Class Pattern

pattern/text/search.py:652–864

GREEDY = "greedy"

class Pattern(object):
    
    def __init__(self, sequence=[], *args, **kwargs):
        """ A sequence of constraints that matches certain phrases in a sentence.
            The given list of Constraint objects can contain nested lists (groups).
        """
        # Parse nested lists and tuples from the sequence into groups.
        # [DT [JJ NN]] => Match.group(1) will yield the JJ NN sequences.
        def _ungroup(sequence, groups=None):
            for v in sequence:
                if isinstance(v, (list, tuple)):
                    if groups is not None:
                        groups.append(list(_ungroup(v, groups=None)))
                    for v in _ungroup(v, groups):
                        yield v
                else: 
                    yield v
        self.groups = []
        self.sequence = list(_ungroup(sequence, groups=self.groups))
        # Assign Constraint.index:
        i = 0
        for constraint in self.sequence:
            constraint.index = i; i+=1
        # There are two search modes: STRICT and GREEDY.
        # - In STRICT, "rabbit" matches only the string "rabbit".
        # - In GREEDY, "rabbit|NN" matches the string "rabbit" tagged "NN".
        # - In GREEDY, "rabbit" matches "the big white rabbit" (the entire chunk is a match).
        # - Pattern.greedy(chunk, constraint) determines (True/False) if a chunk is a match.
        self.strict = kwargs.get("strict", STRICT in args and not GREEDY in args)
        self.greedy = kwargs.get("greedy", lambda chunk, constraint: True)
        
    def __iter__(self):
        return iter(self.sequence)
    def __len__(self):
        return len(self.sequence)
    def __getitem__(self, i):
        return self.sequence[i]
        
    @classmethod
    def fromstring(cls, s, *args, **kwargs):
        """ Returns a new Pattern from the given string.
            Constraints are separated by a space.
            If a constraint contains a space, it must be wrapped in [].
        """
        s = s.replace("\(", "&lparen;")
        s = s.replace("\)", "&rparen;")
        s = s.replace("\[", "[")
        s = s.replace("\]", "]")
        s = s.replace("\{", "&lcurly;")
        s = s.replace("\}", "&rcurly;")
        p = []
        i = 0
        for m in re.finditer(r"\[.*?\]|\(.*?\)", s):
            # Spaces in a range encapsulated in square brackets are encoded.
            # "[Windows Vista]" is one range, don't split on space.
            p.append(s[i:m.start()])
            p.append(s[m.start():m.end()].replace(" ", "&space;")); i=m.end()
        p.append(s[i:])