187 lines
6.8 KiB
Python
187 lines
6.8 KiB
Python
import json
|
||
import MeCab
|
||
import re
|
||
|
||
import requests
|
||
|
||
request_url = "http://127.0.0.1:19633"
|
||
request_timeout = 10
|
||
|
||
# Generate Hiragana and Katakana characters
|
||
hiragana = tuple(chr(i) for i in range(12353, 12436)) # \u3041 - \u3094
|
||
katakana = tuple(chr(i) for i in range(12450, 12532)) # \u30A2 - \u30F4
|
||
punctuation = tuple(chr(i) for i in range(65281, 65382))
|
||
|
||
# Combine both into one tuple
|
||
all_kana = hiragana + katakana + punctuation + (' ', 'ー', '・')
|
||
|
||
KANJI_RE = re.compile(r"[\u4E00-\u9FFF]")
|
||
|
||
def katakana_to_hiragana(text: str) -> str:
|
||
return "".join(
|
||
chr(ord(ch) - 0x60) if 0x30A1 <= ord(ch) <= 0x30F6 else ch
|
||
for ch in text
|
||
)
|
||
|
||
# def get_node_reading(node) -> str:
|
||
# features = node.feature.split(",")
|
||
# print()
|
||
# if len(features) >= 8 and features[7] != "*":
|
||
# return katakana_to_hiragana(features[7])
|
||
# return node.surface
|
||
|
||
def has_kanji(text: str) -> bool:
|
||
return bool(KANJI_RE.search(text))
|
||
|
||
def get_node_reading(node) -> str:
|
||
features = node.feature.split(",")
|
||
## These depend on what dictionary is being used.
|
||
if len(features) > 6 and features[6] != "*":
|
||
return katakana_to_hiragana(features[6])
|
||
return node.surface
|
||
|
||
def annotate_phrase(text: str) -> str:
|
||
tagger = MeCab.Tagger()
|
||
node = tagger.parseToNode(text)
|
||
vocab = ()
|
||
parts = []
|
||
while node:
|
||
surface = node.surface
|
||
if surface:
|
||
vocab += (surface,)
|
||
if has_kanji(surface):
|
||
reading = get_node_reading(node)
|
||
parts.append(f"{surface}({reading})")
|
||
else:
|
||
parts.append(surface)
|
||
node = node.next
|
||
|
||
return {'reading':("".join(parts)),'vocab':vocab}
|
||
|
||
def elide(text: str) -> str:
|
||
elide_max_length = 100
|
||
if len(text) > elide_max_length:
|
||
return text[:100] + "..."
|
||
return text
|
||
|
||
def anki_fields_term(term : str) -> dict:
|
||
params = {
|
||
"text": term,
|
||
"type": "term",
|
||
"markers": ["audio", "cloze-body-kana", "conjugation", "expression", "furigana", "furigana-plain", "glossary", "glossary-brief", "glossary-no-dictionary", "glossary-first", "glossary-first-brief", "glossary-first-no-dictionary", "part-of-speech", "phonetic-transcriptions", "pitch-accents", "pitch-accent-graphs", "pitch-accent-graphs-jj", "pitch-accent-positions", "pitch-accent-categories", "reading", "tags", "clipboard-image", "clipboard-text", "cloze-body", "cloze-prefix", "cloze-suffix", "dictionary", "dictionary-alias", "document-title", "frequencies", "frequency-harmonic-rank", "frequency-harmonic-occurrence", "frequency-average-rank", "frequency-average-occurrence", "screenshot", "search-query", "popup-selection-text", "sentence", "sentence-furigana", "sentence-furigana-plain", "url"],
|
||
"maxEntries": 1,
|
||
"includeMedia": False,
|
||
}
|
||
response = requests.post(request_url + "/ankiFields", json = params, timeout = request_timeout)
|
||
return json.loads(response.text)
|
||
|
||
def term_entries(term :str) -> dict:
|
||
print("Requesting termEntries:")
|
||
params = {
|
||
"term": term,
|
||
"maxEntries": 1,
|
||
"includeMedia": False,
|
||
}
|
||
response = requests.post(request_url + "/termEntries", json = params, timeout = request_timeout)
|
||
return json.loads(response.text)
