def test_interpret(amr_model):
t = codec.parse('(a / A)')
assert interpret(t) == Graph([('a', ':instance', 'A')], top='a')
t = codec.parse('(a / A :consist-of (b / B))')
assert interpret(t) == Graph([('a', ':instance', 'A'),
('b', ':consist', 'a'),
('b', ':instance', 'B')],
top='a')
assert interpret(t, model=amr_model) == Graph([('a', ':instance', 'A'),
('a', ':consist-of', 'b'),
('b', ':instance', 'B')],
top='a')
def reify_edges(g: Graph, model: Model) -> Graph:
"""
Reify all edges in *g* that have reifications in *model*.
Args:
g: a :class:`~penman.graph.Graph` object
model: a model defining reifications
Returns:
A new :class:`~penman.graph.Graph` object with reified edges.
Example:
>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import reify_edges
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode('(c / chapter :mod 7)')
>>> g = reify_edges(g, model)
>>> print(codec.encode(g))
(c / chapter
:ARG1-of (_ / have-mod-91
:ARG2 7))
"""
vars = g.variables()
if model is None:
model = Model()
new_epidata = dict(g.epidata)
new_triples: List[BasicTriple] = []
for triple in g.triples:
if model.is_role_reifiable(triple[1]):
in_triple, node_triple, out_triple = model.reify(triple, vars)
if appears_inverted(g, triple):
in_triple, out_triple = out_triple, in_triple
new_triples.extend((in_triple, node_triple, out_triple))
var = node_triple[0]
vars.add(var)
# manage epigraphical markers
new_epidata[in_triple] = [Push(var)]
old_epis = new_epidata.pop(triple) if triple in new_epidata else []
node_epis, out_epis = _edge_markers(old_epis)
new_epidata[node_triple] = node_epis
new_epidata[out_triple] = out_epis
# we don't know where to put the final POP without configuring
# the tree; maybe this should be a tree operation?
else:
new_triples.append(triple)
g = Graph(new_triples,
epidata=new_epidata,
metadata=g.metadata)
logger.info('Reified edges: %s', g)
return g
def reify_attributes(g: Graph) -> Graph:
"""
Reify all attributes in *g*.
Args:
g: a :class:`~penman.graph.Graph` object
Returns:
A new :class:`~penman.graph.Graph` object with reified
attributes.
Example:
>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import reify_attributes
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode('(c / chapter :mod 7)')
>>> g = reify_attributes(g)
>>> print(codec.encode(g))
(c / chapter
:mod (_ / 7))
"""
variables = g.variables()
new_epidata = dict(g.epidata)
new_triples: List[BasicTriple] = []
i = 2
for triple in g.triples:
source, role, target = triple
if role != CONCEPT_ROLE and target not in variables:
# get unique var for new node
var = '_'
while var in variables:
var = f'_{i}'
i += 1
variables.add(var)
role_triple = (source, role, var)
node_triple = (var, CONCEPT_ROLE, target)
new_triples.extend((role_triple, node_triple))
# manage epigraphical markers
old_epis = new_epidata.pop(triple) if triple in new_epidata else []
role_epis, node_epis = _attr_markers(old_epis)
new_epidata[role_triple] = role_epis + [Push(var)]
new_epidata[node_triple] = node_epis + [POP]
else:
new_triples.append(triple)
g = Graph(new_triples,
epidata=new_epidata,
metadata=g.metadata)
logger.info('Reified attributes: %s', g)
return g
def test_errors(self, mini_amr):
m = Model()
a = Model.from_dict(mini_amr)
# basic roles
g = Graph([('a', ':instance', 'alpha')])
assert m.errors(g) == {}
g = Graph([('a', ':instance', 'alpha'), ('a', ':mod', '1')])
assert m.errors(g) == {('a', ':mod', '1'): ['invalid role']}
assert a.errors(g) == {}
# regex role names
g = Graph([('n', ':instance', 'name'), ('n', ':op1', 'Foo'),
('n', ':op2', 'Bar')])
assert a.errors(g) == {}
# disconnected graph
g = Graph([('a', ':instance', 'alpha'), ('b', ':instance', 'beta')])
assert m.errors(g) == {('b', ':instance', 'beta'): ['unreachable']}
assert a.errors(g) == {('b', ':instance', 'beta'): ['unreachable']}
def test_issue_90():
# https://github.com/goodmami/penman/issues/90
g = Graph([('i', ':instance', 'iota'), ('i2', ':instance', 'i'),
('i', ':ARG0', 'i2')],
top='i')
assert reconfigure(g) == Tree(('i', [('/', 'iota'),
(':ARG0', ('i2', [('/', 'i')]))]))
def indicate_branches(g: Graph, model: Model) -> Graph:
"""
Insert TOP triples in *g* indicating the tree structure.
