] = _ParseResultsWithOffset(dictvalue,i) else: tokenlist[ikey] = _ParseResultsWithOffset(dictvalue[0],i)
if self.resultsName: return [ tokenlist ] else: return tokenlist
""" Converter for ignoring the results of a parsed expression.
Example:: source = "a, b, c,d" wd = Word(alphas) wd_list1 = wd + ZeroOrMore(',' + wd) print(wd_list1.parseString(source))
# often, delimiters that are useful during parsing are just in the # way afterward - use Suppress to keep them out of the parsed output wd_list2 = wd + ZeroOrMore(Suppress(',') + wd) print(wd_list2.parseString(source)) prints:: ['a', ',', 'b', ',', 'c', ',', 'd'] ['a', 'b', 'c', 'd'] (See also L{delimitedList}.) """
return self
""" Wrapper for parse actions, to ensure they are only called once. """ self.callable = _trim_arity(methodCall) self.called = False if not self.called: results = self.callable(s,l,t) self.called = True return results raise ParseException(s,l,"") self.called = False
""" Decorator for debugging parse actions.
When the parse action is called, this decorator will print C{">> entering I{method-name}(line:I{current_source_line}, I{parse_location}, I{matched_tokens})".} When the parse action completes, the decorator will print C{"<<"} followed by the returned value, or any exception that the parse action raised.
Example:: wd = Word(alphas)
@traceParseAction def remove_duplicate_chars(tokens): return ''.join(sorted(set(''.join(tokens))))
wds = OneOrMore(wd).setParseAction(remove_duplicate_chars) print(wds.parseString("slkdjs sld sldd sdlf sdljf")) prints:: >>entering remove_duplicate_chars(line: 'slkdjs sld sldd sdlf sdljf', 0, (['slkdjs', 'sld', 'sldd', 'sdlf', 'sdljf'], {})) <<leaving remove_duplicate_chars (ret: 'dfjkls') ['dfjkls'] """ f = _trim_arity(f) def z(*paArgs): thisFunc = f.__name__ s,l,t = paArgs[-3:] if len(paArgs)>3: thisFunc = paArgs[0].__class__.__name__ + '.' + thisFunc sys.stderr.write( ">>entering %s(line: '%s', %d, %r)\n" % (thisFunc,line(l,s),l,t) ) try: ret = f(*paArgs) except Exception as exc: sys.stderr.write( "<<leaving %s (exception: %s)\n" % (thisFunc,exc) ) raise sys.stderr.write( "<<leaving %s (ret: %r)\n" % (thisFunc,ret) ) return ret try: z.__name__ = f.__name__ except AttributeError: pass return z
# # global helpers # """ Helper to define a delimited list of expressions - the delimiter defaults to ','. By default, the list elements and delimiters can have intervening whitespace, and comments, but this can be overridden by passing C{combine=True} in the constructor. If C{combine} is set to C{True}, the matching tokens are returned as a single token string, with the delimiters included; otherwise, the matching tokens are returned as a list of tokens, with the delimiters suppressed.
Example:: delimitedList(Word(alphas)).parseString("aa,bb,cc") # -> ['aa', 'bb', 'cc'] delimitedList(Word(hexnums), delim=':', combine=True).parseString("AA:BB:CC:DD:EE") # -> ['AA:BB:CC:DD:EE'] """ return Combine( expr + ZeroOrMore( delim + expr ) ).setName(dlName) else:
""" Helper to define a counted list of expressions. This helper defines a pattern of the form:: integer expr expr expr... where the leading integer tells how many expr expressions follow. The matched tokens returns the array of expr tokens as a list - the leading count token is suppressed.
If C{intExpr} is specified, it should be a pyparsing expression that produces an integer value.
Example:: countedArray(Word(alphas)).parseString('2 ab cd ef') # -> ['ab', 'cd']
# in this parser, the leading integer value is given in binary, # '10' indicating that 2 values are in the array binaryConstant = Word('01').setParseAction(lambda t: int(t[0], 2)) countedArray(Word(alphas), intExpr=binaryConstant).parseString('10 ab cd ef') # -> ['ab', 'cd'] """ arrayExpr = Forward() def countFieldParseAction(s,l,t): n = t[0] arrayExpr << (n and Group(And([expr]*n)) or Group(empty)) return [] if intExpr is None: intExpr = Word(nums).setParseAction(lambda t:int(t[0])) else: intExpr = intExpr.copy() intExpr.setName("arrayLen") intExpr.addParseAction(countFieldParseAction, callDuringTry=True) return ( intExpr + arrayExpr ).setName('(len) ' + _ustr(expr) + '...')
