explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-10-03 02:17:00 +03:00
Code Issues Packages Projects Releases Wiki Activity

Add logging for infinite loop

Browse Source
This commit is contained in:
Matthew Honnibal 2020-10-11 19:35:57 +00:00
parent d123e1ec8f
commit 605f0618c4
1 changed files with 15 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
spacy/pipeline/_parser_internals/nonproj.pyx
View File

@ -4,6 +4,7 @@ for doing pseudo-projective parsing implementation uses the HEAD decoration
scheme. scheme.
""" """
from copy import copy from copy import copy
import json
from ...tokens.doc cimport Doc, set_children_from_heads from ...tokens.doc cimport Doc, set_children_from_heads
@ -112,13 +113,18 @@ cpdef deprojectivize(Doc doc):
# Reattach arcs with decorated labels (following HEAD scheme). For each # Reattach arcs with decorated labels (following HEAD scheme). For each
# decorated arc X||Y, search top-down, left-to-right, breadth-first until # decorated arc X||Y, search top-down, left-to-right, breadth-first until
# hitting a Y then make this the new head. # hitting a Y then make this the new head.
new_heads = []
new_labels = []
for i in range(doc.length): for i in range(doc.length):
label = doc.vocab.strings[doc.c[i].dep] label = doc.vocab.strings[doc.c[i].dep]
if DELIMITER in label: if DELIMITER in label:
new_label, head_label = label.split(DELIMITER) new_label, head_label = label.split(DELIMITER)
new_head = _find_new_head(doc[i], head_label) new_head = _find_new_head(doc[i], head_label)
doc.c[i].head = new_head.i - i new_heads.append(new_head.i - i)
doc.c[i].dep = doc.vocab.strings.add(new_label) new_labels.append(doc.vocab.strings.add(new_label))
for i, (head, dep) in enumerate(zip(new_heads, new_labels)):
doc.c[i].head = head
doc.c[i].dep = dep
set_children_from_heads(doc.c, 0, doc.length) set_children_from_heads(doc.c, 0, doc.length)
return doc return doc
@ -166,13 +172,15 @@ def _find_new_head(token, headlabel):
# if there is none, return the current head (no change) # if there is none, return the current head (no change)
queue = [token.head] queue = [token.head]
n_iter = 0 n_iter = 0
headlabel = token.vocab.strings.as_int(headlabel)
heads = token.doc.to_array(["HEAD"]).astype("int64")
labels = [w.dep_ for w in token.doc]
while queue: while queue:
n_iter += 1 n_iter += 1
if n_iter >= len(token.doc): if n_iter >= len(token.doc):
print("Infinite loop") texts = [w.text for w in token.doc]
print([(w.i, w.text, w.head.i, w.dep_) for w in token.doc]) print(json.dumps(list(zip(range(len(token.doc)), texts, heads, labels)), indent=2))
with open("/tmp/doc.bytes", "wb") as file_: raise ValueError("Infinite loop?")
file_.write(token.doc.to_bytes())
next_queue = [] next_queue = []
for qtoken in queue: for qtoken in queue:
for child in qtoken.children: for child in qtoken.children:
@ -180,7 +188,7 @@ def _find_new_head(token, headlabel):
continue continue
if child == token: if child == token:
continue continue
if child.dep_ == headlabel: if child.dep == headlabel:
return child return child
next_queue.append(child) next_queue.append(child)
queue = next_queue queue = next_queue