def collect_variants(text, term, replace="_TERM_"):
"""
This finds all spelling variants of term in text.
>>> text = "I had a Deja-vu, or Déjàvu")
>>> collect_variants(text, "Déjà Vu")
returns {"Deja-vu", "Déjàvu"}
Args:
text: str -- text in which to search for spelling variants
term: str
Returns:
set -- A set of all variants found.
*NB: This is used in the output JSON as an additional index within Wordnik #TODO
"""
squashed_term = squashed(term)
clean_text = unidecode(text).lower()
# This RE allows for up to one non-letter character between all letters
fuzzy_term = ''.join("{}[^a-z0-9]?".format(c) for c in squashed_term[:-1]) + squashed_term[-1]
term_re = r'\b({})s?\b'.format(fuzzy_term) # s? for plurals
collected = set()
for m in re.finditer(term_re, clean_text):
variant = text[m.start():m.end()]
if variant.lower().endswith("s") and not term.lower().endswith("s"):
variant = variant[:-1]
collected.add(variant)
# collected.add(term + 's') # account for finding plurals
# collected.add(term + 'es')
return collected
def collect_variants(text, term, replace="_TERM_"):
"""
This finds all spelling variants of term in text.
>>> text = "I had a Deja-vu, or Déjàvu")
>>> collect_variants(text, "Déjà Vu")
returns {"Deja-vu", "Déjàvu"}
Args:
text: str -- text in which to search for spelling variants
term: str
Returns:
set -- A set of all variants found.
*NB: This is used in the output JSON as an additional index within Wordnik #TODO
"""
squashed_term = squashed(term)
clean_text = unidecode(text).lower()
# This RE allows for up to one non-letter character between all letters
fuzzy_term = ''.join("{}[^a-z0-9]?".format(c)
for c in squashed_term[:-1]) + squashed_term[-1]
term_re = r'\b({})s?\b'.format(fuzzy_term) # s? for plurals
collected = set()
for m in re.finditer(term_re, clean_text):
variant = text[m.start():m.end()]
if variant.lower().endswith("s") and not term.lower().endswith("s"):
variant = variant[:-1]
collected.add(variant)
# collected.add(term + 's') # account for finding plurals
# collected.add(term + 'es')
return collected
def get_html_features(self, html):
"""Detects whether the search term exists is highlighted (bolded, emphasised) or
in quotes. Needs to be called after self.request_page.
"""
if not self.term:
return None
minimal_html = squashed(html, keep='<>/&;')
minimal_term = squashed(self.term)
highlight_re = r"<(em|i|b|strong|span)[^>]*> *{}[ ,:]*</\1>".format(minimal_term)
quote_re = r"<({})[^>]*> *{}[ ,:]*</({})>".format("|".join(self.OPENING_QUOTES), minimal_term, "|".join(self.CLOSING_QUOTES))
self.features = {
"highlighted": bool(re.search(highlight_re, minimal_html, re.IGNORECASE)),
"quotes": bool(re.search(quote_re, minimal_html, re.IGNORECASE)),
}
def get_html_features(self, html):
"""Detects whether the search term exists is highlighted (bolded, emphasised) or
in quotes. Needs to be called after self.request_page.
"""
if not self.term:
return None
minimal_html = squashed(html, keep='<>/&;')
minimal_term = squashed(self.term)
highlight_re = r"<(em|i|b|strong|span)[^>]*> *{}[ ,:]*</\1>".format(
minimal_term)
quote_re = r"<({})[^>]*> *{}[ ,:]*</({})>".format(
"|".join(self.OPENING_QUOTES), minimal_term,
"|".join(self.CLOSING_QUOTES))
self.features = {
"highlighted":
bool(re.search(highlight_re, minimal_html, re.IGNORECASE)),
"quotes":
bool(re.search(quote_re, minimal_html, re.IGNORECASE)),
}
def clean_and_qualify_wordlist(wordlist):
"""Generator that returns cleaned version of a list of words.
Will remove any non-words.
Args:
wordlist: list
Returns:
list
"""
cleaned = filter(bool, map(clean_and_qualify_term, wordlist))
cleaned_squashed = set()
for term in cleaned:
s = squashed(term)
if s not in cleaned_squashed:
cleaned_squashed.add(s)
yield term
def clean_and_qualify_wordlist(wordlist):
"""Generator that returns cleaned version of a list of words.
Will remove any non-words.
Args:
wordlist: list
Returns:
list
"""
cleaned = filter(bool, map(clean_and_qualify_term, wordlist))
cleaned_squashed = set()
for term in cleaned:
s = squashed(term)
if s not in cleaned_squashed:
cleaned_squashed.add(s)
yield term
