def view_calendar(year, month, day):
print('')
print('\033[1m------ CALENDAR ------\033[0m')
cal = Calendar(year, month, day)
line2 = cal.get_calendar_year(year)
line3 = list(map(list, zip(*line2)))
print('')
config = ConfigUtil()
column_num = int(config.get('calendar', 'column_num'))
row_num = math.floor(12 / column_num)
#print('column_num:', column_num, 'row_num:', row_num)
for r in range(row_num):
for x in line3:
print(" ".join(x[column_num * r:column_num * (r + 1)]))
def time_series_range(self, institution, qtr_start, qtr_end):
self.institution = institution
dates = Calendar().quarter_end_list(qtr_start, qtr_end)
frames = []
for date in dates:
df = self.get(date = date, institution=institution) # FIXME; terrible way to call. get all data in one call. cmon...
df = df.sort_values(by=['value'], ascending=False)
frames.append(df)
self.ts = pd.concat(frames).reset_index(drop=True)
self.ts.value = self.ts.value.astype(float) / 1000000
self.ts['_ix'] = self.ts.ticker.astype(str) + '_' + self.ts.securitytype.astype(str)
return self.ts
def main(month='1', year='2012'):
# builds calendar using Python dates
calendar = Calendar(month, year)
cal_html = calendar.build()
# provide values for use in building the Javascript calendar
month = int(month)
months = (
'January', 'February', 'March', 'April',
'May', 'June', 'July', 'August',
'September', 'October', 'November', 'December'
)
month_name = months[month - 1]
prev_year = int(year)
next_year = int(year)
# used for switching between months (links)
prev_month = month - 1
next_month = month + 1
if month == 1:
prev_month = 12
prev_year -= 1
elif month == 12:
next_month = 1
next_year += 1
prev_month_name = months[prev_month - 1]
next_month_name = months[next_month - 1]
# Javascript calculates dates differently, so subtract 1
month -= 1
return render_template('calendar',
year=year, month=month,
month_name=month_name,
prev_month=prev_month,
prev_month_name=prev_month_name,
next_month=next_month, next_year=next_year,
next_month_name=next_month_name,
cal_html=cal_html)
def get(self, years: list = [Calendar().current_year()]):
''' parse US Treasury.gov website for yield curve rates returning concatenated xml content for a list of years
https://home.treasury.gov/resource-center/data-chart-center/interest-rates/TextView?type=daily_treasury_yield_curve&field_tdr_date_value=2022
'''
columns = [
'1 Month', '2 Month', '3 Month', '6 Month', '1 Year', '2 Year',
'3 Year', '5 Year', '7 Year', '10 Year', '20 Year', '30 Year'
]
identifiers = [
'd:NEW_DATE', 'd:BC_1MONTH', 'd:BC_2MONTH', 'd:BC_3MONTH',
'd:BC_6MONTH', 'd:BC_1YEAR', 'd:BC_2YEAR', 'd:BC_3YEAR',
'd:BC_5YEAR', 'd:BC_7YEAR', 'd:BC_10YEAR', 'd:BC_20YEAR',
'd:BC_30YEAR'
]
frames = []
for year in years:
url = f'https://home.treasury.gov/resource-center/data-chart-center/interest-rates/pages/xml?data=daily_treasury_yield_curve&field_tdr_date_value={year}'
document = requests.get(url)
soup = BeautifulSoup(document.content, "lxml-xml")
frames_year = []
for content in soup.find_all('m:properties'):
vlist = []
for text_value in identifiers:
vlist.append(content.find(text_value).text)
df = pd.DataFrame(columns=columns).transpose()
df[vlist[0]] = vlist[1:]
frames_year.append(df.transpose())
df = pd.concat(frames_year)
frames.append(df)
df = pd.concat(frames).reset_index(drop=False).rename(
columns={'index': 'date'})
df.date = pd.to_datetime(df.date)
for c in df.columns:
if c != 'date':
df[c] = pd.to_numeric(df[c])
self.df = df
return self.df
def forecast(request):
ticker = request.POST.get("tickers")
if ticker in ['', None, 'None'] or len(str(ticker)) < 1:
ticker = 'SPY'
print(ticker)
# TODO: call from retreiver
import pandas_datareader as data
import pandas as pd
data_source = 'yahoo'
start_date = '2000-01-01'
end_date = Calendar().today()
stockdata = data.DataReader(ticker, data_source, start_date, end_date)['Close']
stockdata = pd.DataFrame(stockdata).rename(columns={'Close':ticker})
train_start = request.POST.get("train_start")
train_end = request.POST.get("train_end")
test_start = request.POST.get("test_start")
test_end = request.POST.get("test_end")
if train_start in ['', None, 'None']:
train_start = -0.6
train_end = -0.10
test_start = -0.4
test_end = -0.01
# FIXME
ts = TimeSeries(data=stockdata, column=ticker)
ts.decomposition(model='additive')
ts.check_stationarity()
ts.auto_correlation( lags=24 )
ts.prophet_forecast(int(train_start), int(train_end), int(test_start), int(test_end))
img_dirp = os.path.join(os.getcwd(), 'pynance/static/img')
print(f'{ticker}_forecast.png')
context = {
'seasonal_decompose.png':os.path.join(img_dirp, 'seasonal_decompose.png'),
'actual_v_predicted.png':os.path.join(img_dirp, 'actual_v_predicted.png'),
'prophet_forecast.png':os.path.join(img_dirp, 'prophet_forecast.png'),
'autocorrelation.png':os.path.join(img_dirp, 'autocorrelation.png'),
'components.png':os.path.join(img_dirp, 'components.png'),
}
return render(request, 'forecast.html', context)
import pandas as pd
import numpy as np
import nasdaqdatalink
import sys, os
cwd = os.getcwd()
sys.path.append(os.path.dirname(cwd))
from lib.nasdaq import Fundamentals, Metrics, Tickers, Nasdaq
from lib.calendar import Calendar
cal = Calendar()
from dateutil.relativedelta import relativedelta
from postgres import Postgres
import warnings
warnings.simplefilter(action='ignore', category=pd.errors.PerformanceWarning)
warnings.simplefilter(action='ignore', category=RuntimeWarning)
import populate_fundamentals as popfun
'''
calculate quartiles for every metric across each sector and industry for all historical time periods
'''
def init():
calendardates = cal.quarter_end_list('2018-12-31',
cal.previous_quarter_end())
tickers = Tickers().get()
def melt(self, years=[Calendar().current_year()]):
self.melted = self.df.melt(id_vars='date')
self.melted.value = pd.to_numeric(self.melted.value)
return self.melted
def __init__(self):
super().__init__()
self.authenticate()
self.prev_qtr_end = Calendar().previous_quarter_end()
