Signal#

class sigtech.framework.signal.signal_object.Signal

The signal class creates a signal object that uses the same .history() API as the framework.

Keyword arguments:

  • start_date: minimum date output time series.

  • end_date: maximum date output time series.

  • function: the function object that will produce the output.

  • function_name: function address for database persistence.

  • underliers: A dictionary where keys correspond to the inputs of the function argument and value to the name of the object. This object can be of any type with output of timeseries, dataframe or dictionary, e.g. instruments, strategies or other signal objects.

Example: Creates signal object that has a history that is half the SPXT Index.

def func(ts, n):
    return ts / n

example_signal_obj = Signal(
    signal_function=func,
    underliers={'ts': 'SPXT INDEX'},
    params={'n': 2}
)
start_date: Optional[datetime.date]
end_date: Optional[datetime.date]
start_dt: Optional[datetime.datetime]
end_dt: Optional[datetime.datetime]
function: Optional[collections.abc.Callable]
function_name: Optional[str]
underliers: Optional[dict]
params: Optional[dict]
args: Optional[list]
cache_history: Optional[bool]
additional_dependencies: Optional[list]
ticker: Optional[str]
db_ticker: Optional[str]
db_sector: Optional[str]
data_source_all: Optional[list[str]]
has_preloaded_history(field=None, data_point=None)

Called by the Dependencies framework - arguments not used

property dependency_type

Dependency Type

property history_fields: list[str]

Fields for this signal.

property supplementary_fields

Enable the definition of additional fields.

property prime_history_field: str

The main history field returned by default.

property data_point

Field used when retrieving history.

property data_source

Data source of the signal.

history()

Return the signal history. The history is first built if the signal was not already calculated.

Returns

Signal history.

intraday_history(period: datetime.timedelta, start_dt=None, end_dt=None, timezone=None)

Return the signal history. The history is first built if the signal was not already calculated.

Parameters

  • period – period of history requested

  • start_dt – filter the output by datetime range

  • end_dt – filter the output by datetime range

  • timezone – set the timezone of the returned series to this value

Returns

Signal history.

truncate(var, intraday=False)

Set start and end date for historical outputs of time series, dataframes or dictionaries.

Parameters

  • var – Input history.

  • intraday – bool - is var an intra-day time series

Returns

Truncated history.

get_underlying_names(filter_signals=False)

Return the names of signal underliers.

Parameters

filter_signals – If True, signals are filtered out (default is False).

Returns

List of underlier identifiers.

get_underlying_data()

Return a DataFrame of underliers.

map_underliers()

Return a dict of underliers.

get_dependents()

Return the set of dependents.

update(underliers=None, params=None, del_underliers=None, del_params=None, **kwargs)

Update the signal object.

Parameters

  • underliers – Dict of underliers to add.

  • params – Dict of parameters to add.

  • del_underliers – Dict of underliers to remove.

  • del_params – Dict of parameters to remove.

clone(name=None)

Clone the signal object.

Parameters

name – Ticker identifier of the cloned signal.

Returns

Signal object.

deep_clone()

Clone the signal object (deep copy).

insert()

Update all dependents and return a new signal object.

replace_with(node)

Update all dependents given a new node.

Parameters

node – Input node.

explode_initial_data()

Serialise the signal object.

clear_history()

Clear the signal history.

clear_calc_history()

Clear the signal and underliers history.

set_calendar(calendar_name)

Set the signal object calendar.

Parameters

calendar_name – Calendar identifier.

check_signal_parameters()

Validate the signal parameters.

check_all_signal_inputs()

Validate the signal inputs.

store_signal_diagnoistics(df, df_name)

Store the serialised signal items.

Parameters

  • df – Input data.

  • df_name – Storage name.

print_code(lines=None, imports=None, underlier_set=None)

Code print method.

Parameters

  • lines – List of code lines.

  • imports – List of imports.

  • underlier_set – Set of underliers.

create_operator_signal(other, signal_function, input_names=None)

Create a signal operator.

Parameters

  • other – Signal object.

  • signal_function – Signal operation identifier.

  • input_names – List of input identifiers.

Returns

Signal object.

available_data_providers(entitled_only=True) → Optional[List[str]]

Available data providers for this object. Bool to return entitled providers only

property cache_name: str

Cache name of this object.

property class_name

Class name of this object.

property class_short_name

Short name of the object class.

clear_cached_data()

Clear all the cached data of the object.

clone_object(params=None)

Return a clone of the object with amended parameters.

Parameters

params – Optional dictionary of parameters to override.

Returns

New object.

compute_dependencies(root_dependency: Optional[sigtech.framework.internal.infra.mu.graph.registry.factory.Dependency] = None)

Compute these dependencies - triggers MDS requests

static convert_dtypes(clz)

Class decorator to convert classes’ BaseType into variable annotations (PEP-526), and to generate __aliases__ based on BaseType.db_name when required.

data_dict()

Return the object attributes in a dict.

dependencies(input_dependency: Optional[sigtech.framework.internal.infra.mu.graph.registry.factory.Dependency] = None, valuation_currency: Optional[str] = None, use_start: bool = True) → list[sigtech.framework.internal.infra.mu.graph.registry.factory.Dependency]

Returns a list of Dependency

Parameters

  • input_dependency – A Dependency object representing the root of the dependency tree to be returned.

  • valuation_currency – If supplied, include the dependencies required to output the data in this currency.

  • use_start – If False, return only dependencies required to calculate an earliest start date - self.start_date will be None in this case.

