def make_keywords(sentence):
# gooAPIが発行してくれたAPI ID
with open('apikey.json', 'r') as f:
api_data = json.load(f)
app_id = api_data['keyword_api_key']
api = GoolabsAPI(app_id)
# See sample response below.
# 例としてpythonによるデータ分析入門の中表紙の文字をコピペしたものをbodyとして
# キーワード抽出してみた。
# template = u'1.2なぜPythonはデータ分析者におすすめなのか\n私自身を含む多くの人にとって、Pythonという言語は恋に落ちやすい言語です。1991年の登場の時から、\nPythonは、PerlやRubyなどの言語と並び、最も有名な動的プログラミング言語の1つでしたPythonと\nRubyは最近では、多数のWebフレームワーク(たとえば、RubyではRails, PythonではDjango)を使ったWebサイト構築で特に有名です。これらの言語はよくスクリプト言語と呼ばれます。これは、汚くてもすぐに書ける短いプログラム、つまり、スクリプトを書くのに使えるからです。'
template = sentence
sample_response = api.keyword(title="photo01", body=template, max_num=5)
# pprintで、整形された状態でprintできる（sample_responseは辞書型のデータ）
# pprint.pprint(sample_response)
# max_num個のキーワードをリスト型にして出力。一緒に出てくる数字は重要度を表す？
keywords_list = sample_response['keywords']
data = []
for keyword in keywords_list:
data.extend(list(keyword.keys()))
return data
def morph(ctx, app_id, sentence_file, json_flag, sentence, info_filter,
pos_filter, request_id):
# type: (Context, unicode, Optional[IO], bool, unicode, unicode, unicode, unicode) -> None # NOQA
""" Morphological analysis for Japanese."""
app_id = clean_app_id(app_id)
sentence = clean_sentence(sentence, sentence_file)
if info_filter:
info_filter = info_filter.replace(',', '|')
if pos_filter:
pos_filter = pos_filter.replace(',', '|')
api = GoolabsAPI(app_id)
ret = api.morph(
sentence=sentence,
info_filter=info_filter,
pos_filter=pos_filter,
request_id=request_id,
)
if json_flag:
click.echo(format_json(api.response.json()))
return
for words in ret['word_list']:
for word in words:
click.echo(','.join(word))
def keyword(ctx, app_id, body_file, json_flag, title, body, max_num, forcus,
request_id):
# type: (Context, unicode, Optional[IO], bool, unicode, unicode, int, unicode, unicode) -> None # NOQA
"""Extract "keywords" from an input document. """
app_id = clean_app_id(app_id)
body = clean_body(body, body_file)
api = GoolabsAPI(app_id)
ret = api.keyword(
title=title,
body=body,
max_num=max_num,
forcus=forcus,
request_id=request_id,
)
if json_flag:
click.echo(format_json(api.response.json()))
return
for k in ret['keywords']:
k = dict((key.encode('utf-8'), k[key]) for key in k.keys())
for keyword, score in six.iteritems(k):
click.echo(u'{0},{1}'.format(text(keyword), score))
def judge(word):
app_id = "39bc88fcf7da5a2e42e311dbf872353f8a23960f7d4f021b20fefc7504ec76c6"
api = GoolabsAPI(app_id)
