def get_sentences(self, kanji, items_per_page, page_num, allowed_kanji = None):
"""
Implementation of get_sentences.
"""
if len(kanji) != 1 or type(kanji) != unicode or not Restructurer.is_kanji(kanji):
raise ValueError("Invalid value for 'kanji' parameter")
# add the kanji to search to the list of allowed kanji
if not allowed_kanji is None:
allowed_kanji += kanji
try:
query_ids = self.cache[kanji]
except KeyError:
query = [ (k, len(v[0])) for (k, v) in self.sentences.items() \
if kanji in v[0] \
and v[1] != '' \
and ( allowed_kanji is None \
or self.all_allowed(v[0], allowed_kanji) ) ]
# with v[1] != '' we filter bad data (TODO: sanitize bad data)
query.sort(key = lambda x: x[1]) # sort by second argument, length of the sentence
query_ids = [ t[0] for t in query ] # keep only the ids
self.cache[kanji] = query_ids
results = [ self.sentences[id] for id in query_ids ]
init_idx = page_num * items_per_page
return results[init_idx : init_idx + items_per_page]
def all_allowed(self, sentence, allowed):
return all((c in allowed for c in sentence if Restructurer.is_kanji(c)))
def KanjiIterator(string):
for s in string:
if Restructurer.is_kanji(s):
yield s
