class ALCEtmSearcher(object):
def __init__(self):
self.db = WordDB("mydict")
self.n = NLTK()
self.url_unum = "http://home.alc.co.jp/db/owa/etm_sch?unum={unum}&stg=2"
def find_word_with_unum(self, word):
""" Aggressively find word data of ALC etmology dictionary with NLP
1. word itself
2. lemmatize
3. Porter stemmer with lemma
4. Porter stemmer with stem
5. Lancaster stemmer with stem
"""
word = word.lower()
lemma = self.n.lemmatizer(word)
p_stem = self.n.porter_stemmer(word)
l_stem = self.n.lancaster_stemmer(word)
word_data = (
self.db.find_with_unum(word, "lemma")
or self.db.find_with_unum(lemma, "lemma")
or self.db.find_with_unum(p_stem, "lemma")
or self.db.find_with_unum(p_stem, "stem")
or self.db.find_with_unum(l_stem, "stem")
)
return word_data
def text_linker(self, text, is_newtab=None):
# link_format = '<a href="{url}" target="_blank">{w}</a>'.format(
target = 'target="_blank"' if is_newtab else ""
link_format = '<a href="{url}" {target}>{w}</a>'.format(url=self.url_unum, w="{w}", target=target)
# Cache linked word not to look into database more hantwice
# {word: unum}
# NOTE: Should it be with MongoDB or something?
linked_word = {}
sentences = text.splitlines()
sentence_list = []
for sentence in sentences:
word_list = []
for word in self.n.tokenize(sentence):
# Filter stopwords
if len(word) < 3 or word.lower() in self.n.stopwords:
word_list.append(word)
continue
# Not to look into database twice with the same word
if word in linked_word:
word_list.append(link_format.format(unum=linked_word[word], w=word))
continue
# Try to find word in DataBase
word_data = self.find_word_with_unum(word.lower())
if word_data:
linked_word[word] = word_data["alc_etm"]["unum"]
word_list.append(link_format.format(unum=word_data["alc_etm"]["unum"], w=word))
else:
word_list.append(word)
sentence_list.append(" ".join(word_list))
return "<br>".join(sentence_list)
def __init__(self):
self.db = WordDB("mydict")
self.n = NLTK()
self.url_unum = "http://home.alc.co.jp/db/owa/etm_sch?unum={unum}&stg=2"