ret = [] for i in L: if isinstance(i,list): ret.extend(_flatten(i)) else: ret.append(i) return ret
""" Helper to define an expression that is indirectly defined from the tokens matched in a previous expression, that is, it looks for a 'repeat' of a previous expression. For example:: first = Word(nums) second = matchPreviousLiteral(first) matchExpr = first + ":" + second will match C{"1:1"}, but not C{"1:2"}. Because this matches a previous literal, will also match the leading C{"1:1"} in C{"1:10"}. If this is not desired, use C{matchPreviousExpr}. Do I{not} use with packrat parsing enabled. """ rep = Forward() def copyTokenToRepeater(s,l,t): if t: if len(t) == 1: rep << t[0] else: # flatten t tokens tflat = _flatten(t.asList()) rep << And(Literal(tt) for tt in tflat) else: rep << Empty() expr.addParseAction(copyTokenToRepeater, callDuringTry=True) rep.setName('(prev) ' + _ustr(expr)) return rep
""" Helper to define an expression that is indirectly defined from the tokens matched in a previous expression, that is, it looks for a 'repeat' of a previous expression. For example:: first = Word(nums) second = matchPreviousExpr(first) matchExpr = first + ":" + second will match C{"1:1"}, but not C{"1:2"}. Because this matches by expressions, will I{not} match the leading C{"1:1"} in C{"1:10"}; the expressions are evaluated first, and then compared, so C{"1"} is compared with C{"10"}. Do I{not} use with packrat parsing enabled. """ rep = Forward() e2 = expr.copy() rep <<= e2 def copyTokenToRepeater(s,l,t): matchTokens = _flatten(t.asList()) def mustMatchTheseTokens(s,l,t): theseTokens = _flatten(t.asList()) if theseTokens != matchTokens: raise ParseException("",0,"") rep.setParseAction( mustMatchTheseTokens, callDuringTry=True ) expr.addParseAction(copyTokenToRepeater, callDuringTry=True) rep.setName('(prev) ' + _ustr(expr)) return rep
#~ escape these chars: ^-]
""" Helper to quickly define a set of alternative Literals, and makes sure to do longest-first testing when there is a conflict, regardless of the input order, but returns a C{L{MatchFirst}} for best performance.
Parameters: - strs - a string of space-delimited literals, or a collection of string literals - caseless - (default=C{False}) - treat all literals as caseless - useRegex - (default=C{True}) - as an optimization, will generate a Regex object; otherwise, will generate a C{MatchFirst} object (if C{caseless=True}, or if creating a C{Regex} raises an exception)
Example:: comp_oper = oneOf("< = > <= >= !=") var = Word(alphas) number = Word(nums) term = var | number comparison_expr = term + comp_oper + term print(comparison_expr.searchString("B = 12 AA=23 B<=AA AA>12")) prints:: [['B', '=', '12'], ['AA', '=', '23'], ['B', '<=', 'AA'], ['AA', '>', '12']] """ if caseless: isequal = ( lambda a,b: a.upper() == b.upper() ) masks = ( lambda a,b: b.upper().startswith(a.upper()) ) parseElementClass = CaselessLiteral else: isequal = ( lambda a,b: a == b ) masks = ( lambda a,b: b.startswith(a) ) parseElementClass = Literal
symbols = [] if isinstance(strs,basestring): symbols = strs.split() elif isinstance(strs, Iterable): symbols = list(strs) else: warnings.warn("Invalid argument to oneOf, expected string or iterable", SyntaxWarning, stacklevel=2) if not symbols: return NoMatch()
i = 0 while i < len(symbols)-1: cur = symbols[i] for j,other in enumerate(symbols[i+1:]): if ( isequal(other, cur) ): del symbols[i+j+1] break elif ( masks(cur, other) ): del symbols[i+j+1] symbols.insert(i,other) cur = other break else: i += 1
if not caseless and useRegex: #~ print (strs,"->", "|".join( [ _escapeRegexChars(sym) for sym in symbols] )) try: if len(symbols)==len("".join(symbols)): return Regex( "[%s]" % "".join(_escapeRegexRangeChars(sym) for sym in symbols) ).setName(' | '.join(symbols)) else: return Regex( "|".join(re.escape(sym) for sym in symbols) ).setName(' | '.join(symbols)) except Exception: warnings.warn("Exception creating Regex for oneOf, building MatchFirst", SyntaxWarning, stacklevel=2)
# last resort, just use MatchFirst return MatchFirst(parseElementClass(sym) for sym in symbols).setName(' | '.join(symbols))
""" Helper to easily and clearly define a dictionary by specifying the respective patterns for the key and value. Takes care of defining the C{L{Dict}}, C{L{ZeroOrMore}}, and C{L{Group}} tokens in the proper order. The key pattern can include delimiting markers or punctuation, as long as they are suppressed, thereby leaving the significant key text. The value pattern can include named results, so that the C{Dict} results can include named token fields.
Example:: text = "shape: SQUARE posn: upper left color: light blue texture: burlap" attr_expr = (label + Suppress(':') + OneOrMore(data_word, stopOn=label).setParseAction(' '.join)) print(OneOrMore(attr_expr).parseString(text).dump())
attr_label = label attr_value = Suppress(':') + OneOrMore(data_word, stopOn=label).setParseAction(' '.join)
# similar to Dict, but simpler call format result = dictOf(attr_label, attr_value).parseString(text) print(result.dump()) print(result['shape']) print(result.shape) # object attribute access works too print(result.asDict()) prints:: [['shape', 'SQUARE'], ['posn', 'upper left'], ['color', 'light blue'], ['texture', 'burlap']] - color: light blue - posn: upper left - shape: SQUARE - texture: burlap SQUARE SQUARE {'color': 'light blue', 'shape': 'SQUARE', 'posn': 'upper left', 'texture': 'burlap'} """ return Dict( ZeroOrMore( Group ( key + value ) ) )
""" Helper to return the original, untokenized text for a given expression. Useful to restore the parsed fields of an HTML start tag into the raw tag text itself, or to revert separate tokens with intervening whitespace back to the original matching input text. By default, returns astring containing the original parsed text.
