Chore: Make release 1.0.49

gs_quant/target/data.py

@@ -28,7 +28,8 @@ class DataSetType(EnumBase, Enum):
     """Type of the dataset"""
 
     PlotTool_Pro = 'PlotTool Pro'
-    Alloy = 'Alloy'    
+    Alloy = 'Alloy'
+    Snowflake = 'Snowflake'    
 
 
 class DelayExclusionType(EnumBase, Enum):    
@@ -521,6 +522,7 @@ class DataSetParameters(Base):
     support_distribution_list: Optional[Tuple[str, ...]] = field(default=None, metadata=field_metadata)
     apply_market_data_entitlements: Optional[bool] = field(default=None, metadata=field_metadata)
     upload_data_policy: Optional[str] = field(default=None, metadata=field_metadata)
+    database_id: Optional[str] = field(default=None, metadata=field_metadata)
     logical_db: Optional[str] = field(default=None, metadata=field_metadata)
     symbol_strategy: Optional[str] = field(default=None, metadata=field_metadata)
     underlying_data_set_id: Optional[str] = field(default=None, metadata=field_metadata)

gs_quant/target/groups.py

@@ -49,20 +49,6 @@ class UserCoverage(Base):
     guid: Optional[str] = field(default=None, metadata=field_metadata)
 
 
-@handle_camel_case_args
-@dataclass_json(letter_case=LetterCase.CAMEL)
-@dataclass(unsafe_hash=True, repr=False)
-class UserTag(Base):
-    name: str = field(default=None, metadata=field_metadata)
-    added_on: Optional[datetime.datetime] = field(default=None, metadata=field_metadata)
-    added_by_id: Optional[str] = field(default=None, metadata=field_metadata)
-    removed: Optional[bool] = field(default=None, metadata=field_metadata)
-    removed_on: Optional[datetime.datetime] = field(default=None, metadata=field_metadata)
-    removed_by_id: Optional[str] = field(default=None, metadata=field_metadata)
-    removal_reason: Optional[str] = field(default=None, metadata=field_metadata)
-    category: Optional[str] = field(default=None, metadata=field_metadata)
-
-
 @handle_camel_case_args
 @dataclass_json(letter_case=LetterCase.CAMEL)
 @dataclass(unsafe_hash=True, repr=False)

gs_quant/target/indices.py

@@ -426,6 +426,7 @@ class CustomBasketsRebalanceInputs(Base):
     allow_in_position_rebalance: Optional[bool] = field(default=False, metadata=field_metadata)
     action_date: Optional[datetime.date] = field(default=None, metadata=field_metadata)
     preferred_risk_model: Optional[str] = field(default=None, metadata=field_metadata)
+    on_behalf_of: Optional[str] = field(default=None, metadata=field_metadata)
     name: Optional[str] = field(default=None, metadata=name_metadata)
 
 

gs_quant/target/positions_v2_pricing.py

@@ -30,6 +30,7 @@ class PositionsPricingParameters(Base):
     weighting_strategy: str = field(default=None, metadata=field_metadata)
     carryover_positions_for_missing_dates: Optional[bool] = field(default=False, metadata=field_metadata)
     should_reweight: Optional[bool] = field(default=False, metadata=field_metadata)
+    allow_fractional_shares: Optional[bool] = field(default=True, metadata=field_metadata)
     name: Optional[str] = field(default=None, metadata=name_metadata)
 
 
@@ -40,6 +41,7 @@ class PositionsRequest(Base):
     asset_id: str = field(default=None, metadata=field_metadata)
     weight: Optional[float] = field(default=None, metadata=field_metadata)
     quantity: Optional[float] = field(default=None, metadata=field_metadata)
+    notional: Optional[float] = field(default=None, metadata=field_metadata)
     tags: Optional[Tuple[PositionTag, ...]] = field(default=None, metadata=field_metadata)
     name: Optional[str] = field(default=None, metadata=name_metadata)
 

gs_quant/target/secmaster.py

@@ -287,6 +287,7 @@ class SecMasterGetRequestPathSchema(Base):
     fields: Optional[Tuple[str, ...]] = field(default=None, metadata=field_metadata)
     as_of_time: Optional[Tuple[datetime.datetime, ...]] = field(default=None, metadata=field_metadata)
     is_primary: Optional[Tuple[str, ...]] = field(default=None, metadata=field_metadata)
+    all_listings: Optional[Tuple[str, ...]] = field(default=None, metadata=field_metadata)
     effective_date: Optional[Tuple[datetime.date, ...]] = field(default=None, metadata=field_metadata)
     limit: Optional[Tuple[str, ...]] = field(default=None, metadata=field_metadata)
     offset: Optional[Tuple[str, ...]] = field(default=None, metadata=field_metadata)