# See sample response below.
ret = api.similarity(query_pair=[word, favoriteword])
print ret['score']
return ret['score']
def __init__(self, text: str, goolab_api_key: str):
"""
GoolabAPIを使用して文字列から問題作成をするクラス
:param text: 問題作成用文字列
:param goolab_api_key: goolabのAPI Key
"""
self.text = text
self.goolab = GoolabsAPI(goolab_api_key)
self.keywords = self.__get_keywords()
self.questions = self.__create_questions()
def get_posChecker(keyword_list):
app_id = "72557413b523d38db2d1de26f8095928d43d6d0882707ed41249f9edb643db45"
api = GoolabsAPI(app_id)
check_list = []
for keyword in keyword_list:
check_list.append(
str(api.morph(sentence=keyword)["word_list"][0][0][1]))
if len(list(set(check_list))):
return True
else:
return False
def similarity(ctx, app_id, json_flag, query_pair, request_id):
# type: (Context, unicode, bool, List[unicode], unicode) -> None
""" Scoring the similarity of two words. """
app_id = clean_app_id(app_id)
api = GoolabsAPI(app_id)
ret = api.similarity(query_pair=query_pair, request_id=request_id)
if json_flag:
click.echo(format_json(api.response.json()))
return
click.echo('{0:.16f}'.format(ret['score']))
def hiragana(ctx, app_id, sentence_file, json_flag, sentence, output_type,
request_id):
# type: (Context, unicode, Optional[IO], bool, unicode, unicode, unicode) -> None # NOQA
""" Convert the Japanese to Hiragana or Katakana. """
app_id = clean_app_id(app_id)
sentence = clean_sentence(sentence, sentence_file)
api = GoolabsAPI(app_id)
ret = api.hiragana(sentence=sentence,
output_type=output_type,
request_id=request_id)
if json_flag:
click.echo(format_json(api.response.json()))
return
click.echo(ret['converted'])
def shortsum(ctx, app_id, review_file, json_flag, review, length, request_id):
# type: (Context, unicode, Optional[IO], bool, unicode, unicode, unicode) -> None # NOQA
"""Summarize reviews into a short summary."""
app_id = clean_app_id(app_id)
review_list = clean_review(review, review_file)
length_int = clean_length(length) # type: Optional[int]
api = GoolabsAPI(app_id)
ret = api.shortsum(
review_list=review_list,
length=length_int,
request_id=request_id,
)
if json_flag:
click.echo(format_json(api.response.json()))
return
click.echo(ret['summary'])
def chrono(ctx, app_id, sentence_file, json_flag, sentence, doc_time,
request_id):
# type: (Context, unicode, Optional[IO], bool, unicode, unicode, unicode) -> None # NOQA
"""Extract expression expressing date and time and normalize its value """
app_id = clean_app_id(app_id)
sentence = clean_sentence(sentence, sentence_file)
api = GoolabsAPI(app_id)
ret = api.chrono(
sentence=sentence,
doc_time=doc_time,
request_id=request_id,
)
if json_flag:
click.echo(format_json(api.response.json()))
return
for pair in ret['datetime_list']:
click.echo(u'{0}: {1}'.format(text(pair[0]), pair[1]))
def entity(ctx, app_id, sentence_file, json_flag, sentence, class_filter,
request_id):
# type: (Context, unicode, Optional[IO], bool, unicode, unicode, unicode) -> None # NOQA
""" Extract unique representation from sentence. """
app_id = clean_app_id(app_id)
sentence = clean_sentence(sentence, sentence_file)
if class_filter:
class_filter = class_filter.replace(',', '|')
api = GoolabsAPI(app_id)
ret = api.entity(sentence=sentence,
class_filter=class_filter,
request_id=request_id)
if json_flag:
click.echo(format_json(api.response.json()))
return
for ne in ret['ne_list']:
click.echo(','.join(ne))
class QuestionGenerator(object):
def __init__(self, text: str, goolab_api_key: str):
"""
GoolabAPIを使用して文字列から問題作成をするクラス
:param text: 問題作成用文字列
:param goolab_api_key: goolabのAPI Key
"""
self.text = text
self.goolab = GoolabsAPI(goolab_api_key)
self.keywords = self.__get_keywords()
self.questions = self.__create_questions()
