def _get_share_filename(share, extension=None):
_check_type(share,
bb.Share,
raise_err=True,
expected_type_name='bulbea.Share')
if extension is not None:
_check_str(extension, raise_err=True)
source = share.source
ticker = share.ticker
start = _get_datetime_str(share.data.index.min(), format_='%Y%m%d')
end = _get_datetime_str(share.data.index.max(), format_='%Y%m%d')
filename = '{source}_{ticker}_{start}_{end}'.format(source=source,
ticker=ticker,
start=start,
end=end)
if extension:
filename = '{filename}.{extension}'.format(filename=filename,
extension=extension)
return filename
def _get_share_filename(share, extension = None):
_check_type(share, bb.Share, raise_err = True, expected_type_name = 'bulbea.Share')
if extension is not None:
_check_str(extension, raise_err = True)
source = share.source
ticker = share.ticker
start = _get_datetime_str(share.data.index.min(), format_ = '%Y%m%d')
end = _get_datetime_str(share.data.index.max(), format_ = '%Y%m%d')
filename = '{source}_{ticker}_{start}_{end}'.format(
source = source,
ticker = ticker,
start = start,
end = end
)
if extension:
filename = '{filename}.{extension}'.format(
filename = filename,
extension = extension
)
return filename
def split(share,
attrs = 'Close',
window = 0.01,
train = 0.60,
shift = 1,
normalize = False,
provider = 'quandl'):
'''
:param attrs: `str` or `list` of attribute names of a share, defaults to *Close* attribute
:type attrs: :obj: `str`, :obj:`list`
'''
_check_type(share, type_ = bb.Share, raise_err = True, expected_type_name = 'bulbea.Share')
_check_iterable(attrs, raise_err = True)
_check_int(shift, raise_err = True)
_check_real(window, raise_err = True)
_check_real(train, raise_err = True)
_validate_in_range(window, 0, 1, raise_err = True)
_validate_in_range(train, 0, 1, raise_err = True)
if(provider == 'alphavantage'):
df = pd.DataFrame(share.data)
data = df.ix[3]
data = pd.to_numeric(data)
else:
data = share.data[attrs]
length = len(share)
window = int(np.rint(length * window))
offset = shift - 1
splits = np.array([data[i if i is 0 else i + offset: i + window] for i in range(length - window)])
if normalize:
splits = np.array([_get_cummulative_return(split) for split in splits])
size = len(splits)
split = int(np.rint(train * size))
train = splits[:split,:]
test = splits[split:,:]
Xtrain, Xtest = train[:,:-1], test[:,:-1]
ytrain, ytest = train[:, -1], test[:, -1]
return (Xtrain, Xtest, ytrain, ytest)
def split(share,
attrs = 'Close',
window = 0.01,
train = 0.60,
shift = 1,
normalize = False):
'''
:param attrs: `str` or `list` of attribute names of a share, defaults to *Close* attribute
:type attrs: :obj: `str`, :obj:`list`
'''
_check_type(share, type_ = bb.Share, raise_err = True, expected_type_name = 'bulbea.Share')
_check_iterable(attrs, raise_err = True)
_check_int(shift, raise_err = True)
_check_real(window, raise_err = True)
_check_real(train, raise_err = True)
_validate_in_range(window, 0, 1, raise_err = True)
_validate_in_range(train, 0, 1, raise_err = True)
data = share.data[attrs]
length = len(share)
window = int(np.rint(length * window))
if window <= 1: window = 2
offset = shift - 1
splits = np.array([data[i if i is 0 else i + offset: i + window] for i in range(length - window)])
print("{} {}".format(window,len(splits)))
if normalize:
splits = np.array([_get_cummulative_return(split) for split in splits])
size = len(splits)
split = int(np.rint(train * size))
train = splits[:split,:]
test = splits[split:,:]
Xtrain, Xtest = train[:,:-1], test[:,:-1]
ytrain, ytest = train[:, -1], test[:, -1]
return (Xtrain, Xtest, ytrain, ytest)
def split(share,
attrs = 'Close',
window = 0.01,
train = 0.60,
shift = 1,
normalize = False):
'''
:param attrs: `str` or `list` of attribute names of a share, defaults to *Close* attribute
:type attrs: :obj: `str`, :obj:`list`
'''
_check_type(share, type_ = bb.Share, raise_err = True, expected_type_name = 'bulbea.Share')
_check_iterable(attrs, raise_err = True)
_check_int(shift, raise_err = True)
_check_real(window, raise_err = True)
_check_real(train, raise_err = True)
_validate_in_range(window, 0, 1, raise_err = True)
_validate_in_range(train, 0, 1, raise_err = True)
data = share.data[attrs]
length = len(share)
window = int(np.rint(length * window))
offset = shift - 1
splits = np.array([data[i if i is 0 else i + offset: i + window] for i in range(length - window)])
if normalize:
splits = np.array([_get_cummulative_return(split) for split in splits])
size = len(splits)
split = int(np.rint(train * size))
train = splits[:split,:]
test = splits[split:,:]
Xtrain, Xtest = train[:,:-1], test[:,:-1]
ytrain, ytest = train[:, -1], test[:, -1]
return (Xtrain, Xtest, ytrain, ytest)
def test__check_type():
with pytest.raises(TypeError):
_check_type('foo', type_ = int, raise_err = True, expected_type_name = 'str')
with pytest.raises(TypeError):
_check_type(12345, type_ = str, raise_err = True, expected_type_name = 'int')
assert _check_type('bar', type_ = str) == True
assert _check_type('foo', type_ = int) == False