gs_quant/test/timeseries/test_measures_reports.py

@@ -64,7 +64,7 @@
     {
         'date': '2020-11-23',
         'reportId': 'report_id',
-        'factor': 'factor_id',
+        'factor': 'Factor Name',
         'factorCategory': 'CNT',
         'pnl': 11.23,
         'exposure': -11.23,
@@ -73,7 +73,7 @@
     {
         'date': '2020-11-24',
         'reportId': 'report_id',
-        'factor': 'factor_id',
+        'factor': 'Factor Name',
         'factorCategory': 'CNT',
         'pnl': 11.24,
         'exposure': -11.24,
@@ -82,11 +82,38 @@
     {
         'date': '2020-11-25',
         'reportId': 'report_id',
-        'factor': 'factor_id',
+        'factor': 'Factor Name',
         'factorCategory': 'CNT',
         'pnl': 11.25,
         'exposure': -11.25,
         'proportionOfRisk': 3
+    },
+    {
+        'date': '2020-11-23',
+        'reportId': 'report_id',
+        'factor': 'Total',
+        'factorCategory': 'CNT',
+        'pnl': 19.23,
+        'exposure': -11.23,
+        'proportionOfRisk': 1
+    },
+    {
+        'date': '2020-11-24',
+        'reportId': 'report_id',
+        'factor': 'Total',
+        'factorCategory': 'CNT',
+        'pnl': 14.24,
+        'exposure': -11.24,
+        'proportionOfRisk': 2
+    },
+    {
+        'date': '2020-11-25',
+        'reportId': 'report_id',
+        'factor': 'Total',
+        'factorCategory': 'CNT',
+        'pnl': 21.25,
+        'exposure': -11.25,
+        'proportionOfRisk': 3
     }
 ]
 
@@ -499,7 +526,7 @@ def test_factor_pnl():
         factor_data_copy.insert(0, {
             'date': '2020-11-22',
             'reportId': 'report_id',
-            'factor': 'factor_id',
+            'factor': 'Factor Name',
             'factorCategory': 'CNT',
             'pnl': 0,
             'exposure': -11.23,
@@ -519,6 +546,8 @@ def test_factor_pnl():
 def test_factor_pnl_percent():
     replace = Replacer()
 
+    aum = {'2020-11-22': 200, '2020-11-23': 400, '2020-11-24': 400, '2020-11-25': 400}
+
     # mock getting risk model entity()
     mock = replace('gs_quant.api.gs.risk_models.GsRiskModelApi.get_risk_model', Mock())
     mock.return_value = risk_model
@@ -553,10 +582,11 @@ def test_factor_pnl_percent():
         'factorCategory': 'Factor Name'
     }]
 
-    pnl = {entry['date']: entry['pnl'] for entry in factor_data}
-    aum = {'2020-11-22': 200, '2020-11-23': 400, '2020-11-24': 400, '2020-11-25': 400}
+    mock = replace('gs_quant.markets.report.PerformanceReport.get_aum', Mock())
+    mock.return_value = aum
+
     # Each team uses pnl today/aum yesterday
-    expected_values = compute_geometric_aggregation_calculations(aum, pnl, list(aum.keys()))
+    expected_values = [0, 0, 3.022984004, 6.0231948]
 
     with DataContext(datetime.date(2020, 11, 23), datetime.date(2020, 11, 25)):
         # mock getting aum source
@@ -1074,7 +1104,7 @@ def test_pnl_percent():
 
     # mock PerformanceReport.get_pnl()
     mock = replace('gs_quant.markets.report.PerformanceReport.get_pnl', Mock())
-    mock.return_value = pandas.DataFrame.from_dict({'date': pnl.keys(), 'pnl': pnl.values()})
+    mock.return_value = pandas.DataFrame.from_dict({'date': list(pnl.keys()), 'pnl': list(pnl.values())})
 
     mock = replace('gs_quant.markets.report.PerformanceReport.get_aum_source', Mock())
     mock.return_value = RiskAumSource.Long
@@ -1083,6 +1113,9 @@ def test_pnl_percent():
     mock = replace('gs_quant.markets.portfolio_manager.PortfolioManager.get_performance_report', Mock())
     mock.return_value = PerformanceReport(id='ID')
 
+    mock = replace('gs_quant.markets.report.PerformanceReport.get_aum', Mock())
+    mock.return_value = aum
+
     # Each team uses pnl today/aum yesterday
     expected_values = compute_geometric_aggregation_calculations(aum, pnl, list(aum.keys()))
 

gs_quant/timeseries/measures_reports.py

@@ -18,6 +18,7 @@
 from typing import Optional
 
 import pandas as pd
+import numpy as np
 from pandas.tseries.offsets import BDay
 from pydash import decapitalize
 
