DefaultUniverses

class sigtech.framework.experimental.default_universes.default_universes.DefaultUniverses

Access the default universes of returns series for quick assessment of signal potential.

Constructor args are passed to the file utilities class used to cache returns series.

Example usage:

from sigtech.framework.experimental.default_universes import DefaultUniverses, AssetType

du = DefaultUniverses(force_local=True, local_path='du_store')

# Print out a pd.DataFrame of available returns series
du.overview()

# Fetch (and cache) all predefined PreciousMetals assets
du.universe("PreciousMetals")

Parameters

• force_local – Boolean to override the cache to be set to local rather than use s3 locations.

• local_path – Local path to use for the file cache.

static overview(return_df: bool = False) → Optional[pandas.core.frame.DataFrame]

Return a pd.DataFrame of the available default universes. Universes are split by asset type.

Example usage:

du = DeafultUniverses()

# Print out a pd.DataFrame of available returns series
du.overview()

# Return the overview dataframe for manual filtering
du.overview(return_df=True).loc['Energy']

Parameters

return_df – If True the overview dataframe is returned from this method for filtering. Otherwise, if False the dataframe is displayed and nothing is returned.

Returns

pd.DataFrame if return_df else None.

universe(universe_name: str, asset_types: Optional[list[sigtech.framework.internal.default_universes.asset_type_mapping.AssetType]] = None, start_date: datetime.date = datetime.date(2010, 1, 1), end_date: Optional[datetime.date] = None, eq_freq: str = 'Q', eq_ffields: Optional[list[str]] = None, ccy: bool = 'USD', use_cache: bool = True, tickers: Optional[Union[str, list[str]]] = None, progress: bool = False, raise_exceptions: bool = True) → pandas.core.frame.DataFrame

Generate the requested default universe and return a DataFrame of universe returns. Most AssetTypes simply use the default_strategy_objects but the SingleNameEquitys use a filter on the Russell 3000 universe for stocks in the requested sector. This filtering can take some time and requires some additional kwargs outlined below.

..Note::

Higher eq_freq frequencies will take some time to analyse. Suggestion is 'Q'.

Example usage:

from sigtech.framework.experimental.default_universes import DefaultUniverses, AssetType

du = DefaultUniverses(force_local=True, local_path='du_store')

# Return (and cache) all returns series relating to the U.S.
df_usa = du.universe('USA')

# Return (and cache) only ETFs and Futures from the "Footfall" category
df_footfall = du.universe(universe_name='Footfall', asset_types=[AssetType.ETF, AssetType.Future])

# Return the top 10 stocks (by market cap) in the "Industrials" sector, rebalancing each quarter
# In this example we don't use the cache even if available but the new data will be saved into the cache
df_industrial_stocks = du.universe(
    universe_name='Industrials',
    asset_type=[AssetType.SingleNameEquity],
    start_date=dtm.date(2020, 1, 1),
    end_date=None,  # Defaults to today,
    eq_freq='Q',
    use_cache=False
)

# Get reinvestment strategy returns series for specific stock tickers
df_stocks = du.universe(
    'SpecificStocks',
    tickers=[
        '1000045.SINGLE_STOCK.TRADABLE',
        '1001962.SINGLE_STOCK.TRADABLE'
    ]
)

# Custom universes can be defined by inheriting from ``Universe``.
# The string keys for the assets can be found in the ``loader`` files like:
# sigtech/framework/internal/default_universes/loaders/

from sigtech.framework.internal.default_universes.universe import Universe

class CustomUniverse(Universe):
    name = 'MyCustomUniverse'
    strategy_names = {
        AssetType.Future: [
            "Crude Oil", "Gas Oil"
        ]
    }
du.universe("MyCustomUniverse")

Parameters

• asset_types – List of Types for which the returns series will be provided.

• start_date – Starting date for the rates and single name equities filter. No effect on other AssetTypes.

• end_date – End date for the rates and single name equities filter. No effect on other AssetTypes.

• eq_freq – Pandas frequency on which to sample equity single stock universes between start_date and end_date.

• eq_ffields – List of equity fundamental fields to pass to return single name equity fundamentals.

• tickers – List of single name equity tickers to pass to AssetType.SpecificStocks (e.g. “1000045.SINGLE_STOCK.TRADABLE”) or list of rates keys (e.g. “USD 2Y”) to pass to AssetType.FairSwapRate. Ignored for any other AssetType.

• use_cache – Boolean whether to use cache files. NOTE: use_cache=False will save new data to the cache even if previous versions exist. The cache will not be cleaned (previous versions removed) until the same query is re-run with use_cache=True.

• progress – If True a progress bar is displayed, incrementing each time an asset is loaded.

• raise_exceptions – If False exceptions raised during asset data loading will be captured and logged. Otherwise, if True (default) exceptions are raised from this method.

Returns

DataFrame of universe returns.