Note:
This depends on *g* containing the epigraphical layout markers
from parsing; it will not work with programmatically
constructed Graph objects or those whose epigraphical data
were removed.
Args:
g: a :class:`~penman.graph.Graph` object
model: a model defining the TOP role
Returns:
A new :class:`~penman.graph.Graph` object with TOP roles
indicating tree branches.
Example:
>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import indicate_branches
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode('''
... (w / want-01
... :ARG0 (b / boy)
... :ARG1 (g / go-02
... :ARG0 b))''')
>>> g = indicate_branches(g, model)
>>> print(codec.encode(g))
(w / want-01
:TOP b
:ARG0 (b / boy)
:TOP g
:ARG1 (g / go-02
:ARG0 b))
"""
new_triples: List[BasicTriple] = []
for t in g.triples:
push = next((epi for epi in g.epidata.get(t, [])
if isinstance(epi, Push)),
None)
if push is not None:
if push.variable == t[2]:
new_triples.append((t[0], model.top_role, t[2]))
elif push.variable == t[0]:
assert isinstance(t[2], str)
new_triples.append((t[2], model.top_role, t[0]))
new_triples.append(t)
g = Graph(new_triples,
epidata=g.epidata,
metadata=g.metadata)
logger.info('Indicated branches: %s', g)
return g
def node_contexts(g: Graph) -> List[Union[Variable, None]]:
"""
Return the list of node contexts corresponding to triples in *g*.
If a node context is unknown, the value ``None`` is substituted.
Example:
>>> from penman import decode, layout
>>> g = decode('''
... (a / alpha
... :attr val
... :ARG0 (b / beta :ARG0 (g / gamma))
... :ARG0-of g)''')
>>> for ctx, trp in zip(layout.node_contexts(g), trp):
... print(ctx, ':', trp)
...
a : ('a', ':instance', 'alpha')
a : ('a', ':attr', 'val')
a : ('a', ':ARG0', 'b')
b : ('b', ':instance', 'beta')
b : ('b', ':ARG0', 'g')
g : ('g', ':instance', 'gamma')
a : ('g', ':ARG0', 'a')
"""
variables = g.variables()
stack = [g.top]
contexts: List[Union[Variable, None]] = [None] * len(g.triples)
for i, triple in enumerate(g.triples):
eligible: List[Variable] = [triple[0]]
if triple[1] != CONCEPT_ROLE and triple[2] in variables:
eligible.append(cast(Variable, triple[2]))
if stack[-1] not in eligible:
break
else:
contexts[i] = stack[-1]
pushed = get_pushed_variable(g, triple)
if pushed:
stack.append(pushed)
try:
for epi in g.epidata[triple]:
if epi is POP:
stack.pop()
except IndexError:
break # more POPs than contexts in stack
return contexts
def interpret(t: Tree, model: Model = None) -> Graph:
"""
Interpret tree *t* as a graph using *model*.
Tree interpretation is the process of transforming the nodes and
edges of a tree into a directed graph. A semantic model determines
which edges are inverted and how to deinvert them. If *model* is
not provided, the default model will be used.