@@ -352,14 +353,16 @@ def _get_factor_data(report_id: str, factor_name: str, query_type: QueryType, un
     col_name = query_type.value.replace(' ', '')
     col_name = decapitalize(col_name)
     data_type = decapitalize(col_name[6:]) if col_name.startswith('factor') else col_name
-
+    include_total_factor = factor_name != 'Total' and unit == Unit.PERCENT and query_type == QueryType.FACTOR_PNL
+    factors_to_query = [factor_name, 'Total'] if include_total_factor else [factor_name]
     factor_data = report.get_results(
-        factors=[factor_name],
+        factors=factors_to_query,
         start_date=DataContext.current.start_date,
         end_date=DataContext.current.end_date,
         return_format=ReturnFormat.JSON
     )
-    factor_data = [d for d in factor_data if d.get(data_type) is not None]
+    total_data = [d for d in factor_data if d.get(data_type) is not None and d.get('factor') == 'Total']
+    factor_data = [d for d in factor_data if d.get(data_type) is not None and d.get('factor') == factor_name]
     if unit == Unit.PERCENT:
         if report.position_source_type != PositionSourceType.Portfolio:
             raise MqValueError('Unit can only be set to percent for portfolio reports')
@@ -372,9 +375,11 @@ def _get_factor_data(report_id: str, factor_name: str, query_type: QueryType, un
             aum_as_dict = performance_report.get_aum(start_date=prev_business_date(start_date), end_date=end_date)
             aum_df = pd.DataFrame(aum_as_dict.items(), columns=['date', 'aum'])
             pnl_df = pd.DataFrame(factor_data)[['date', 'pnl']]
+            total_returns_df = pd.DataFrame(total_data)[['date', 'pnl']]
+            total_returns_df = total_returns_df.rename(columns={'pnl': 'totalPnl'})
+            pnl_df = pd.merge(pnl_df, total_returns_df, how='inner', on=['date'])
             is_first_data_point_on_start_date = pnl_df['date'].iloc[[0]].values[0] == start_date.strftime('%Y-%m-%d')
-            return_series = _generate_daily_returns(aum_df, pnl_df, 'aum', 'pnl', is_first_data_point_on_start_date)
-            return geometrically_aggregate(return_series).multiply(100)
+            return _generate_daily_returns(aum_df, pnl_df, 'aum', 'pnl', is_first_data_point_on_start_date)
         else:
             aum = performance_report.get_aum(start_date=start_date, end_date=end_date)
             for data in factor_data:
@@ -421,5 +426,35 @@ def _generate_daily_returns(aum_df: pd.DataFrame, pnl_df: pd.DataFrame, aum_col_
     df.sort_index(inplace=True)
     df.loc[:, [aum_col_key]] = df.loc[:, [aum_col_key]].fillna(method='ffill')
     df['return'] = df[pnl_col_key].div(df[aum_col_key].shift(1))
+    if 'totalPnl' in list(df.columns.values):
+        df['totalPnl'] = df['totalPnl'].div(df[aum_col_key].shift(1))
+        df = df.fillna(0)
+        df['return'] = __smooth_percent_returns(df['return'].to_numpy(), df['totalPnl'].to_numpy()).tolist()
     return_series = pd.Series(df['return'], name="return").dropna()
     return return_series
+
+
+def __smooth_percent_returns(daily_factor_returns: np.array, daily_total_returns: np.array) -> np.array:
+    """
+    When attribution (in weights) are decomposed among multiple factors (like a group of risk model factors or
+    categories), simple geometric aggregation will not preserve additivity. In other words, the geometric sum of factor
+    PnL from Factor and Specific will NOT add up to the factor PnL from Total. The Carino log linking formula
+    calculates the coefficients for each node, and after that multiplication is done, the results can be aggregated
+    to calculate cumulative PnL:
+    https://rdrr.io/github/R-Finance/PortfolioAttribution/man/Carino.html
+
+    βt = A * αt
+    where βt is the linking coefficient on date t
+    where A is the log scaling factor (which is constant throughout the timeseries)
+    where αt is the perturbation factor on date t
+
+    A = (Rp - Rb) / (ln(1 + Rp) - ln(1 + Rb)) where Rp is the total portfolio return ; Rb is the total benchmark return
+    αt = (ln(1 + Rpt) - ln(1 + Rbt)) / (Rpt - Rbt) where Rpt is portfolio return on t and Rbp is benchmark return on t
+
+    For this use case benchmark returns are set to 0 for every day.
+    """
+    total_return = np.prod(daily_total_returns + 1) - 1
+    log_scaling_factor = (total_return) / (np.log(1 + total_return)) if total_return != 0 else 1
+    perturbation_factors = np.log(1 + daily_total_returns) / daily_total_returns
+    perturbation_factors = np.nan_to_num(perturbation_factors, nan=1)
+    return np.cumsum(daily_factor_returns * log_scaling_factor * perturbation_factors * 100)