def __get_keywords(self):
"""
textパラメータからGoolabAPIを使用してキーワード抽出
:return: キーワードリスト
"""
keywords = []
if not self.text:
return []
ret = self.goolab.entity(sentence=self.text,
class_filter=u"PSN|ORG|ART|DAT")
for idx in range(len(ret['ne_list'])):
key = (ret['ne_list'][idx][0], ret['ne_list'][idx][1])
keywords.append(key)
return keywords
def __create_questions(self):
"""
抽出したキーワードに文章を付加して問題文を生成する
生成した問題文を”リスト”で返すことに注意
:param keyword: 抽出したキーワード
:return questions: 生成した問題文
"""
# カテゴリ別で問題を作成
questions = []
for key in self.keywords:
if key[1] == 'PSN':
questions.append(key[0] + 'は何をしたか？')
questions.append(key[0] + 'について説明せよ。')
elif key[1] == 'ART':
questions.append(key[0] + 'とは何か？')
questions.append(key[0] + 'は誰が考案したか？')
elif key[1] == 'DAT':
questions.append(key[0] + 'には何が起こったか？')
elif key[1] == 'ORG':
questions.append(key[0] + 'はいつ作られたか？')
questions.append(key[0] + 'について説明せよ。')
return questions
def get_questions(self, max_questions: int = 3):
"""
問題文リストからランダムに問題を返す
:param max_questions: 問題数
:return: ランダムに選択された問題文リスト
"""
n_ques = len(self.questions)
if max_questions > n_ques:
max_questions = n_ques
return random.sample(self.questions, max_questions)
from goolabs import GoolabsAPI
import configparser
config = configparser.ConfigParser()
config.read('config.ini')
app_id = config['gooAPI']['id']
api = GoolabsAPI(app_id)
ret = api.entity(sentence=u"鈴木さんがきょうの9時30分に横浜に行きます。")
print(ret)
def get_request(text):
api = GoolabsAPI(key)
response = api.morph(sentence=text)
return response
def get_promise_content(letter):
api = GoolabsAPI(GOO_API_KEY)
index = 0
# letter = "クリスマスに岩見と一緒に東京駅に来て"
# 時刻情報正規化API
# chrono_response = api.chrono(sentence=letter)
# 固有表現抽出API
entity_response = api.entity(sentence=letter)
# 形態素解析API
# morph_response = api.morph(sentence=letter)
date = ''
hour = ''
min = ''
place = ''
for i in range(len(entity_response["ne_list"])):
# 日付の抽出
if entity_response["ne_list"][i][1] == "DAT":
day_response = api.chrono(sentence=entity_response["ne_list"][i][0])
if len(day_response['datetime_list']) != 0:
date = day_response["datetime_list"][0][1]
# 約束内容の要素番号を算出
day = entity_response["ne_list"][i][0]
index = letter.find(day) + len(day)
# 時間の抽出
elif entity_response["ne_list"][i][1] == "TIM":
time_len = len(entity_response["ne_list"][i][0])
for j in range(time_len):
if entity_response["ne_list"][i][0][j] == u"時":
hour = entity_response["ne_list"][i][0][0:j]
hour_n = j
# 約束内容の要素番号を算出
time = entity_response["ne_list"][i][0]
index = letter.find(time) + len(time)
if j != time_len - 1:
if entity_response["ne_list"][i][0][j + 1] == u"半":
min = "30"
elif entity_response["ne_list"][i][0][j] == u"分":
min = entity_response["ne_list"][i][0][hour_n + 1:j]
elif entity_response["ne_list"][i][1] == "LOC":
place = entity_response["ne_list"][i][0]
# 約束内容を出力
# 約束内容の最初の文字に格助詞が入ってるか調べる
morph_response = api.morph(sentence=letter[index:len(letter)])
if morph_response["word_list"][0][0][1] == u"格助詞" or morph_response["word_list"][0][0][1] == u"読点":
par_len = len(morph_response["word_list"][0][0][0])
content = letter[index + par_len:len(letter)]
else:
content = letter[index:len(letter)]
return {'date': date, 'hour': hour, 'min': min, 'content': content, 'place': place}
APP_ROOT = os.path.dirname(os.path.abspath(__file__))
# 設定ファイルの読み込み
config = configparser.ConfigParser()
config.read(APP_ROOT + '/config_local.ini')
# 設定ファイルからgooラボAPIに関する情報を取得
Goo_API_APPLICATION_ID = config.get("Goo_API", "ApplicationId")
# 設定ファイルからRECRUITのAPIに関する情報を取得
RECRUITE_API_PROOFREADING_API_KEY = config.get("RECRUITE_API",
"ProofreadingAPIKey")
RECRUITE_API_PROOFREADING_URL = config.get("RECRUITE_API", "ProofreadingURL")
# JSON送信用のヘッダー
headers = {"Content-type": "application/json"}
