def test_invalid_args(self):
with pytest.raises(ValueError):
mod.dt_to_iso8601(dtdt(2016, 1, 2, 3, 4, 5, 123456), 'foo', 'ms')
with pytest.raises(ValueError):
mod.dt_to_iso8601(dtdt(2016, 1, 2, 3, 4, 5, 123456), 'ext', 'foo')
with pytest.raises(ValueError):
mod.dt_to_iso8601('foo', 'ext', 'ms')
def create_stock_csv_file(df, category, num=10):
print("Start to create init stock data - " + category)
start_date = dtdt(1980, 1, 1)
for i in range(num):
# 순서대로 종목 stock_item 가져오기
try:
code = str(df.ix[i]['종목코드']).zfill(6)
except:
print("ERROR - Unregisterd Stock!")
continue
if check_csv_file(category, code):
print("Already exist: " + category + '_' + code + '.csv')
continue
stock_item = str(code).zfill(6)
print(category, stock_item)
start_idx_dates = pd.date_range(start_date.strftime('%Y/%m/%d'),
periods=1)
init_data = np.empty((1, len(columns)))
init_data[:] = np.NAN # NaN 값으로 초기화
init_df = DataFrame(init_data, index=start_idx_dates, columns=columns)
get_df = get_info_with_web_scrap(stock_item, init_df, start_idx_dates)
get_df = get_df.dropna().sort_index()
get_df.to_csv('./data/csv/' + category + '_' + stock_item + '.csv')
print("Complete to create init stock data!")
def fomcdates():
################################################################################
# Grabs the FOMC meeting dates
# Description: The FOMC releases a statment after their meetings (historically
# only when a rate decision was made). The FOMC posts its meeting materials on
# its website. This script parses the webpage for the link to the minutes and
# saves them.
# TODO : We should honestly porobably be using the meeting dates displayed on
# the webpage and not the minute links but this seems to work.
# Also need to check to make sure statements are always released at 2:30
################################################################################
# Grab the non-historical meetings first
print('Reading FOMC meeting dates')
url = 'https://www.federalreserve.gov/monetarypolicy/fomccalendars.htm'
raw = urllib.request.urlopen(url).read()
datesRaw = re.findall('monetarypolicy/fomcminutes[0-9]{8}.htm', str(raw))
datesStr = [re.findall('[0-9]{8}', dd)[0] for dd in datesRaw]
dates = [dtdt.strptime(dd, '%Y%m%d') for dd in datesStr]
# Code to get historical meetings
# Get years first
url = 'https://www.federalreserve.gov/monetarypolicy/fomc_historical_year.htm'
start = 1965
end = min(dates).year
# The historical data has a seperate page for each year. Loop through them
for year in range(start, end):
print('Reading FOMC meeting dates for ' + str(year))
url = 'https://www.federalreserve.gov/monetarypolicy/fomchistorical' + \
str(year) + '.htm'
raw = urllib.request.urlopen(url).read()
datesRaw = re.findall('monetarypolicy/files/FOMC[0-9]{8}Agenda.pdf',
str(raw))
datesStr.extend([re.findall('[0-9]{8}', dd)[0] for dd in datesRaw])
dates = [dtdt.strptime(dd, '%Y%m%d') for dd in datesStr]
dates.sort()
releasedays = [x.day for x in dates]
releasemonths = [x.month for x in dates]
releaseyears = [x.year for x in dates]
coveredday = [(x - dtdt(x.year, 1, 1)).days + 1 for x in dates]
return pd.DataFrame({
'release': 'FOMC meeting',
'releaseyear': releaseyears,
'releasemonth': releasemonths,
'releaseday': releasedays,
'releasehour': 14,
'releaseminute': 30,
'coveredyear': releaseyears,
'coveredperiod': coveredday,
'freq': 365
})
def movepath(path):
flag = re.search(cvtpath, path)
if flag is not None and len(flag.groups()) == 6:
date = flag.groups()
date = list(map(lambda x: int(x), date))
day = dtdt(*date).strftime("%a")
dateemb = date[3] * 60 + date[4]
jigen = None
for key in jikanwariemb.keys():
if between(dateemb, *jikanwariemb[key]):
jigen = key
break
if jigen is not None and not target.loc[jigen, day]:
dirpath = os.path.join(savedirname, jikanwari.loc[jigen, day])
os.makedirs(dirpath, exist_ok=True)
if not test:
shutil.move(path, dirpath)
def main():
df = blsjobsdays()
df = df.append(getBLScalendars(), ignore_index=True)
# Some data is pulled twice
df.drop_duplicates()
#df = df.append(fomcdates(), ignore_index=True)
