def write_date():
_results=get_udemy()
sub = _results['subscribers']
rev = _results['reviews']
date=datetime.date.today()
row = Data(date=date, subscribers=subscribers, reviews=reviews)
db_session.add(row)
db_session.commit()
def write_data():
# 新規データの読み込み
_results = get_udemy_info()
# 書き込むデータ
date = datetime.date.today()
subscribers = _results['n_subscribers']
reviews = _results['n_reviews']
row = Data(date=date, subscribers=subscribers, reviews=reviews)
db_session.add(row)
db_session.commit()
def read_data():
from assets import models
df = pd.read_csv('assets/data.csv')
date = datetime.datetime.strptime(df.iloc[0, 0], '%Y/%m/%d').date()
for index, _df in df.iterrows():
date = datetime.datetime.strptime(_df['date'], '%Y/%m/%d').date()
row = models.Data(date=date,
subscribers=_df['subscribers'],
reviews=_df['reviews'])
db_session.add(row)
# print(date)
db_session.commit()
def update_data():
#スクレイピングデータの読込み
_results = get_info()
#DBに書き込むデータ
date = datetime.date.today() #datetime形式
subscribers = _results['n_subscribers']
reviews = _results['n_reviews']
row = Data(date=date, subscribers=subscribers, reviews=reviews)
db_session.add(row)
db_session.commit()
def write_data():
# get_udemy_info関数内のresults内のデータを見るための準備（新規のデータ）
_results = get_udemy_info()
# 書き込むデータ
date = datetime.date.today()
subscribers = _results['n_subscribers']
reviews = _results['n_reviews']
row = Data(date=date, subscribers=subscribers, reviews=reviews)
db_session.add(row)
db_session.commit()
def write_data():
#df = pd.read_csv("./data.csv")
#新規データ
_results = get_udemy_info()
#書き込みデータ
date = datetime.date.today()
subscribers = _results["n_subscribers"]
reviews = _results["n_reviews"]
#データベースに追加
row = Data(date=date , subscribers=subscribers,reviews=reviews)
db_session.add(row)
db_session.commit()
def write_data():
# df = pd.read_csv('assets/data.csv')
# 新規のデータ
_results = get_udemy_info()
# 書き込むデータ
date = datetime.date.today()
subscribers = _results['n_subscribers']
reviews = _results['n_reviews']
# results = pd.DataFrame([[date, subscribers, reviews]], columns=['date', 'subscribers', 'reviews'])
# df = pd.concat([df, results])
# df.to_csv('assets/data.csv', index=False)
row = Data(date=date, subscribers=subscribers, reviews=reviews)
db_session.add(row)
db_session.commit()
def predict_nextday():
# 価格情報を取得しデータフレームに格納
today_data = db_session.query(Data).all()[-1]
today_data.real_price = get_price() # 上書き（上書きする前はNoneになっている）
db_session.add(today_data)
db_session.commit()
# windowを作成し標準化
window = 5
data_five = db_session.query(Data).all()[-window:]
input_will = []
for price in data_five:
input_will.append(price.real_price)
input_will = np.array(input_will).reshape(1, window, 1)
input_mean = np.mean(input_will)
input_std = np.std(input_will)
sta_input = (input_will - input_mean) / input_std
# 学習済みモデルを取得
with open("lstm_model.pickle", "rb") as f:
model = pickle.load(f)
# 翌日の価格を予測、標準化してあるので元に戻す
future_price = model.predict(sta_input)
future_price = future_price.reshape(-1)
future_price = future_price * input_std + input_mean
future_price = float(future_price)
# 翌日の日付を取得
today = db_session.query(Data).all()[-1].date
next_day = today + timedelta(days=1)
# 翌日の日付、予測した価格をデータベースに格納
row = Data(date=next_day, real_price=None, pred_price=future_price)
db_session.add(row)
db_session.commit()
from assets.database import db_session from assets.database import init_db from assets.models import Data import datetime init_db() #create date = datetime.date.today() row1 = Data(date=date, subscribers=6500, reviews=210) row2 = Data(date=date, subscribers=1500, reviews=220) db_session.add(row1) db_session.add(row2) db_session.commit() #commitでデータベースに保存 #read read1 = db_session.query(Data).all()[0].subscribers #update datum = db_session.query(Data).all()[0] datum.subscribers = 1000 db_session.add(datum) db_session.commit() #dalete datum = db_session.query(Data).filter_by(subscribers=1000).one() db_session.delete(datum) db_session.commit() #全削除 #Data.query.delete() print(db_session.query(Data).all())
