Source code for penman.graph

# -*- coding: utf-8 -*-

"""
Data structures for Penman graphs and triples.
"""

from typing import (Union, Optional, Mapping, List, Dict, Set, NamedTuple)
from collections import defaultdict
import copy

from penman.exceptions import GraphError
from penman.types import (
    Variable,
    Constant,
    Role,
    Target,
    BasicTriple,
    Triples,
)
from penman.epigraph import Epidata


CONCEPT_ROLE = ':instance'


[docs]class Triple(NamedTuple): """ A relation between nodes or between a node and an constant. Args: source: the source variable of the triple role: the edge label between the source and target target: the target variable or constant """ source: Variable """The source variable of the triple.""" role: Role """The edge label between the source and target.""" target: Target """The target variable or constant."""
[docs]class Instance(Triple): """A relation indicating the concept of a node.""" target: Constant """The node concept."""
[docs]class Edge(Triple): """A relation between nodes.""" target: Variable """The target variable."""
[docs]class Attribute(Triple): """A relation between a node and a constant.""" target: Constant """The target constant."""
[docs]class Graph(object): """ A basic class for modeling a rooted, directed acyclic graph. A Graph is defined by a list of triples, which can be divided into two parts: a list of graph edges where both the source and target are variables (node identifiers), and a list of node attributes where only the source is a variable and the target is a constant. The raw triples are available via the :attr:`triples` attribute, while the :meth:`instances`, :meth:`edges` and :meth:`attributes` methods return only those that are concept relations, relations between nodes, or relations between a node and a constant, respectively. Args: triples: an iterable of triples (:class:`Triple` or 3-tuples) top: the variable of the top node; if unspecified, the source of the first triple is used epidata: a mapping of triples to epigraphical markers metadata: a mapping of metadata types to descriptions Example: >>> from penman.graph import Graph >>> Graph([('b', ':instance', 'bark-01'), ... ('d', ':instance', 'dog'), ... ('b', ':ARG0', 'd')]) <Graph object (top=b) at ...> """ def __init__(self, triples: Triples = None, top: Variable = None, epidata: Mapping[BasicTriple, Epidata] = None, metadata: Mapping[str, str] = None): if not triples: triples = [] if not epidata: epidata = {} if not metadata: metadata = {} # the following (a) creates a new list (b) validates that # they are triples, and (c) ensures roles begin with : self.triples = [(src, _ensure_colon(role), tgt) for src, role, tgt in triples] self._top = top self.epidata = dict(epidata) self.metadata = dict(metadata) def __repr__(self): name = self.__class__.__name__ return f'<{name} object (top={self.top}) at {id(self)}>' def __str__(self): triples = '[{}]'.format(',\n '.join(map(repr, self.triples))) epidata = '{{{}}}'.format(',\n '.join( map('{0[0]!r}: {0[1]!r}'.format, self.epidata.items()))) return f'Graph(\n {triples},\n epidata={epidata})' def __eq__(self, other): if not isinstance(other, Graph): return NotImplemented return (self.top == other.top and len(self.triples) == len(other.triples) and set(self.triples) == set(other.triples)) def __or__(self, other): if isinstance(other, Graph): g = copy.deepcopy(self) g.metadata.clear() g |= other return g else: return NotImplemented def __ior__(self, other): if isinstance(other, Graph): new = set(other.triples) - set(self.triples) self.triples.extend(t for t in other.triples if t in new) for t in new: if t in other.epidata: self.epidata[t] = list(other.epidata[t]) self.epidata.update(other.epidata) return self else: return NotImplemented def __sub__(self, other): if isinstance(other, Graph): g = copy.deepcopy(self) g.metadata.clear() g -= other return g else: return NotImplemented def __isub__(self, other): if isinstance(other, Graph): removed = set(other.triples) self.triples[:] = [t for t in self.triples if t not in removed] for t in removed: if t in self.epidata: del self.epidata[t] possible_variables = set(v for t in self.triples for v in t[::2]) if self._top not in possible_variables: self._top = None return self else: return NotImplemented @property def top(self) -> Union[Variable, None]: """ The top variable. """ top = self._top if top is None and len(self.triples) > 0: top = self.triples[0][0] # implicit top return top @top.setter def top(self, top: Union[Variable, None]): if top is not None and top not in self.variables(): raise GraphError('top must be a valid node') self._top = top # check if top is valid variable?
[docs] def variables(self) -> Set[Variable]: """ Return the set of variables (nonterminal node identifiers). """ vs = set(src for src, _, _ in self.triples) if self._top is not None: vs.add(self._top) return vs
[docs] def instances(self) -> List[Instance]: """ Return instances (concept triples). """ return [Instance(*t) for t in self._filter_triples(None, CONCEPT_ROLE, None)]
[docs] def edges(self, source: Optional[Variable] = None, role: Role = None, target: Variable = None) -> List[Edge]: """ Return edges filtered by their *source*, *role*, or *target*. Edges don't include terminal triples (concepts or attributes). """ variables = self.variables() return [Edge(*t) for t in self._filter_triples(source, role, target) if t[1] != CONCEPT_ROLE and t[2] in variables]
[docs] def attributes(self, source: Optional[Variable] = None, role: Role = None, target: Constant = None) -> List[Attribute]: """ Return attributes filtered by their *source*, *role*, or *target*. Attributes don't include concept triples or those where the target is a nonterminal. """ variables = self.variables() return [Attribute(*t) for t in self._filter_triples(source, role, target) if t[1] != CONCEPT_ROLE and t[2] not in variables]
def _filter_triples(self, source: Optional[Variable], role: Optional[Role], target: Optional[Target]) -> List[BasicTriple]: """ Filter triples based on their source, role, and/or target. """ if source is role is target is None: triples = list(self.triples) else: triples = [ t for t in self.triples if ((source is None or source == t[0]) and (role is None or role == t[1]) and (target is None or target == t[2])) ] return triples
[docs] def reentrancies(self) -> Dict[Variable, int]: """ Return a mapping of variables to their re-entrancy count. A re-entrancy is when more than one edge selects a node as its target. These graphs are rooted, so the top node always has an implicit entrancy. Only nodes with re-entrancies are reported, and the count is only for the entrant edges beyond the first. Also note that these counts are for the interpreted graph, not for the linearized form, so inverted edges are always re-entrant. """ entrancies: Dict[Variable, int] = defaultdict(int) if self.top is not None: entrancies[self.top] += 1 # implicit entrancy to top for t in self.edges(): entrancies[t.target] += 1 return dict((v, cnt - 1) for v, cnt in entrancies.items() if cnt >= 2)
def _ensure_colon(role): if not role.startswith(':'): return ':' + role return role