#df = df.append(minutes_dates(), ignore_index=True)
df['releasedate'] = pd.Series([
dtdt(df.loc[i, 'releaseyear'], df.loc[i, 'releasemonth'],
df.loc[i, 'releaseday'], df.loc[i, 'releasehour'],
df.loc[i, 'releaseminute']) for i in range(len(df))
])
df = df[[
'releasedate', 'release', 'releaseyear', 'releasemonth', 'releaseday',
'releasehour', 'releaseminute', 'coveredyear', 'coveredperiod', 'freq'
]]
df = df.set_index(['releasedate', 'release'])
df.sort_index(inplace=True)
df.to_csv('/app/output/bls_release_dates.csv')
return df
def reset_jikanwari():
jigen = {
1: "08:30~10:25",
2: "10:25~12:10",
3: "13:00~14:55",
4: "14:55~16:50",
5: "16:50~18:45",
6: "18:45~20:30",
}
jigen = pd.DataFrame(jigen, index=["jigen"]).T
dirname = []
for i in range(22, 29):
dirname.append(dtdt(*[2020, 6, i]).strftime("%a"))
jikanwari = [[d + str(i) for d in dirname] for i in range(1, 7)]
jikanwari = pd.DataFrame(jikanwari,
columns=dirname,
index=[i for i in range(1, 7)])
home = expanduser("~")
base = os.path.join(home, ".myscreenshot")
os.makedirs(base, exist_ok=True)
jikanwari_path = os.path.join(base, "jikanwari.csv")
pd.concat([jigen, jikanwari], axis=1).to_csv(jikanwari_path)
def test_ext_ms(self):
assert mod.dt_to_iso8601(dtdt(2016, 1, 2, 3, 4, 5, 123456), 'ext', 'ms') == '2016-01-02T03:04:05.123456'
def test_ext_second(self):
assert mod.dt_to_iso8601(dtdt(2016, 1, 2, 3, 4, 5), 'ext', 'second') == '2016-01-02T03:04:05'
def test_basic_ms(self):
assert mod.dt_to_iso8601(dtdt(2016, 1, 2, 3, 4, 5, 123456), 'basic', 'ms') == '20160102T030405123456'
def test_basic_second(self):
assert mod.dt_to_iso8601(dtdt(2016, 1, 2, 3, 4, 5), 'basic', 'second') == '20160102T030405'
def update_stock_data(item, kospi, kosdaq):
try:
code = kospi[kospi['기업명'] == item]['종목코드'].values[0]
category = 'KRX'
except:
try:
code = kosdaq[kosdaq['기업명'] == item]['종목코드'].values[0]
category = 'KOSDAQ'
except:
sys.exit("ERROR - Unregisterd Stock!")
stock_item = str(code).zfill(6)
print(category, item, stock_item)
# 조회 종목 CSV 파일 확인
file_path = './data/csv/'
file_name = category + '_' + stock_item + '.csv'
csv_file = True
if check_csv_file(category, stock_item):
print("Exist csv file:" + file_name)
try:
df = pd.read_csv(file_path + file_name, index_col='Unnamed: 0')
except:
df = pd.read_csv(file_path + file_name, index_col='Date')
df = df.drop(df[-10:].index) # TEST1
#df = df.drop(['2016.12.06', '2016.12.07', '2016.12.08']) # TEST2
else:
df = DataFrame(default_data, columns=columns, index=['1980.1.1'])
csv_file = False
df.index.name = 'Date'
# CSV 파일의 최종 날짜 정보와 조회시점의 정보 누락분 확인
#print("Check delta date")
d = [int(x) for x in df.index[-1].split('.')]
end_date = dtdt(d[0], d[1], d[2])
print(end_date)
check_hour = dtdt.now().hour
check_min = dtdt.now().minute
print(str(check_hour) + ':' + str(check_min))
market_closed = True
if check_hour == 15:
if (check_min > 30):
today = dt.date.today()
elif check_hour > 15:
today = dt.date.today()
else:
market_closed = False
today = dt.date.today() - dt.timedelta(days=1)
delta = dtdt(today.year, today.month, today.day) - end_date
# 구글 파이낸스에서 해당 종목 정보 받기
print("delta days:" + str(delta.days))
if delta.days <= 0:
delta_dates = pd.date_range(end_date.strftime('%Y.%m.%d'), periods=1)
else:
delta_dates = pd.date_range(end_date.strftime('%Y.%m.%d'),
periods=delta.days)
need_web_scrap = True
add_df = DataFrame(default_data, columns=columns)
if csv_file:
if delta.days > 30:
try:
add_df = data.DataReader(
category + ":" + stock_item, "google", end_date,
dtdt(today.year, today.month, today.day))
add_df.index = pd.to_datetime(
add_df.index).strftime('%Y.%m.%d')
need_web_scrap = False
except:
print("Exception to get info via GOOGLE")
if (need_web_scrap):
add_df = get_info_with_web_scrap(stock_item, add_df, delta_dates)
add_df = add_df.dropna().sort_index()
if (market_closed):
print("Market closed")
else:
add_df.drop(add_df[-1:].index)
print("Complete to get stock info via NAVER")
if delta.days > 0:
df = df.append(add_df).dropna()
#if not os.path.isfile(file_path+file_name):
df.to_csv(file_path + file_name)
