def strlib_unary_features(span):
"""
Structural-related features for a single span
"""
if not span.sentence.is_structural():
return
yield "TAG_" + get_tag(span), DEF_VALUE
for attr in get_attributes(span):
yield "HTML_ATTR_" + attr, DEF_VALUE
yield "PARENT_TAG_" + get_parent_tag(span), DEF_VALUE
prev_tags = get_prev_sibling_tags(span)
if len(prev_tags):
yield "PREV_SIB_TAG_" + prev_tags[-1], DEF_VALUE
yield "NODE_POS_" + str(len(prev_tags) + 1), DEF_VALUE
else:
yield "FIRST_NODE", DEF_VALUE
next_tags = get_next_sibling_tags(span)
if len(next_tags):
yield "NEXT_SIB_TAG_" + next_tags[0], DEF_VALUE
else:
yield "LAST_NODE", DEF_VALUE
yield "ANCESTOR_CLASS_[%s]" % " ".join(
get_ancestor_class_names(span)), DEF_VALUE
yield "ANCESTOR_TAG_[%s]" % " ".join(
get_ancestor_tag_names(span)), DEF_VALUE
yield "ANCESTOR_ID_[%s]" % " ".join(get_ancestor_id_names(span)), DEF_VALUE
def _strlib_unary_features(span: SpanMention) -> Iterator[Tuple[str, int]]:
"""Structural-related features for a single span."""
if not span.sentence.is_structural():
return
yield f"TAG_{get_tag(span)}", DEF_VALUE
for attr in get_attributes(span):
yield f"HTML_ATTR_{attr}", DEF_VALUE
yield f"PARENT_TAG_{get_parent_tag(span)}", DEF_VALUE
prev_tags = get_prev_sibling_tags(span)
if len(prev_tags):
yield f"PREV_SIB_TAG_{prev_tags[-1]}", DEF_VALUE
yield f"NODE_POS_{len(prev_tags) + 1}", DEF_VALUE
else:
yield "FIRST_NODE", DEF_VALUE
next_tags = get_next_sibling_tags(span)
if len(next_tags):
yield f"NEXT_SIB_TAG_{next_tags[0]}", DEF_VALUE
else:
yield "LAST_NODE", DEF_VALUE
yield f"ANCESTOR_CLASS_[{' '.join(get_ancestor_class_names(span))}]", DEF_VALUE
yield f"ANCESTOR_TAG_[{' '.join(get_ancestor_tag_names(span))}]", DEF_VALUE
yield f"ANCESTOR_ID_[{' '.join(get_ancestor_id_names(span))}]", DEF_VALUE