# gooラボAPIのAPIクライアント設定
gooAPI = GoolabsAPI(Goo_API_APPLICATION_ID)
## 辞書データ取得
json_open = open(APP_ROOT + '/dict.json', 'r')
HumbleLangDict = json.load(json_open)
## 名詞用辞書データ取得
json_open = open(APP_ROOT + '/noun.json', 'r')
HumbleNounDict = json.load(json_open)
# 探索の省略が可能な品詞（Part of speech to omit）
Posto = ['句点', '読点', '空白', '格助詞', '終助詞', '括弧', '助数詞', '助助数詞', '冠数詞']
# 人名と会社名をリストで取得する関数
def get_list_people_companies(sentence):
""" データベースに接続する"""
connector = MySQLdb.connect(
unix_socket="/Applications/MAMP/tmp/mysql/mysql.soc\
k",
host="localhost",
db="Wordrop",
user="******",
passwd="root",
charset="utf8")
cur = connector.cursor()
cur.execute("select * from home;")
app_id = "2d84d0d734ebefeb1f4dcf8ae106ec9d2f3b72a5be084a1014d6e27a9002ffef"
api = GoolabsAPI(app_id)
response = api.morph(pos_filter="名詞", sentence=tex) #sentenceにある言葉を形態素解析する
value = response["word_list"] #形態素解析の結果を代入 \
response = api.entity(sentence=value) #名詞の固有表現を求める \
list = response.values()
pp(value)
response = api.entity(sentence=value[0][0][0])
print type(value)
ss = []
from goolabs import GoolabsAPI
import configparser
config = configparser.ConfigParser()
config.read('config.ini')
app_id = config['gooAPI']['id']
api = GoolabsAPI(app_id)
ret = api.keyword(
request_id="keyword-req001",
title="「和」をコンセプトとする 匿名性コミュニケーションサービス「MURA」",
body="NTTレゾナント株式会社（本社：東京都港区、代表取締役社長：若井 昌宏",
max_num=10,
forcus="ORG",
)
print(ret)
import os
import subprocess
sys.stdin = io.TextIOWrapper(sys.stdin.buffer, encoding='utf-8')
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')
sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding='utf-8')
# チャット画面で生成されたテキストファイルを読み込む
chat_text = ""
with open('./data/chat_text.txt', newline='', encoding='utf-8') as f:
for row in f:
chat_text = row
print(row)
app_id = "13346dab6a67b41a884de45bcbf4bdc523ced4da711c35346f0a8501ef0b23fe"
api = GoolabsAPI(app_id)
echat_text = chat_text.encode('shift-jis')
uchat_text = echat_text.decode('shift-jis')
# # See sample response below.
# ret = api.entity(sentence=uchat_text)
# with open('./data/entity_response.json', 'w', newline='', encoding='utf-8') as f:
# f.write(str(ret))
# See sample response below.
ret = api.morph(sentence=uchat_text)
with open('./data/morph_response.json', 'w', newline='',
encoding='utf-8') as f:
f.write(str(ret))
# print (json.dumps(ret, sort_keys = True, indent = 4))
# {'word_list': [[['トイレットペーパー', '名詞', 'トイレットペーパー'], \
from goolabs import GoolabsAPI
import json
# api取得
app_id = "9707a9ca41154956524fe5ef01ba774b4305ccc701adfb6be574a87ba4a5687b"
api = GoolabsAPI(app_id)
# 元のテキストデータ
f = open('before.txt', 'r', encoding='UTF-8')
data = f.read()
OriginalText = data
# 単語探索関数
def SearchForWords(sentence):
for start in range(len(sentence)):
for end in range(len(sentence) - 1, start - 1, -1):
testKey = ''
for check in range(start, end + 1):
testKey += sentence[check][0]
if testKey in HumbleLangDict:
if testKey not in HitWordList:
HitWordList.append(testKey)
# 単語置き換え用関数
def ChangeWord(text, HitWordList):
ConvertedText = text
for word in HitWordList:
ConvertedText = ConvertedText.replace(word, HumbleLangDict[word])
return ConvertedText
マッチした文字列の元であるkeyを返す
'''
for key in adict.keys():
if re.search(key, text):
return key
def one_xlat(match):
return adict[dedictkey(match.group(0))]
return rx.sub(one_xlat, text)
original_text = '大変だ！抽選10回もできるぞ！（基本5回、ツイートで5回）'
# csvファイルから辞書データを作成
with open('dict.csv', 'r') as f:
csvdata = csv.reader(f)
data = [x for x in csvdata]
datadic = dict(data)
APPID = sys.argv[1]
api = GoolabsAPI(APPID)
req = api.hiragana(sentence=original_text, output_type='hiragana')
after = multiple_replace(req['converted'], datadic)
print('Before:\t' + original_text)
print('After:\t' + after)