If the optional C{asString} argument is passed as C{False}, then the return value is a C{L{ParseResults}} containing any results names that were originally matched, and a single token containing the original matched text from the input string. So if the expression passed to C{L{originalTextFor}} contains expressions with defined results names, you must set C{asString} to C{False} if you want to preserve those results name values.
Example:: src = "this is test <b> bold <i>text</i> </b> normal text " for tag in ("b","i"): opener,closer = makeHTMLTags(tag) patt = originalTextFor(opener + SkipTo(closer) + closer) print(patt.searchString(src)[0]) prints:: ['<b> bold <i>text</i> </b>'] ['<i>text</i>'] """ else: def extractText(s,l,t): t[:] = [s[t.pop('_original_start'):t.pop('_original_end')]]
""" Helper to undo pyparsing's default grouping of And expressions, even if all but one are non-empty. """ return TokenConverter(expr).setParseAction(lambda t:t[0])
""" Helper to decorate a returned token with its starting and ending locations in the input string. This helper adds the following results names: - locn_start = location where matched expression begins - locn_end = location where matched expression ends - value = the actual parsed results
Be careful if the input text contains C{<TAB>} characters, you may want to call C{L{ParserElement.parseWithTabs}}
Example:: wd = Word(alphas) for match in locatedExpr(wd).searchString("ljsdf123lksdjjf123lkkjj1222"): print(match) prints:: [[0, 'ljsdf', 5]] [[8, 'lksdjjf', 15]] [[18, 'lkkjj', 23]] """ locator = Empty().setParseAction(lambda s,l,t: l) return Group(locator("locn_start") + expr("value") + locator.copy().leaveWhitespace()("locn_end"))
# convenience constants for positional expressions
r""" Helper to easily define string ranges for use in Word construction. Borrows syntax from regexp '[]' string range definitions:: srange("[0-9]") -> "0123456789" srange("[a-z]") -> "abcdefghijklmnopqrstuvwxyz" srange("[a-z$_]") -> "abcdefghijklmnopqrstuvwxyz$_" The input string must be enclosed in []'s, and the returned string is the expanded character set joined into a single string. The values enclosed in the []'s may be: - a single character - an escaped character with a leading backslash (such as C{\-} or C{\]}) - an escaped hex character with a leading C{'\x'} (C{\x21}, which is a C{'!'} character) (C{\0x##} is also supported for backwards compatibility) - an escaped octal character with a leading C{'\0'} (C{\041}, which is a C{'!'} character) - a range of any of the above, separated by a dash (C{'a-z'}, etc.) - any combination of the above (C{'aeiouy'}, C{'a-zA-Z0-9_$'}, etc.) """ except Exception: return ""
""" Helper method for defining parse actions that require matching at a specific column in the input text. """ def verifyCol(strg,locn,toks): if col(locn,strg) != n: raise ParseException(strg,locn,"matched token not at column %d" % n) return verifyCol
""" Helper method for common parse actions that simply return a literal value. Especially useful when used with C{L{transformString<ParserElement.transformString>}()}.
Example:: num = Word(nums).setParseAction(lambda toks: int(toks[0])) na = oneOf("N/A NA").setParseAction(replaceWith(math.nan)) term = na | num
OneOrMore(term).parseString("324 234 N/A 234") # -> [324, 234, nan, 234] """ return lambda s,l,t: [replStr]
""" Helper parse action for removing quotation marks from parsed quoted strings.
Example:: # by default, quotation marks are included in parsed results quotedString.parseString("'Now is the Winter of our Discontent'") # -> ["'Now is the Winter of our Discontent'"]
# use removeQuotes to strip quotation marks from parsed results quotedString.setParseAction(removeQuotes) quotedString.parseString("'Now is the Winter of our Discontent'") # -> ["Now is the Winter of our Discontent"] """ return t[0][1:-1]
""" Helper to define a parse action by mapping a function to all elements of a ParseResults list.If any additional args are passed, they are forwarded to the given function as additional arguments after the token, as in C{hex_integer = Word(hexnums).setParseAction(tokenMap(int, 16))}, which will convert the parsed data to an integer using base 16.