Returns

List of Dependency objects representing the current level in the dependency tree.

property env: sigtech.framework.config.config.ConfiguredEnvironment

Return the configured environment.

env_date_change(old_env_dt: datetime.datetime, new_env_dt: datetime.datetime, live_data_update=None)

Routine called after the environment date is changed.

Parameters

  • old_env_dt – Old datetime (tz aware).

  • new_env_dt – New datetime (tz aware).

  • live_data_update – Live streaming data associated with the environment date change.

ewma(ts)

EWMA operator.

Parameters

  • window – Size of the window.

  • ts – Input timeseries.

Returns

pandas Series or DataFrame.

finalize_for_comparison()

Method to call to ensure all data dict values are finalized prior to doing object comparisons.

classmethod from_dictionary(dct: dict[str, Any], cache: bool = True, identifier: Optional[sigtech.framework.infra.data_adapter.identifier.Identifier] = None, env: Optional[sigtech.framework.config.config.ConfiguredEnvironment] = None, **kwargs)

Factory method to create object using data dictionary

classmethod get_names(sort_by_group: Optional[bool] = False, include_db: Optional[bool] = True, include_local: Optional[bool] = True, include_children: Optional[bool] = False, ignore_unmapped: Optional[bool] = True) → list[str]

Return an ordered list of object names associated with the class.

Parameters

  • sort_by_group – If set, the list is first ordered by sector/group, if applies, e.g. commodity or index futures (default is False).

  • include_db – If set, include objects available from the database (default is True).

  • include_local – If set, include objects available in the local environment (default is True).

  • include_children – If set, include objects available from child classes (default is False).

  • ignore_unmapped – If set, ignore errors due to unmapped database objects (default is True).

Returns

List of object names.

property identifier: sigtech.framework.infra.data_adapter.identifier.Identifier

Return the object identifier.

info() → dict

Return a dictionary with useful object information.

property input_parameters

Values of initial input parameters entered when creating the instance.

property internal_id

Ticker - i.e. part of name determining content of the class. For types with saved reference data - ticker will be self._ticker. For software types - ticker should be calculated.

property is_unmapped

Check if the object is not mapped.

property live_supported: bool

Flag for live supported objects.

ma(ts)

Rolling average operator.

Parameters

  • window – Size of the window.

  • ts – Input timeseries.

Returns

pandas Series or DataFrame.

macd(short_window: int = 12, long_window: int = 26) → pandas.core.series.Series

Moving Average Convergence Divergence Indicator.

See macd() for details.

Parameters

  • ts – Close price timeseries.

  • short_window – Short moving average periods.

  • long_window – Long moving average periods.

Returns

Series of indicator values.

property name

Primary name by which the FrameworkObject is identified. If reference information (or e.g. corresponding time series) are stored in the DB, then this name will be used.

normalize_weights() → pandas.core.frame.DataFrame

Normalise weights to have 100% gross notional.

Parameters

ts – Input timeseries.

Returns

pandas Dataframe.

allocation.normalize_weights(
    pd.DataFrame({'A': [2, 2, 2], 'B': [1, 2, 1]})
)

Returns:

A

B

0

0.67

0.33

1

0.5

0.5

2

0.67

0.33
normalize_weights_zero_filled() → pandas.core.frame.DataFrame

Normalise weights to have 100% gross notional and fill empty values with zeros.

Parameters

ts – Input timeseries.

Returns

pandas Dataframe.

normalize_weights_zero_filled(
    pd.DataFrame({'A': [1,np.nan,2], 'B': [0.5, np.nan, np.nan]})
)

Returns:

Returned allocations:#

A

B

0

0.67

0.33

1

0.0

0.0

2

1.0

0.0
percentile()

Expand the percentile rank of a timeseries.

Parameters

ts – pandas Series or DataFrame.

Returns

pandas Series or DataFrame.

print_dependencies(root_dependency: Optional[sigtech.framework.internal.infra.mu.graph.registry.factory.Dependency] = None, resolve_future_dependencies: bool = True, fields: Optional[list] = None)

Pretty print dependency tree :param root_dependency: starting dependency of the tree. :param resolve_future_dependencies: resolve future dependencies before printing. If ‘false’ nothing get printed. :param fields: additional fields to extract from dependencies object. Default fields are product_type, currency, frequency and data_source. To get the list of available fields for an object you can use the data_dict() method.

property product_type

Return the product_type property of this object.

property sector

Sector used on BBG, or the class ID for SIG instruments.

sigmoid(ts)

Sigmoid operator.

Parameters

  • factor – Factor value.

  • ts – Input timeseries.

Returns

pandas Series or DataFrame.

simple_ma_signal(long_span, df)

Create a signal based on rolling average cross operator.

Parameters

  • short_span – Size of the short span window.

  • long_span – Size of the long span window.

  • df – Input DataFrame.

Returns

Signal object.

static sort_key_static(name)

Convert a name to the sort key as of Portfolio Presentation Guidelines for performance reasons.

Parameters

name – Input name.

Returns

Tuple (sort key, length of sort key, input name).

textual_representation()

Return a printable representation of this object.

validate()

Validation routine adding checks that will be run on object creation.

vol(*, ts, factor=252, **kwargs)

Volatility operator.

Parameters

  • ts – pandas Series or DataFrame.

  • factor – Volatility factor (default is 252).

Returns

pandas Series or DataFrame.

z_score(centered=True, min_periods=None, ts=None)

Z-score calculation.

Parameters

  • period – Size of the moving window.

  • centered – If set to True, do not consider the rolling mean (default is True).

  • min_periods – Minimum number of observations in window required to have a value.

  • ts – Input timeseries.

instrument_id: Optional[int]