def _create_data():
data = {}
data['quarterly'] = SF1QuarterlyData(config['sf1_data_path'])
data['base'] = SF1BaseData(config['sf1_data_path'])
data['daily'] = SF1DailyData(config['sf1_data_path'])
return data
def main(data_path :str=None):
'''
Download quarterly fundamental data from
https://www.quandl.com/databases/SF1/data
Note:
SF1 is paid, so you need to subscribe
and paste quandl token to `~/.ml_investment/secrets.json`
``quandl_api_key``
Parameters
----------
data_path:
path to folder in which downloaded data will be stored.
OR ``None`` (downloading path will be as ``sf1_data_path`` from
`~/.ml_investment/config.json`
'''
if data_path is None:
config = load_config()
data_path = config['sf1_data_path']
downloader = QuandlDownloader(sleep_time=0.8)
downloader.zip_download('datatables/SHARADAR/TICKERS?qopts.export=true',
'{}/tickers.zip'.format(data_path))
base_df = SF1BaseData(data_path).load()
tickers = base_df['ticker'].unique().tolist()
downloader.ticker_download('datatables/SHARADAR/SF1?ticker={ticker}', tickers,
save_dirpath='{}/core_fundamental'.format(data_path),
skip_exists=False, batch_size=10, n_jobs=4)
downloader.ticker_download('datatables/SHARADAR/DAILY?ticker={ticker}', tickers,
save_dirpath='{}/daily'.format(data_path),
skip_exists=False, batch_size=5, n_jobs=4)
def test_load(self, tickers):
data_loader = SF1BaseData(config['sf1_data_path'])
df = data_loader.load(tickers)
assert type(df) == pd.DataFrame
assert len(df) > 0
assert 'ticker' in df.columns
assert len(set(df['ticker'].values).difference(set(tickers))) == 0
def test_load_no_param(self):
data_loader = SF1BaseData(config['sf1_data_path'])
df = data_loader.load()
assert type(df) == pd.DataFrame
assert len(df) > 0
assert 'ticker' in df.columns
assert df['ticker'].isnull().max() == False
def _create_data():
data = {}
data['quarterly'] = SF1QuarterlyData(config['sf1_data_path'])
data['base'] = SF1BaseData(config['sf1_data_path'])
data['daily'] = SF1DailyData(config['sf1_data_path'])
data['commodities'] = QuandlCommoditiesData(
config['commodities_data_path'])
return data
def main():
'''
Default model training. Resulted model weights directory path
can be changed in `~/.ml_investment/config.json` ``models_path``
'''
pipeline = MarketcapDownStdSF1(pretrained=False)
base_df = SF1BaseData(config['sf1_data_path']).load()
tickers = base_df[(base_df['currency'] == CURRENCY) &\
(base_df['scalemarketcap'].apply(lambda x: x in SCALE_MARKETCAP))
]['ticker'].values
result = pipeline.fit(tickers[:200], median_absolute_relative_error)
print(result)
path = '{}/{}'.format(config['models_path'], OUT_NAME)
pipeline.export_core(path)
SF1DailyData
from synthetic_data import GenQuarterlyData, GenBaseData, GenDailyData
config = load_config()
gen_data = {
'quarterly': GenQuarterlyData(),
'base': GenBaseData(),
'daily': GenDailyData(),
}
datas = [gen_data]
if os.path.exists(config['sf1_data_path']):
sf1_data = {
'quarterly': SF1QuarterlyData(config['sf1_data_path']),
'base': SF1BaseData(config['sf1_data_path']),
'daily': SF1DailyData(config['sf1_data_path']),
}
datas.append(sf1_data)
tickers = ['AAPL', 'TSLA', 'K', 'MAC', 'NVDA']
class TestBasePipeline:
def _create_base_components(self):
columns = ['revenue', 'netinc', 'ncf', 'ebitda', 'debt', 'fcf']
f1 = QuarterlyFeatures(data_key='quarterly',
columns=columns,
quarter_counts=[2, 10],
max_back_quarter=1)