Example (compare the last to example in L{ParserElement.transformString}:: hex_ints = OneOrMore(Word(hexnums)).setParseAction(tokenMap(int, 16)) hex_ints.runTests(''' 00 11 22 aa FF 0a 0d 1a ''')
upperword = Word(alphas).setParseAction(tokenMap(str.upper)) OneOrMore(upperword).runTests(''' my kingdom for a horse ''')
wd = Word(alphas).setParseAction(tokenMap(str.title)) OneOrMore(wd).setParseAction(' '.join).runTests(''' now is the winter of our discontent made glorious summer by this sun of york ''') prints:: 00 11 22 aa FF 0a 0d 1a [0, 17, 34, 170, 255, 10, 13, 26]
my kingdom for a horse ['MY', 'KINGDOM', 'FOR', 'A', 'HORSE']
now is the winter of our discontent made glorious summer by this sun of york ['Now Is The Winter Of Our Discontent Made Glorious Summer By This Sun Of York'] """ return [func(tokn, *args) for tokn in t]
getattr(func, '__class__').__name__) except Exception: func_name = str(func)
"""(Deprecated) Helper parse action to convert tokens to upper case. Deprecated in favor of L{pyparsing_common.upcaseTokens}"""
"""(Deprecated) Helper parse action to convert tokens to lower case. Deprecated in favor of L{pyparsing_common.downcaseTokens}"""
"""Internal helper to construct opening and closing tag expressions, given a tag name""" resname = tagStr tagStr = Keyword(tagStr, caseless=not xml) else:
tagAttrValue = dblQuotedString.copy().setParseAction( removeQuotes ) openTag = Suppress("<") + tagStr("tag") + \ Dict(ZeroOrMore(Group( tagAttrName + Suppress("=") + tagAttrValue ))) + \ Optional("/",default=[False]).setResultsName("empty").setParseAction(lambda s,l,t:t[0]=='/') + Suppress(">") else: Dict(ZeroOrMore(Group( tagAttrName.setParseAction(downcaseTokens) + \ Optional( Suppress("=") + tagAttrValue ) ))) + \ Optional("/",default=[False]).setResultsName("empty").setParseAction(lambda s,l,t:t[0]=='/') + Suppress(">")
""" Helper to construct opening and closing tag expressions for HTML, given a tag name. Matches tags in either upper or lower case, attributes with namespaces and with quoted or unquoted values.
Example:: text = '<td>More info at the <a href="http://pyparsing.wikispaces.com">pyparsing</a> wiki page</td>' # makeHTMLTags returns pyparsing expressions for the opening and closing tags as a 2-tuple a,a_end = makeHTMLTags("A") link_expr = a + SkipTo(a_end)("link_text") + a_end
for link in link_expr.searchString(text): # attributes in the <A> tag (like "href" shown here) are also accessible as named results print(link.link_text, '->', link.href) prints:: pyparsing -> http://pyparsing.wikispaces.com """
""" Helper to construct opening and closing tag expressions for XML, given a tag name. Matches tags only in the given upper/lower case.
Example: similar to L{makeHTMLTags} """ return _makeTags( tagStr, True )
""" Helper to create a validating parse action to be used with start tags created with C{L{makeXMLTags}} or C{L{makeHTMLTags}}. Use C{withAttribute} to qualify a starting tag with a required attribute value, to avoid false matches on common tags such as C{<TD>} or C{<DIV>}.
Call C{withAttribute} with a series of attribute names and values. Specify the list of filter attributes names and values as: - keyword arguments, as in C{(align="right")}, or - as an explicit dict with C{**} operator, when an attribute name is also a Python reserved word, as in C{**{"class":"Customer", "align":"right"}} - a list of name-value tuples, as in ( ("ns1:class", "Customer"), ("ns2:align","right") ) For attribute names with a namespace prefix, you must use the second form. Attribute names are matched insensitive to upper/lower case.
If just testing for C{class} (with or without a namespace), use C{L{withClass}}.
To verify that the attribute exists, but without specifying a value, pass C{withAttribute.ANY_VALUE} as the value.
Example:: html = ''' <div> Some text <div type="grid">1 4 0 1 0</div> <div type="graph">1,3 2,3 1,1</div> <div>this has no type</div> </div>
''' div,div_end = makeHTMLTags("div")
# only match div tag having a type attribute with value "grid" div_grid = div().setParseAction(withAttribute(type="grid")) grid_expr = div_grid + SkipTo(div | div_end)("body") for grid_header in grid_expr.searchString(html): print(grid_header.body)
# construct a match with any div tag having a type attribute, regardless of the value div_any_type = div().setParseAction(withAttribute(type=withAttribute.ANY_VALUE)) div_expr = div_any_type + SkipTo(div | div_end)("body") for div_header in div_expr.searchString(html): print(div_header.body) prints:: 1 4 0 1 0
1 4 0 1 0 1,3 2,3 1,1 """ if args: attrs = args[:] else: attrs = attrDict.items() attrs = [(k,v) for k,v in attrs] def pa(s,l,tokens): for attrName,attrValue in attrs: if attrName not in tokens: raise ParseException(s,l,"no matching attribute " + attrName) if attrValue != withAttribute.ANY_VALUE and tokens[attrName] != attrValue: raise ParseException(s,l,"attribute '%s' has value '%s', must be '%s'" % (attrName, tokens[attrName], attrValue)) return pa
""" Simplified version of C{L{withAttribute}} when matching on a div class - made difficult because C{class} is a reserved word in Python.
