def common_index(self):
all_indices = [data_item.index for data_item in self]
all_indices_flattened = flatten_list(all_indices)
common_unique_index = list(set(all_indices_flattened))
common_unique_index.sort()
return common_unique_index
def common_columns(self):
all_columns = [data_item.columns for data_item in self]
all_columns_flattened = flatten_list(all_columns)
common_unique_columns = list(set(all_columns_flattened))
common_unique_columns.sort()
return common_unique_columns
def as_df(self):
instrument_codes = [
instrument_object.instrument_code for instrument_object in self
]
meta_data_keys = [
instrument_object.meta_data.as_dict().keys() for instrument_object in self
]
meta_data_keys_flattened = flatten_list(meta_data_keys)
meta_data_keys_unique = list(set(meta_data_keys_flattened))
meta_data_as_lists = dict(
[
(
metadata_name,
[
getattr(instrument_object.meta_data, metadata_name)
for instrument_object in self
],
)
for metadata_name in meta_data_keys_unique
]
)
meta_data_as_dataframe = pd.DataFrame(
meta_data_as_lists, index=instrument_codes
)
return meta_data_as_dataframe
def _get_average_return_in_dict_for_column(returns_dict: dictOfReturnsForOptimisation,
column: str) -> float:
## all daily data so can take an average
series_of_returns = [returns_series[column].values
for returns_series in returns_dict.values()]
all_returns = flatten_list(series_of_returns)
return np.nanmean(all_returns)
def _calculate_pseudo_fills(self) -> list:
list_of_years_in_data = years_in_data(self.positions)
fills_by_year = [
self._pseudo_fills_for_year(year) for year in list_of_years_in_data
]
fills_as_single_list = flatten_list(fills_by_year)
return fills_as_single_list
def assets_with_missing_data(self) -> list:
missing_correlations = self.correlation.assets_with_missing_data()
missing_means = self.correlation.assets_with_missing_data()
missing_stdev = self.stdev.assets_with_missing_data()
missing_assets = list(
set(flatten_list([missing_stdev, missing_means, missing_correlations]))
)
return missing_assets
def from_list_of_subportfolios(portfolioWeights,
list_of_portfolio_weights):
list_of_unique_asset_names = \
list(set(flatten_list([list(subportfolio.keys()) for subportfolio in list_of_portfolio_weights])))
portfolio_weights = portfolioWeights.allzeros(
list_of_unique_asset_names)
for subportfolio_weights in list_of_portfolio_weights:
for asset_name in list(subportfolio_weights.keys()):
portfolio_weights[asset_name] = portfolio_weights[
asset_name] + subportfolio_weights[asset_name]
return portfolio_weights
def assets_in_cluster_order(corr_matrix: correlationEstimate,
cluster_size: int = 2):
clusters_as_names = cluster_correlation_matrix(corr_matrix,
cluster_size=cluster_size)
return flatten_list(clusters_as_names)
