def print_plaintext(text):
for pair in split_furigana(text):
if len(pair)==2:
kanji,hira = pair
print("%s(%s)" % (kanja,hira), end='')
else:
print(pair[0], end='')
print('')
def convert_text_to_furigana_html(text):
for pair in split_furigana(text):
if len(pair) == 2:
kanji, hira = pair
yield ("<ruby><rb>{0}</rb><rt>{1}</rt></ruby>".format(kanji, hira))
else:
yield pair[0]
yield ''
def get_furi(s):
temp=[]
for x in s.split(" "):
try:
temp.append(conv_final(split_furigana(x)))
except:
temp.append(x)
return " ".join(temp)
def return_html(text):
output = ""
for pair in split_furigana(text):
if len(pair)==2:
kanji,hira = pair
output += "<ruby><rb>{0}</rb><rt>{1}</rt></ruby>".
format(kanji, hira)
else:
output += pair[0]
def print_html(text):
for pair in split_furigana(text):
if len(pair)==2:
kanji,hira = pair
print("<ruby><rb>{0}</rb><rt>{1}</rt></ruby>".
format(kanji, hira), end='')
else:
print(pair[0], end='')
print('')
def pronounce(text: str) -> str:
result = []
for pair in split_furigana(text):
if len(pair) == 2:
kanji, hira = pair
result.append(f"<ruby><rb>{kanji}</rb><rt>{hira}</rt></ruby>")
else:
result.append(pair[0])
return "".join(result)
def get_hira_kata(s):
S=""
for t in s.split(" "):
try:
temp=split_furigana(t)
except:
return ""
for x in temp:
if len(x)==1:
S=S+x[0]
else:
S=S+x[1]
S=S+" "
return " ".join([x for x in re.sub(u"([\u3040-\u30ff\u3400-\u4dbf\u4e00-\u9fff\uf900-\ufaff\uff66-\uff9f-\u3300-\u33ff\ufe30-\ufe4f\uf900-\ufaff\U0002f800-\U0002fa1f\u30a0-\u30ff\u2e80-\u2eff\u4e00-\u9fff\u3400-\u4dbf\U00020000-\U0002a6df\U0002a700-\U0002b73f\U0002b740-\U0002b81f\U0002b820-\U0002ceaf]+)", r" \1 ", S).strip().capitalize().split(" ") if x!=""])
def add_furigana(japanese_string):
"""Add furigana to all kanji in a string."""
from furigana.furigana import split_furigana
ruby_markup = ""
# TODO: make the index access sane
for ch in japanese_string:
try:
pair = split_furigana(ch)
except:
pair = [(ch, ), (ch, )]
if len(pair[0]) == 2:
kanji, hira = pair[1]
ruby_markup += "<ruby><rb>{}</rb><rt>{}</rt></ruby>".format(
kanji, hira)
else:
ruby_markup += ("{}".format(pair[0][0]))
return ruby_markup
def tokenize(text):
tokens = map(list, split_furigana(text))
return tokens