|
||
|
||
# perhaps set up a filter parameter.
|
||
def flatten_string(node):
|
||
if isinstance(node,str):
|
||
yield node
|
||
elif isinstance(node, list):
|
||
for item in node:
|
||
yield from flatten_string(item)
|
||
elif isinstance(node,dict):
|
||
if isinstance(node.get("content"), str):
|
||
yield node["content"]
|
||
else:
|
||
for v in node.values():
|
||
yield from flatten_string(v)
|
||
|
||
def extract_text_by_type(node, target_type):
|
||
if isinstance(node,dict):
|
||
data = node.get("data")
|
||
if isinstance(data,dict) and data.get("content") == target_type:
|
||
yield from flatten_string(node.get("content"))
|
||
#yield from ",".join(node.get("content"))
|
||
for value in node.values():
|
||
yield from extract_text_by_type(value, target_type)
|
||
elif isinstance(node,list):
|
||
for item in node:
|
||
yield from extract_text_by_type(item,target_type)
|
||
|
||
def get_glossary(word :str) -> list:
|
||
glossary = list()
|
||
dict_items = term_entries(word)
|
||
|
||
entry = next(iter(dict_items.get("dictionaryEntries",[])),None)
|
||
if entry:
|
||
definition = next(iter(entry.get("definitions",[])),None)
|
||
if definition:
|
||
def_entry = next(iter(definition.get("entries", [])),None)
|
||
if def_entry:
|
||
glossary.extend( list(extract_text_by_type(def_entry, "glossary")) )
|
||
return glossary[:2]
|
||
|
||
#print(annotate_phrase("私は学校へ行きます"))
|
||
|
||
with open("dump2.json", "r", encoding="utf-16") as f:
|
||
dump = json.load(f)
|
||
|
||
vocab_dict = {}
|
||
## A micab phrase based word parser
|
||
for cd in dump:
|
||
print(dump[cd]['name_J'])
|
||
# node = tagger.parseToNode(dump[cd]['name_J'])
|
||
# while node:
|
||
# # node.surface is the word; node.feature contains POS tags
|
||
# if node.surface:
|
||
# print(f"{node.surface}\t{node.feature}")
|
||
# node = node.next
|
||
anottated = annotate_phrase(dump[cd]['name_J'])
|
||
if anottated['reading'] != dump[cd]['name_J']:
|
||
print(anottated['reading'])
|
||
dump[cd]['kana'] = anottated['reading']
|
||
cd_nm = dump[cd]['name_J']
|
||
for word in anottated['vocab']:
|
||
# Skip punctuation words hopefully
|
||
if len(word) == 1 and has_kanji(word) == False:
|
||
continue
|
||
if vocab_dict.get(word) is None:
|
||
vocab_dict[word] = {
|
||
'count': 0,
|
||
'Examples': (cd_nm,)
|
||
}
|
||
else:
|
||
vocab_dict[word]['count'] = vocab_dict[word]['count'] + 1
|
||
if cd_nm not in vocab_dict[word]['Examples']:
|
||
vocab_dict[word]['Examples'] += (cd_nm,)
|
||
print()
|
||
phrase = dump[cd]['desc_J']
|
||
anottated = annotate_phrase(phrase)
|
||
if anottated['reading'] != phrase:
|
||
#print(anottated['reading'])
|
||
dump[cd]['desc_Furi'] = anottated['reading']
|
||
for word in anottated['vocab']:
|
||
if len(word) == 1 and has_kanji(word) == False:
|
||
continue
|
||
if vocab_dict.get(word) is None:
|
||
vocab_dict[word] = {
|
||
'count': 0,
|
||
'Examples': (cd_nm,)
|
||
}
|
||
else:
|
||
vocab_dict[word]['count'] = vocab_dict[word]['count'] + 1
|
||
if len(vocab_dict[word]['Examples']) < 5:
|
||
if cd_nm not in vocab_dict[word]['Examples']:
|
||
vocab_dict[word]['Examples'] += (cd_nm,)
|
||
|
||
vocab_dict = sorted(vocab_dict.items(), key= lambda x: x[1]['count'], reverse=True)
|
||
|
||
with open("dump2.json", "w", encoding="utf-16") as outfile:
|
||
json.dump(dump, outfile, ensure_ascii=False, indent=2)
|
||
|
||
with open("dump-vocab.json", "w", encoding="utf-16") as outfile:
|
||
json.dump(vocab_dict, outfile, ensure_ascii=False, indent=2)
|