Example:: html = ''' <div> Some text <div class="grid">1 4 0 1 0</div> <div class="graph">1,3 2,3 1,1</div> <div>this <div> has no class</div> </div>
''' div,div_end = makeHTMLTags("div") div_grid = div().setParseAction(withClass("grid"))
grid_expr = div_grid + SkipTo(div | div_end)("body") for grid_header in grid_expr.searchString(html): print(grid_header.body)
div_any_type = div().setParseAction(withClass(withAttribute.ANY_VALUE)) div_expr = div_any_type + SkipTo(div | div_end)("body") for div_header in div_expr.searchString(html): print(div_header.body) prints:: 1 4 0 1 0
1 4 0 1 0 1,3 2,3 1,1 """ classattr = "%s:class" % namespace if namespace else "class" return withAttribute(**{classattr : classname})
""" Helper method for constructing grammars of expressions made up of operators working in a precedence hierarchy. Operators may be unary or binary, left- or right-associative. Parse actions can also be attached to operator expressions. The generated parser will also recognize the use of parentheses to override operator precedences (see example below).
Note: if you define a deep operator list, you may see performance issues when using infixNotation. See L{ParserElement.enablePackrat} for a mechanism to potentially improve your parser performance.
Parameters: - baseExpr - expression representing the most basic element for the nested - opList - list of tuples, one for each operator precedence level in the expression grammar; each tuple is of the form (opExpr, numTerms, rightLeftAssoc, parseAction), where: - opExpr is the pyparsing expression for the operator; may also be a string, which will be converted to a Literal; if numTerms is 3, opExpr is a tuple of two expressions, for the two operators separating the 3 terms - numTerms is the number of terms for this operator (must be 1, 2, or 3) - rightLeftAssoc is the indicator whether the operator is right or left associative, using the pyparsing-defined constants C{opAssoc.RIGHT} and C{opAssoc.LEFT}. - parseAction is the parse action to be associated with expressions matching this operator expression (the parse action tuple member may be omitted); if the parse action is passed a tuple or list of functions, this is equivalent to calling C{setParseAction(*fn)} (L{ParserElement.setParseAction}) - lpar - expression for matching left-parentheses (default=C{Suppress('(')}) - rpar - expression for matching right-parentheses (default=C{Suppress(')')})
Example:: # simple example of four-function arithmetic with ints and variable names integer = pyparsing_common.signed_integer varname = pyparsing_common.identifier
arith_expr = infixNotation(integer | varname, [ ('-', 1, opAssoc.RIGHT), (oneOf('* /'), 2, opAssoc.LEFT), (oneOf('+ -'), 2, opAssoc.LEFT), ])
arith_expr.runTests(''' 5+3*6 (5+3)*6 -2--11 ''', fullDump=False) prints:: 5+3*6 [[5, '+', [3, '*', 6]]]
(5+3)*6 [[[5, '+', 3], '*', 6]]
-2--11 [[['-', 2], '-', ['-', 11]]] """ ret = Forward() lastExpr = baseExpr | ( lpar + ret + rpar ) for i,operDef in enumerate(opList): opExpr,arity,rightLeftAssoc,pa = (operDef + (None,))[:4] termName = "%s term" % opExpr if arity < 3 else "%s%s term" % opExpr if arity == 3: if opExpr is None or len(opExpr) != 2: raise ValueError("if numterms=3, opExpr must be a tuple or list of two expressions") opExpr1, opExpr2 = opExpr thisExpr = Forward().setName(termName) if rightLeftAssoc == opAssoc.LEFT: if arity == 1: matchExpr = FollowedBy(lastExpr + opExpr) + Group( lastExpr + OneOrMore( opExpr ) ) elif arity == 2: if opExpr is not None: matchExpr = FollowedBy(lastExpr + opExpr + lastExpr) + Group( lastExpr + OneOrMore( opExpr + lastExpr ) ) else: matchExpr = FollowedBy(lastExpr+lastExpr) + Group( lastExpr + OneOrMore(lastExpr) ) elif arity == 3: matchExpr = FollowedBy(lastExpr + opExpr1 + lastExpr + opExpr2 + lastExpr) + \ Group( lastExpr + opExpr1 + lastExpr + opExpr2 + lastExpr ) else: raise ValueError("operator must be unary (1), binary (2), or ternary (3)") elif rightLeftAssoc == opAssoc.RIGHT: if arity == 1: # try to avoid LR with this extra test if not isinstance(opExpr, Optional): opExpr = Optional(opExpr) matchExpr = FollowedBy(opExpr.expr + thisExpr) + Group( opExpr + thisExpr ) elif arity == 2: if opExpr is not None: matchExpr = FollowedBy(lastExpr + opExpr + thisExpr) + Group( lastExpr + OneOrMore( opExpr + thisExpr ) ) else: matchExpr = FollowedBy(lastExpr + thisExpr) + Group( lastExpr + OneOrMore( thisExpr ) ) elif arity == 3: matchExpr = FollowedBy(lastExpr + opExpr1 + thisExpr + opExpr2 + thisExpr) + \ Group( lastExpr + opExpr1 + thisExpr + opExpr2 + thisExpr ) else: raise ValueError("operator must be unary (1), binary (2), or ternary (3)") else: raise ValueError("operator must indicate right or left associativity") if pa: if isinstance(pa, (tuple, list)): matchExpr.setParseAction(*pa) else: matchExpr.setParseAction(pa) thisExpr <<= ( matchExpr.setName(termName) | lastExpr ) lastExpr = thisExpr ret <<= lastExpr return ret
"""(Deprecated) Former name of C{L{infixNotation}}, will be dropped in a future release."""
