def load_file(filename):
assert 'daily_' in filename
ticker = filename.split('_')[1]
full_path = op.join(BASE_DIR, filename)
data = readCSV(full_path, date_name = 'Date', date_format='%Y%m%d')
data = np.sort(data, 0, order=['Date'])
data = data[np.where(data['Date'] > datetime.date(2004, 1, 1))]
iso_week, new_data = rearrange(data)
return ticker, iso_week, new_data
def get_sig_data(filename, sig_funcs):
data = readCSV(filename, datetime_name="Datetime", datetime_format="%Y%m%d%H%M")
data = np.sort(data, 0, order=["Datetime"])
dates = np.array([dt.date() for dt in data["Datetime"]])
unique_dates = np.unique(dates)
print len(unique_dates)
print unique_dates
result = []
for ii in range(0, len(unique_dates)):
tmp = data[np.where(dates == unique_dates[ii])]
sig = [sf(tmp) for sf in sig_funcs]
row = sig
result.append(row)
return np.array(result)
def load_all_data(self, filenames, min_date='20040101'):
result = {}
dates = []
for fname in filenames:
assert 'daily_' in fname
ticker = op.split(fname)[-1].split('_')[1]
rec = readCSV(fname, date_name='Date', date_format='%Y%m%d')
rec = np.sort(rec, 0, order=['Date'])
result[ticker] = rec
dates = np.append(dates, rec['Date'])
self._result = result
self._dates = [dt.strftime('%Y%m%d') for dt in np.unique(dates) if dt.strftime('%Y%m%d') >= min_date ]
self._dates.sort()
print "Tickers: "
print " ", ", ".join(result.keys())
def get_data(filename, sig_funcs, fut_func):
data = readCSV(filename, datetime_name="Datetime", datetime_format="%Y%m%d%H%M")
data = np.sort(data, 0, order=["Datetime"])
dates = np.array([dt.date() for dt in data["Datetime"]])
unique_dates = np.unique(dates)
print len(unique_dates)
print unique_dates
result = []
for ii in range(1, len(unique_dates) - 1):
tmp = data[np.where(dates == unique_dates[ii - 1])]
sig = [sf(tmp) for sf in sig_funcs]
fut_tmp = data[np.where((dates == unique_dates[ii]) | (dates == unique_dates[ii + 1]))]
fut = fut_func(tmp, fut_tmp)
row = sig + [fut]
result.append(row)
return np.array(result)
def load_all_data(self, filenames, ret_func, min_date='20040101'):
result = {}
dates = []
for fname in filenames:
assert 'daily_' in fname
ticker = op.split(fname)[-1].split('_')[1]
rec = readCSV(fname, date_name='Date', date_format='%Y%m%d')
rec = np.sort(rec, 0, order=['Date'])
result[ticker] = ret_func(rec, ticker)
tickers, recs = zip(*result.items())
self._tickers = tickers
self._result = recs_inner_join('Date', recs,
postfixes=tickers)
self._result = np.sort(self._result, 0, order=['Date'])
self._dates = [dt.strftime('%Y%m%d') for dt in self._result['Date'] if dt.strftime('%Y%m%d') >= min_date ]
print "Tickers: "
print " ", ", ".join(result.keys())
