def getPredFromDB():
predData = pred.query.order_by(desc(pred.date)).limit(61).all()
date, predPrice = [], []
for i in predData:
date.insert(0, i.date)
predPrice.insert(0, i.closePrice)
closePrice = getPriceFromDB()
return (date, closePrice, predPrice)
def getCryptoHist(coin):
dates = three_months_back()
print(dates)
url = 'https://coinmarketcap.com/currencies/' + coin + '/historical-data/?start=%s&end=%s' % (
dates[1], dates[0])
print(url)
resp = requests.get(url)
soup = bs.BeautifulSoup(resp.text)
sevenDay = soup.find("div", {"id": "historical-data"})
tab = sevenDay.find('table')
body = tab.find('tbody')
rows = body.findAll('tr')
date = []
openPrice = []
high = []
low = []
close = []
volume = []
marketCap = []
for row in rows: #pack lists
col = row.findAll('td')
col0 = col[0].text.strip()
date.insert(0, col0)
col1 = col[1].text.strip()
openPrice.insert(0, float(col1))
col2 = col[2].text.strip()
high.insert(0, float(col2))
col3 = col[3].text.strip()
low.insert(0, float(col3))
col4 = col[4].text.strip()
close.insert(0, float(col4))
col5 = col[5].text.strip()
col5 = col5.replace(',', '')
volume.insert(0, int(col5))
col6 = col[6].text.strip()
col6 = col6.replace(',', '')
marketCap.insert(0, int(col6))
tableDict = {
'date': date,
'open': openPrice,
'high': high,
'low': low,
'close': close,
'volume': volume,
'marketCap': marketCap
}
#df = pd.DataFrame(tableDict)
return tableDict
def prediction():
# load model
json_file = open('model.json', 'r')
loaded_model_json = json_file.read()
json_file.close()
loaded_model = model_from_json(loaded_model_json)
# load weights into new model
loaded_model.load_weights("model.h5")
# Get Latest Historical Data
testData = hp.query.order_by(desc(hp.date)).limit(121).all()
# get only close price
date, testPrice = [], []
for i in testData:
date.insert(0, i.date)
testPrice.insert(0, i.closePrice)
date.append(nextTradingDay())
# Change to np array to reshape
testPrice = np.asarray(testPrice)
# reshape array to apply sc transform
testPrice = testPrice.reshape(-1, 1)
sc = load('std_scaler.bin')
testPrice = sc.transform(testPrice)
# prepare data for prediction
X_test = []
for i in range(60, testPrice.shape[0]):
X_test.append(testPrice[i - 60:i, 0])
X_test = np.array(X_test)
X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1))
# perform prediction
result = loaded_model.predict(X_test)
# Return actual result
normal_result = sc.inverse_transform(result)
normal_result = np.array(normal_result).tolist()
predPrice = []
for i in normal_result:
predPrice.append(round(i[0], 4))
return (date[61:122], predPrice)
def country_wise():
url_cases = 'https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv'
url_death = 'https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv'
url_recovered = "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_recovered_global.csv"
df_cases = pd.read_csv(url_cases)
df_death = pd.read_csv(url_death)
df_recovered = pd.read_csv(url_recovered)
#year = (datetime.now()).strftime('%Y')
data_cases = df_cases.groupby(['Country/Region']).sum()
data_death = df_death.groupby(['Country/Region']).sum().reset_index()
data_recovered = df_recovered.groupby(['Country/Region']).sum().reset_index()
data_cases.drop(columns=['Lat','Long'],inplace=True)
data_death.drop(columns=['Lat','Long'],inplace=True)
data_recovered.drop(columns=['Lat','Long'],inplace=True)
date = []
for i in data_cases.columns:
a = i.split('/')
b = a[1]+ ' '+ calendar.month_abbr[int(a[0])] + ' 20'+ a[2]
date.append(b)
date.insert(0,'Country')
data_cases = data_cases.reset_index()
X_cases = data_cases.iloc[:,:].values
X_deaths = data_death.iloc[:,:].values
X_rec = data_recovered.iloc[:,:].values
new_cases = pd.DataFrame(X_cases,columns = [date])
new_deaths = pd.DataFrame(X_deaths,columns = [date])
new_rec = pd.DataFrame(X_rec,columns = [date])
new_cases.to_csv('data/country_wise_cases.csv')
new_deaths.to_csv('data/country_wise_deaths.csv')
new_rec.to_csv('data/country_wise_recovered.csv')
def formatDate(dates):
date = dates.split()
if len(date) == 1: # if it only contains the year
date.insert(0, '1')
date.insert(1, 'JAN')
if len(date) == 2: # if it only contains month and year
date.insert(0, '1')
if (int(date[0]) in range(1, 10)):
date[0] = "0" + date[0]
month = months[date[1]]
year = date[2]
return (year + '-' + month + '-' + date[0])