Regex(r"'(?:[^'\n\r\\]|(?:'')|(?:\\(?:[^x]|x[0-9a-fA-F]+)))*")+"'").setName("quotedString using single or double quotes")
""" Helper method for defining nested lists enclosed in opening and closing delimiters ("(" and ")" are the default).
Parameters: - opener - opening character for a nested list (default=C{"("}); can also be a pyparsing expression - closer - closing character for a nested list (default=C{")"}); can also be a pyparsing expression - content - expression for items within the nested lists (default=C{None}) - ignoreExpr - expression for ignoring opening and closing delimiters (default=C{quotedString})
If an expression is not provided for the content argument, the nested expression will capture all whitespace-delimited content between delimiters as a list of separate values.
Use the C{ignoreExpr} argument to define expressions that may contain opening or closing characters that should not be treated as opening or closing characters for nesting, such as quotedString or a comment expression. Specify multiple expressions using an C{L{Or}} or C{L{MatchFirst}}. The default is L{quotedString}, but if no expressions are to be ignored, then pass C{None} for this argument.
Example:: data_type = oneOf("void int short long char float double") decl_data_type = Combine(data_type + Optional(Word('*'))) ident = Word(alphas+'_', alphanums+'_') number = pyparsing_common.number arg = Group(decl_data_type + ident) LPAR,RPAR = map(Suppress, "()")
code_body = nestedExpr('{', '}', ignoreExpr=(quotedString | cStyleComment))
c_function = (decl_data_type("type") + ident("name") + LPAR + Optional(delimitedList(arg), [])("args") + RPAR + code_body("body")) c_function.ignore(cStyleComment)
source_code = ''' int is_odd(int x) { return (x%2); }
int dec_to_hex(char hchar) { if (hchar >= '0' && hchar <= '9') { return (ord(hchar)-ord('0')); } else { return (10+ord(hchar)-ord('A')); } } ''' for func in c_function.searchString(source_code): print("%(name)s (%(type)s) args: %(args)s" % func)
prints:: is_odd (int) args: [['int', 'x']] dec_to_hex (int) args: [['char', 'hchar']] """ if opener == closer: raise ValueError("opening and closing strings cannot be the same") if content is None: if isinstance(opener,basestring) and isinstance(closer,basestring): if len(opener) == 1 and len(closer)==1: if ignoreExpr is not None: content = (Combine(OneOrMore(~ignoreExpr + CharsNotIn(opener+closer+ParserElement.DEFAULT_WHITE_CHARS,exact=1)) ).setParseAction(lambda t:t[0].strip())) else: content = (empty.copy()+CharsNotIn(opener+closer+ParserElement.DEFAULT_WHITE_CHARS ).setParseAction(lambda t:t[0].strip())) else: if ignoreExpr is not None: content = (Combine(OneOrMore(~ignoreExpr + ~Literal(opener) + ~Literal(closer) + CharsNotIn(ParserElement.DEFAULT_WHITE_CHARS,exact=1)) ).setParseAction(lambda t:t[0].strip())) else: content = (Combine(OneOrMore(~Literal(opener) + ~Literal(closer) + CharsNotIn(ParserElement.DEFAULT_WHITE_CHARS,exact=1)) ).setParseAction(lambda t:t[0].strip())) else: raise ValueError("opening and closing arguments must be strings if no content expression is given") ret = Forward() if ignoreExpr is not None: ret <<= Group( Suppress(opener) + ZeroOrMore( ignoreExpr | ret | content ) + Suppress(closer) ) else: ret <<= Group( Suppress(opener) + ZeroOrMore( ret | content ) + Suppress(closer) ) ret.setName('nested %s%s expression' % (opener,closer)) return ret
""" Helper method for defining space-delimited indentation blocks, such as those used to define block statements in Python source code.
Parameters: - blockStatementExpr - expression defining syntax of statement that is repeated within the indented block - indentStack - list created by caller to manage indentation stack (multiple statementWithIndentedBlock expressions within a single grammar should share a common indentStack) - indent - boolean indicating whether block must be indented beyond the the current level; set to False for block of left-most statements (default=C{True})
A valid block must contain at least one C{blockStatement}.
