def adapt(markup):
spans = list(markup.spans)
spans = list(split_overlapping_spans(spans))
spans = list(
strip_spans(spans, markup.text, QUOTES + BRACKETS + DASHES + SPACES))
spans = list(filter_empty_spans(spans))
spans = list(adapt_spans(spans, markup.text, TYPES))
return Markup(markup.text, spans)
def adapt(markup):
# Чувашской Республики".
# ----------------------
# год Чарльза Дарвина»
# ----------------
spans = list(strip_spans(markup.spans, markup.text, QUOTES + DOT + SPACES))
spans = list(adapt_spans(spans, markup.text, TYPES))
return Markup(markup.text, spans)
def adapt(markup):
# extra spaces + dots in spans
# News Corp .
# -----------
# « Русал »
# ---------
spans = strip_spans(markup.spans, markup.text, DOT + SPACES)
spans = strip_spans_bounds(spans, markup.text, QUOTES + SPACES)
spans = adapt_spans(spans, markup.text, TYPES)
return Markup(markup.text, list(spans))
def adapt_spans(spans, text, types):
spans = select_type_spans(spans, types)
spans = convert_span_types(spans, types)
# in mitie and sometimes in deeppavlov
spans = list(strip_spans(spans, text, QUOTES))
# ne5 typos is span.stop
# Magna Internationa -> Magna International
# Горсове -> Горсовет
# tokenizer errors
# поезд Москва-Баку
# Yahoo!.
tokens = list(tokenize(text))
spans = list(filter_misaligned_spans(spans, tokens))
# ne5 bug
# Бражский район Подмосковья
# --------------
# -----------------
spans = list(filter_overlapping(spans))
return spans
def adapt(markup):
spans = strip_spans(markup.spans, markup.text, QUOTES + SPACES + DOT)
spans = adapt_spans(spans, markup.text, TYPES)
return Markup(markup.text, list(spans))
def adapt(markup):
spans = list(strip_spans(markup.spans, markup.text, QUOTES + SPACES))
spans = list(adapt_spans(spans, markup.text, TYPES))
return Markup(markup.text, spans)
def adapt_spans(spans, text):
spans = list(adapt_overlapping_spans(spans, text))
spans = list(strip_spans(spans, text, QUOTES + BRACKETS + DASHES))
spans = list(filter_empty_spans(spans))
return adapt_spans_(list(spans), text, TYPES)
def adapt_overlapping_spans(spans, text):
spans = split_overlapping_spans(spans)
spans = strip_spans(spans, text, SPACES)
return filter_empty_spans(spans)