Merge pull request #261 from ta-oliver/enable_allocation_limits_for_api

Add limits by default to the API position calculations.

infertrade/api.py

@@ -25,7 +25,7 @@
 
 # InferTrade packages
 from infertrade.algos import algorithm_functions, ta_adaptor
-from infertrade.utilities.operations import ReturnsFromPositions
+from infertrade.utilities.operations import ReturnsFromPositions, restrict_allocation, limit_allocation
 from infertrade.PandasEnum import PandasEnum
 
 
@@ -159,13 +159,14 @@ def _get_raw_callable(name_of_strategy_or_signal: str) -> callable:
 
     @staticmethod
     def calculate_allocations(
-        df: pd.DataFrame, name_of_strategy: str, name_of_price_series: str = PandasEnum.MID.value
+        df: pd.DataFrame, name_of_strategy: str, name_of_price_series: str = PandasEnum.MID.value,
+            allocation_lower_limit: float = -1.0, allocation_upper_limit: float = 1.0
     ) -> pd.DataFrame:
         """Calculates the allocations using the supplied strategy."""
         if name_of_price_series is not PandasEnum.MID.value:
             df[PandasEnum.MID.value] = df[name_of_price_series]
         rule_function = Api._get_raw_callable(name_of_strategy)
-        df_with_positions = rule_function(df)
+        df_with_positions = limit_allocation(rule_function(df), allocation_lower_limit, allocation_upper_limit)
         return df_with_positions
 
     @staticmethod
@@ -176,10 +177,10 @@ def calculate_returns(df: pd.DataFrame) -> pd.DataFrame:
 
     @staticmethod
     def calculate_allocations_and_returns(
-        df: pd.DataFrame, name_of_strategy: str, name_of_price_series: str = PandasEnum.MID.value
+        df: pd.DataFrame, name_of_strategy: str, name_of_price_series: str = PandasEnum.MID.value, *args, **kwargs
     ) -> pd.DataFrame:
         """Calculates the returns using the supplied strategy."""
-        df_with_positions = Api.calculate_allocations(df, name_of_strategy, name_of_price_series)
+        df_with_positions = Api.calculate_allocations(df, name_of_strategy, name_of_price_series, *args, **kwargs)
         df_with_returns = ReturnsFromPositions().transform(df_with_positions)
         return df_with_returns
 

tests/test_allocations.py

@@ -37,6 +37,7 @@
 from infertrade.algos import algorithm_functions
 from infertrade.algos.community import allocations
 from infertrade.algos.community.allocations import create_infertrade_export_allocations
+from infertrade.api import Api
 
 
 def test_under_minimum_length_to_calculate():
@@ -115,3 +116,10 @@ def test_create_infertrade_export_allocations():
     """Checks that a valid dictionary can be created."""
     dictionary_algorithms = create_infertrade_export_allocations()
     assert isinstance(dictionary_algorithms, dict)  # could add checks for contents too
+
+
+def test_all_allocations_list_required_series():
+    """Checks that all allocation rules list required series."""
+    for ii_rule in Api.available_algorithms(filter_by_category="allocation"):
+        assert isinstance(Api.required_inputs_for_algorithm(ii_rule), list)
+

tests/test_api.py

@@ -19,6 +19,8 @@
 """
 
 # External imports
+import copy
+
 import pandas as pd
 import pytest
 
@@ -54,7 +56,7 @@ def test_get_available_algorithms(algorithm):
     assert Api.determine_package_of_algorithm(algorithm) in Api.available_packages()
     try:
         Api.determine_package_of_algorithm("not_available_algo")
-    except (NameError):
+    except NameError:
         pass
 
     inputs = Api.required_inputs_for_algorithm(algorithm)
@@ -172,8 +174,8 @@ def test_return_representations(algorithm):
         )
     for representation in dict_of_properties[algorithm]["available_representation_types"]:
         assert (
-            returned_representations[representation]
-            == dict_of_properties[algorithm]["available_representation_types"][representation]
+                returned_representations[representation]
+                == dict_of_properties[algorithm]["available_representation_types"][representation]
         )
 
     # Check if the if the function returns the correct representation when given a string
@@ -185,8 +187,8 @@ def test_return_representations(algorithm):
                 type(returned_representations),
             )
         assert (
-            returned_representations[representation]
-            == dict_of_properties[algorithm]["available_representation_types"][representation]
+                returned_representations[representation]
+                == dict_of_properties[algorithm]["available_representation_types"][representation]
         )
 
     # Check if the function returns the correct representations when given a list
@@ -198,8 +200,8 @@ def test_return_representations(algorithm):
         )
     for representation in algorithm_representations:
         assert (
-            returned_representations[representation]
-            == dict_of_properties[algorithm]["available_representation_types"][representation]
+                returned_representations[representation]
+                == dict_of_properties[algorithm]["available_representation_types"][representation]
         )
 
 
@@ -357,3 +359,23 @@ def test_export_cross_prediction():
     )
 
     assert isinstance(sorted_dict, dict)
+
+
+@pytest.mark.parametrize("test_df", test_dfs)
+def test_allocation_limit(test_df):
+    """Test used to see if calculated allocation values are inside of specified limit"""
+
+    test_df_copy = copy.deepcopy(test_df)
+    df_with_allocations = Api.calculate_allocations(
+        df=test_df_copy, name_of_strategy=available_allocation_algorithms[0], name_of_price_series="close",
+        allocation_lower_limit=0, allocation_upper_limit=0
+    )
+    if not all(df_with_allocations["allocation"] == 0.0):
+        raise ValueError("Allocation limits breached")
+
+    df_with_allocations = Api.calculate_allocations(
+        df=test_df_copy, name_of_strategy=available_allocation_algorithms[0], name_of_price_series="close",
+        allocation_lower_limit=-0.1, allocation_upper_limit=0.1
+    )
+    if any(-0.1 > df_with_allocations["allocation"]) or any(df_with_allocations["allocation"] > 0.1):
+        raise ValueError("Allocation limits breached")