Example:: data = ''' def A(z): A1 B = 100 G = A2 A2 A3 B def BB(a,b,c): BB1 def BBA(): bba1 bba2 bba3 C D def spam(x,y): def eggs(z): pass '''
indentStack = [1] stmt = Forward()
identifier = Word(alphas, alphanums) funcDecl = ("def" + identifier + Group( "(" + Optional( delimitedList(identifier) ) + ")" ) + ":") func_body = indentedBlock(stmt, indentStack) funcDef = Group( funcDecl + func_body )
rvalue = Forward() funcCall = Group(identifier + "(" + Optional(delimitedList(rvalue)) + ")") rvalue << (funcCall | identifier | Word(nums)) assignment = Group(identifier + "=" + rvalue) stmt << ( funcDef | assignment | identifier )
module_body = OneOrMore(stmt)
parseTree = module_body.parseString(data) parseTree.pprint() prints:: [['def', 'A', ['(', 'z', ')'], ':', [['A1'], [['B', '=', '100']], [['G', '=', 'A2']], ['A2'], ['A3']]], 'B', ['def', 'BB', ['(', 'a', 'b', 'c', ')'], ':', [['BB1'], [['def', 'BBA', ['(', ')'], ':', [['bba1'], ['bba2'], ['bba3']]]]]], 'C', 'D', ['def', 'spam', ['(', 'x', 'y', ')'], ':', [[['def', 'eggs', ['(', 'z', ')'], ':', [['pass']]]]]]] """ def checkPeerIndent(s,l,t): if l >= len(s): return curCol = col(l,s) if curCol != indentStack[-1]: if curCol > indentStack[-1]: raise ParseFatalException(s,l,"illegal nesting") raise ParseException(s,l,"not a peer entry")
def checkSubIndent(s,l,t): curCol = col(l,s) if curCol > indentStack[-1]: indentStack.append( curCol ) else: raise ParseException(s,l,"not a subentry")
def checkUnindent(s,l,t): if l >= len(s): return curCol = col(l,s) if not(indentStack and curCol < indentStack[-1] and curCol <= indentStack[-2]): raise ParseException(s,l,"not an unindent") indentStack.pop()
NL = OneOrMore(LineEnd().setWhitespaceChars("\t ").suppress()) INDENT = (Empty() + Empty().setParseAction(checkSubIndent)).setName('INDENT') PEER = Empty().setParseAction(checkPeerIndent).setName('') UNDENT = Empty().setParseAction(checkUnindent).setName('UNINDENT') if indent: smExpr = Group( Optional(NL) + #~ FollowedBy(blockStatementExpr) + INDENT + (OneOrMore( PEER + Group(blockStatementExpr) + Optional(NL) )) + UNDENT) else: smExpr = Group( Optional(NL) + (OneOrMore( PEER + Group(blockStatementExpr) + Optional(NL) )) ) blockStatementExpr.ignore(_bslash + LineEnd()) return smExpr.setName('indented block')
"""Helper parser action to replace common HTML entities with their special characters""" return _htmlEntityMap.get(t.entity)
# it's easy to get these comment structures wrong - they're very common, so may as well make them available "Comment of the form C{/* ... */}"
"Comment of the form C{<!-- ... -->}"
"Comment of the form C{// ... (to end of line)}"
"Comment of either form C{L{cStyleComment}} or C{L{dblSlashComment}}"
"Same as C{L{cppStyleComment}}"
"Comment of the form C{# ... (to end of line)}"
Optional( Word(" \t") + ~Literal(",") + ~LineEnd() ) ) ).streamline().setName("commaItem") """(Deprecated) Predefined expression of 1 or more printable words or quoted strings, separated by commas. This expression is deprecated in favor of L{pyparsing_common.comma_separated_list}."""
# some other useful expressions - using lower-case class name since we are really using this as a namespace """ Here are some common low-level expressions that may be useful in jump-starting parser development: - numeric forms (L{integers<integer>}, L{reals<real>}, L{scientific notation<sci_real>}) - common L{programming identifiers<identifier>} - network addresses (L{MAC<mac_address>}, L{IPv4<ipv4_address>}, L{IPv6<ipv6_address>}) - ISO8601 L{dates<iso8601_date>} and L{datetime<iso8601_datetime>} - L{UUID<uuid>} - L{comma-separated list<comma_separated_list>} Parse actions: - C{L{convertToInteger}} - C{L{convertToFloat}} - C{L{convertToDate}} - C{L{convertToDatetime}} - C{L{stripHTMLTags}} - C{L{upcaseTokens}} - C{L{downcaseTokens}}
Example:: pyparsing_common.number.runTests(''' # any int or real number, returned as the appropriate type 100 -100 +100 3.14159 6.02e23 1e-12 ''')
pyparsing_common.fnumber.runTests(''' # any int or real number, returned as float 100 -100 +100 3.14159 6.02e23 1e-12 ''')
pyparsing_common.hex_integer.runTests(''' # hex numbers 100 FF ''')
pyparsing_common.fraction.runTests(''' # fractions 1/2 -3/4 ''')
pyparsing_common.mixed_integer.runTests(''' # mixed fractions 1 1/2 -3/4 1-3/4 ''')
import uuid pyparsing_common.uuid.setParseAction(tokenMap(uuid.UUID)) pyparsing_common.uuid.runTests(''' # uuid 12345678-1234-5678-1234-567812345678 ''') prints:: # any int or real number, returned as the appropriate type 100 [100]
-100 [-100]
+100 [100]
3.14159 [3.14159]
6.02e23 [6.02e+23]
1e-12 [1e-12]
# any int or real number, returned as float 100 [100.0]
-100 [-100.0]
+100 [100.0]
3.14159 [3.14159]
6.02e23 [6.02e+23]
1e-12 [1e-12]
# hex numbers 100 [256]
FF [255]
# fractions 1/2 [0.5]
-3/4 [-0.75]
# mixed fractions 1 [1]
1/2 [0.5]
-3/4 [-0.75]