Args:
t: the :class:`~penman.tree.Tree` to interpret
model: the :class:`~penman.model.Model` used to interpret *t*
Returns:
The interpreted :class:`~penman.graph.Graph`.
Example:
>>> from penman.tree import Tree
>>> from penman import layout
>>> t = Tree(
... ('b', [
... ('/', 'bark-01'),
... ('ARG0', ('d', [
... ('/', 'dog')]))]))
>>> g = layout.interpret(t)
>>> for triple in g.triples:
... print(triple)
...
('b', ':instance', 'bark-01')
('b', ':ARG0', 'd')
('d', ':instance', 'dog')
"""
if model is None:
model = _default_model
variables = {v for v, _ in t.nodes()}
top, triples, epidata = _interpret_node(t.node, variables, model)
epimap = {}
for triple, epis in epidata:
if triple in epimap:
logger.warning(
f'ignoring epigraph data for duplicate triple: {triple}'
)
else:
epimap[triple] = epis
g = Graph(triples, top=top, epidata=epimap, metadata=t.metadata)
logger.info('Interpreted: %s', g)
return g
def dereify_edges(g: Graph, model: Model) -> Graph:
"""
Dereify edges in *g* that have reifications in *model*.
Args:
g: a :class:`~penman.graph.Graph` object
Returns:
A new :class:`~penman.graph.Graph` object with dereified
edges.
Example:
>>> from penman.codec import PENMANCodec
>>> from penman.models.amr import model
>>> from penman.transform import dereify_edges
>>> codec = PENMANCodec(model=model)
>>> g = codec.decode(
... '(c / chapter'
... ' :ARG1-of (_ / have-mod-91'
... ' :ARG2 7))')
>>> g = dereify_edges(g, model)
>>> print(codec.encode(g))
(c / chapter
:mod 7)
"""
if model is None:
model = Model()
agenda = _dereify_agenda(g, model)
new_epidata = dict(g.epidata)
new_triples: List[BasicTriple] = []
for triple in g.triples:
var = triple[0]
if var in agenda:
first, dereified, epidata = agenda[var]
# only insert at the first triple so the dereification
# appears in the correct location
if triple == first:
new_triples.append(dereified)
new_epidata[dereified] = epidata
if triple in new_epidata:
del new_epidata[triple]
else:
new_triples.append(triple)
g = Graph(new_triples,
epidata=new_epidata,
metadata=g.metadata)
logger.info('Dereified edges: %s', g)
return g
def appears_inverted(g: Graph, triple: BasicTriple) -> bool:
"""
Return ``True`` if *triple* appears inverted in serialization.
More specifically, this function returns ``True`` if *triple* has
a :class:`Push` epigraphical marker in graph *g* whose associated
variable is the source variable of *triple*. This should be
accurate when testing a triple in a graph interpreted using
:func:`interpret` (including :meth:`PENMANCodec.decode
<penman.codec.PENMANCodec.decode>`, etc.), but it does not
guarantee that a new serialization of *g* will express *triple* as
inverted as it can change if the graph or its epigraphical markers
are modified, if a new top is chosen, etc.
Args:
g: a :class:`~penman.graph.Graph` containing *triple*
triple: the triple that does or does not appear inverted
Returns:
``True`` if *triple* appears inverted in graph *g*.
"""
variables = g.variables()
if triple[1] == CONCEPT_ROLE or triple[2] not in variables:
# attributes and instance triples should never be inverted
return False
else:
# edges may appear inverted...
variable = get_pushed_variable(g, triple)
if variable is not None:
# ... when their source is pushed
return variable == triple[0]
else:
# ... or when their target is the current node context
for variable, _triple in zip(node_contexts(g), g.triples):
if variable is None:
break # we can no longer guess the node context
elif _triple == triple:
return triple[2] == variable
return False
def deserialize(self, gstring):
node_stack = [] # list of previously instantiated nodes
node_depth = 0 # number of left parens encountered and not destacked
triple = []
# Tokenize and some logging (the system has a lot of unbalanced parentheses)
tokens = self.graph_tokenize(gstring)
# left_parens = tokens.count('(')
# right_parens = tokens.count(')')
# if left_parens != right_parens:
# logger.warning('gid=%s has %d left parens and %d right parens' % (self.gid, left_parens, right_parens))
# Loop through all tokens and parse the string
for tnum, token in enumerate(tokens):
#### Big case statement to classify parts of the graph string ####
ttype = self.token_type(token)
# Mostly ignored but can be used for error checking
if token == '(':
node_depth += 1
# Find the source for the triple
elif len(triple) == 0 and ttype == TType.concept:
# This path should only happen for a new graph. Make a somewhat arbitrary choice to
# either stop parsing or to clear out the existing triples to prevent disconnected graphs.
if len(self.triples) > 0:
logger.error('gid=%s Initial node constructed when triples not empty.' % (self.gid))
if len(self.triples) > len(tokens)/4: # if > half done (on average ~2 tokens per triple)
break
else:
self.triples = []
variable, concept, is_new_node = self.get_var_concept(token)
triple.append(variable)
if is_new_node:
node_stack.append( variable )
# Some error logging
if is_new_node and tokens[tnum-1] != '(':
logger.warning('gid=%s Missing starting paren for node %s/%s' % (self.gid, variable, concept))
if not is_new_node and tokens[tnum-1] == '(':
logger.warning('gid=%s Start paren present but %s is not a new concept' % (self.gid, concept))
elif len(triple) == 0 and ttype == TType.role:
variable = node_stack[-1]
triple.append(variable)
triple.append(token)
# Look for the role (aka edge)
elif len(triple) == 1 and ttype == TType.role:
triple.append(token)
# Look for the target
elif len(triple) == 2 and ttype == TType.attrib:
triple.append(token)
elif len(triple) == 2 and ttype == TType.concept:
variable, concept, is_new_node = self.get_var_concept(token)
if is_new_node:
node_stack.append( variable )
# Some error logging
if is_new_node and tokens[tnum-1] != '(':
logger.warning('gid=%s Missing starting paren for node %s/%s' % (self.gid, variable, concept))
if not is_new_node and tokens[tnum-1] == '(':
logger.warning('gid=%s Start paren present but %s is not a new concept' % (self.gid, concept))
triple.append(variable)
# De-stack the root nodes based on closing parens, but don't destack past the top var
# Log an error if we're trying to empty the stack and it's not the very last token
elif token == ')':
if len(node_stack) > 1:
node_stack.pop()
node_depth -= 1
elif tnum < len(self.triples)-1:
logger.warning('gid=%s Trying to destack past top node' % self.gid)
# Unknown situation (should never get here)
else:
logger.warning('gid=%s Unhandled token %s' % (self.gid, token))
#### Save the triple if complete ####
if len(triple) == 3:
self.triples.append( tuple(triple) )
triple = []
# Do a little post-processing check on the triples and fix attribs if needed
# I haven't found instances that requires this but it could be useful
for i, triple in enumerate(self.triples):
if triple[1] == self.INSTANCE:
continue
target = triple[2]
# Check if this is a varible
if self.re_var.fullmatch(target) or self.re_ii.fullmatch(target):
continue
# If it's an attrib enforce attribute syntax
else:
if (target.startswith('"') and target.endswith('"')) or self.is_num(target) or \
(target in set(['-', '+', 'interrogative', 'imperative', 'expressive'])):
continue
else:
new_target = '"' + target.replace('"', '') + '"'
self.triples[i] = tuple([triple[0], triple[1], new_target])
logger.warning('gid=%s Replacing attrib %s with %s' % (self.gid, target, new_target))
# Now convert to a penman graph and then back to a string
pgraph = Graph(self.triples)
# Catch malformed graphs, including disconnected ones, incorrectly quoted attibs, etc..
try:
self.gstring = penman.encode(pgraph, indent=6, model=NoOpModel())
self.pgraph = penman.decode(self.gstring, model=NoOpModel())
except:
self.gstring = None
self.pgraph = None