def phrases(ccg):
for s in paren_utils.paren_iter(ccg,
bottom_up=True,
lparen='{',
rparen='}'):
if Readible.Phrase_RE.match(s):
yield Readible.Phrase(s)
def clean_parens(tag):
marked_tag = paren_utils.mark_depth(tag)
for t in paren_utils.paren_iter(f'({tag})', bottom_up=True):
depth = paren_utils.depth_at(tag, tag.index(t)) + 1
t = paren_utils.mark_depth(t)
j = 2
while f'<{j}>' in t:
t = t.replace(f'<{j}>', '(').replace(f'</{j}>', ')')
j += 1
mod = CCG_Tag._Modifier_RE.match(t)
if mod and CCG_Tag.remove_features(
mod.group('a')) == CCG_Tag.remove_features(mod.group('b')):
a = mod.group('a')
b = mod.group('b')
slash = mod.group('slash')
marked_tag = marked_tag.replace(
f'<{depth}>{a}</{depth}>{slash}<{depth}>{b}</{depth}>',
f'({a}){slash}({b})', 1)
elif CCG_Tag._Left_RE.search(t):
x = CCG_Tag._Left_RE.match(t)
a = x.group('a')
marked_tag = marked_tag.replace(f'<{depth}>{a}</{depth}>', a,
1)
elif CCG_Tag._Right_RE.search(t):
x = CCG_Tag._Right_RE.search(t)
a = x.group('a')
marked_tag = marked_tag.replace(f'<{depth}>{a}</{depth}>',
f'({a})', 1)
return paren_utils.unmark_depth(marked_tag)
def edge_triples(self):
if self.edges_are_memoized:
return self.edge_triples_memoized
memo = []
text = str(self)
idx = [(e.start(), e.group()) for e in self.EDGE_RE.finditer(text)]
for rel_position, rel in idx:
Target_RE = re.compile(f'{rel}\s*[(]?\s*{self.NODE_RE.pattern}')
for t in paren_utils.paren_iter(text):
pos = rel_position - text.index(t)
if t.startswith(rel, pos) and paren_utils.depth_at(t,
pos) == 0:
root = self.NODE_RE.match(t)
source = root.group().split('/')[0].strip()
x = Target_RE.match(t, pos=pos)
if x:
target = x.group('id')
memo.append((source, rel, target))
else:
print('Missing target node! ', rel,
re.sub('\s+', ' ', t))
memo.append((source, rel, '?'))
break
self.edge_triples_memoized = memo
self.edges_are_memoized = True
return memo
def latex(text):
amr = AMR(text)
text = str(amr)
for x in re.findall('x[0-9]+ ?/ ?[^()\s]+', text):
text = text.replace(x, '(' + x + ')')
edges = [(e, id) for e, id in zip(amr.edges(), amr.edge_ids())]
elems = []
max_depth = paren_utils.max_depth(text)
prev_depth = 0
depth = 0
i = 0
node_depth = {}
for t in paren_utils.paren_iter(text):
node = amr.NODE_RE.match(t).group()
id = node.split('/')[0].strip()
# clean node
if re.match('x[0-9]+/', node):
node = node.split('/')[1]
node = node.replace('"', '``', 1).replace('"', "''", 1)
prev_depth = depth
depth = paren_utils.depth_at(text, text.index(t))
if depth > prev_depth:
i = 0
node_depth[id] = depth
num_nodes = paren_utils.mark_depth(text).count(f'<{depth}>')
x = AMR_Latex.get_x(i, num_nodes)
y = AMR_Latex.get_y(depth, max_depth)
color = AMR_Latex.get_color(i)
elems.append(f'\t\\node[{color}]({id}) at ({x},{y}) {{{node}}};')
i += 1
for edge, id in edges:
source = id.split('_')[0]
target = id.split('_')[2]
dir1 = 'south'
dir2 = 'north'
if node_depth[source] > node_depth[target]:
dir1 = 'north'
dir2 = 'south'
if node_depth[source] == node_depth[target]:
dir1 = 'north'
dir2 = 'north'
elems.append(
f'\t\draw[->, thick] ({source}.{dir1}) -- ({target}.{dir2}) node[midway, above, sloped] {{{edge}}};'
)
latex = '\n\\begin{tikzpicture}[\n'
latex += 'red/.style={rectangle, draw=red!60, fill=red!5, very thick, minimum size=7mm},\n'
latex += 'blue/.style={rectangle, draw=blue!60, fill=blue!5, very thick, minimum size=7mm},\n'
latex += 'green/.style={rectangle, draw=green!60, fill=green!5, very thick, minimum size=7mm},\n'
latex += 'purple/.style={rectangle, draw=purple!60, fill=purple!5, very thick, minimum size=7mm},\n'
latex += 'orange/.style={rectangle, draw=orange!60, fill=orange!5, very thick, minimum size=7mm},\n'
latex += ']\n'
latex += '\n'.join(elems)
latex += '\n\end{tikzpicture}\n'
return latex
def named_entities(self):
NE_RE = re.compile(
f'(?P<root>{self.NODE_RE.pattern}).*:name\s+<1>(?P<name>.*?)</1>',
re.DOTALL)
for t in paren_utils.paren_iter(str(self)):
t = paren_utils.mark_depth(t)
x = NE_RE.match(t)
if x:
root = x.group('root')
name = x.group('name')
yield AMR(f'({root} :name ({name}) )')
def phrases(ccg):
ccg = ccg.replace('{', '-LBR-').replace('}', '-RBR-')
ccg = ccg.replace('(', '{').replace(')', '}')
for t in re.finditer('<[LT] (?P<tag>[^\s>]+) .*?>', ccg):
tag = t.group('tag')
ccg = ccg.replace(
f' {tag} ',
' ' + tag.replace('{', '(').replace('}', ')') + ' ')
for s in paren_utils.paren_iter(ccg, bottom_up=True):
p = CCGBank.Phrase_RE.match(s)
if not p: continue
yield CCGBank.Phrase(s)
def phrases_and_indices(ccg):
phrases = Readible.phrases(ccg)
ccg_ids = ccg
ID_RE = re.compile('[*](?P<n>[0-9]+)[*]')
for i, w in enumerate(Readible.words(ccg)):
word = w.text
ccg_ids = ccg_ids.replace(word, f'{{*{i}*}}', 1)
for s in paren_utils.paren_iter(ccg_ids,
bottom_up=True,
lparen='{',
rparen='}'):
if Readible.Phrase_RE.match(s):
indices = [i.group('n') for i in ID_RE.finditer(s)]
yield (next(phrases), f'{indices[0]}-{indices[-1]}')
def sub_amrs(self):
for t in paren_utils.paren_iter(self.text):
yield AMR('(' + t + ')')