1-3/4 [1.75]
# uuid 12345678-1234-5678-1234-567812345678 [UUID('12345678-1234-5678-1234-567812345678')] """
""" Parse action for converting parsed integers to Python int """
""" Parse action for converting parsed numbers to Python float """
"""expression that parses an unsigned integer, returns an int"""
"""expression that parses a hexadecimal integer, returns an int"""
"""expression that parses an integer with optional leading sign, returns an int"""
"""fractional expression of an integer divided by an integer, returns a float"""
"""mixed integer of the form 'integer - fraction', with optional leading integer, returns float"""
"""expression that parses a floating point number and returns a float"""
"""expression that parses a floating point number with optional scientific notation and returns a float"""
# streamlining this expression makes the docs nicer-looking """any numeric expression, returns the corresponding Python type"""
"""any int or real number, returned as float"""
"""typical code identifier (leading alpha or '_', followed by 0 or more alphas, nums, or '_')"""
"IPv4 address (C{0.0.0.0 - 255.255.255.255})"
"IPv6 address (long, short, or mixed form)"
"MAC address xx:xx:xx:xx:xx (may also have '-' or '.' delimiters)"
""" Helper to create a parse action for converting parsed date string to Python datetime.date
Params - - fmt - format to be passed to datetime.strptime (default=C{"%Y-%m-%d"})
Example:: date_expr = pyparsing_common.iso8601_date.copy() date_expr.setParseAction(pyparsing_common.convertToDate()) print(date_expr.parseString("1999-12-31")) prints:: [datetime.date(1999, 12, 31)] """ def cvt_fn(s,l,t): try: return datetime.strptime(t[0], fmt).date() except ValueError as ve: raise ParseException(s, l, str(ve)) return cvt_fn
""" Helper to create a parse action for converting parsed datetime string to Python datetime.datetime
Params - - fmt - format to be passed to datetime.strptime (default=C{"%Y-%m-%dT%H:%M:%S.%f"})
Example:: dt_expr = pyparsing_common.iso8601_datetime.copy() dt_expr.setParseAction(pyparsing_common.convertToDatetime()) print(dt_expr.parseString("1999-12-31T23:59:59.999")) prints:: [datetime.datetime(1999, 12, 31, 23, 59, 59, 999000)] """ def cvt_fn(s,l,t): try: return datetime.strptime(t[0], fmt) except ValueError as ve: raise ParseException(s, l, str(ve)) return cvt_fn
"ISO8601 date (C{yyyy-mm-dd})"
"ISO8601 datetime (C{yyyy-mm-ddThh:mm:ss.s(Z|+-00:00)}) - trailing seconds, milliseconds, and timezone optional; accepts separating C{'T'} or C{' '}"
"UUID (C{xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx})"
def stripHTMLTags(s, l, tokens): """ Parse action to remove HTML tags from web page HTML source
Example:: # strip HTML links from normal text text = '<td>More info at the <a href="http://pyparsing.wikispaces.com">pyparsing</a> wiki page</td>' td,td_end = makeHTMLTags("TD") table_text = td + SkipTo(td_end).setParseAction(pyparsing_common.stripHTMLTags)("body") + td_end
print(table_text.parseString(text).body) # -> 'More info at the pyparsing wiki page' """ return pyparsing_common._html_stripper.transformString(tokens[0])
+ Optional( White(" \t") ) ) ).streamline().setName("commaItem") """Predefined expression of 1 or more printable words or quoted strings, separated by commas."""
"""Parse action to convert tokens to upper case."""
if __name__ == "__main__":
selectToken = CaselessLiteral("select") fromToken = CaselessLiteral("from")
ident = Word(alphas, alphanums + "_$")
columnName = delimitedList(ident, ".", combine=True).setParseAction(upcaseTokens) columnNameList = Group(delimitedList(columnName)).setName("columns") columnSpec = ('*' | columnNameList)
tableName = delimitedList(ident, ".", combine=True).setParseAction(upcaseTokens) tableNameList = Group(delimitedList(tableName)).setName("tables")
simpleSQL = selectToken("command") + columnSpec("columns") + fromToken + tableNameList("tables")
# demo runTests method, including embedded comments in test string simpleSQL.runTests(""" # '*' as column list and dotted table name select * from SYS.XYZZY
# caseless match on "SELECT", and casts back to "select" SELECT * from XYZZY, ABC
# list of column names, and mixed case SELECT keyword Select AA,BB,CC from Sys.dual
# multiple tables Select A, B, C from Sys.dual, Table2
# invalid SELECT keyword - should fail Xelect A, B, C from Sys.dual
# incomplete command - should fail Select
# invalid column name - should fail Select ^^^ frox Sys.dual
""")
pyparsing_common.number.runTests(""" 100 -100 +100 3.14159 6.02e23 1e-12 """)
# any int or real number, returned as float pyparsing_common.fnumber.runTests(""" 100 -100 +100 3.14159 6.02e23 1e-12 """)
pyparsing_common.hex_integer.runTests(""" 100 FF """)
import uuid pyparsing_common.uuid.setParseAction(tokenMap(uuid.UUID)) pyparsing_common.uuid.runTests(""" 12345678-1234-5678-1234-567812345678 """) |