per-asset bounds

docs/EfficientFrontier.rst

@@ -43,8 +43,8 @@ magnitude, I will definitely consider switching.
 
             .. note::
 
-                As a rule of thumb, any parameters that can apply to all optimisers
+                As of v0.5.0, you can pass a collection (list or tuple) of (min, max) pairs
-                are instance variables (passed when you are initialising the object).
+                representing different bounds for different assets.
 
 .. caution::
 

pypfopt/base_optimizer.py

@@ -80,16 +80,17 @@ class BaseScipyOptimizer(BaseOptimizer):
     - ``n_assets`` - int
     - ``tickers`` - str list
     - ``weights`` - np.ndarray
-    - ``bounds`` - (float tuple) list
+    - ``bounds`` - float tuple OR (float tuple) list
     - ``initial_guess`` - nnp.ndarray
     - ``constraints`` - dict list
     """
 
     def __init__(self, n_assets, tickers=None, weight_bounds=(0, 1)):
         """
-        :param weight_bounds: minimum and maximum weight of an asset, defaults to (0, 1).
+        :param weight_bounds: minimum and maximum weight of each asset OR single min/max pair
-                              Must be changed to (-1, 1) for portfolios with shorting.
+                              if all identical, defaults to (0, 1). Must be changed to (-1, 1)
-        :type weight_bounds: tuple, optional
+                              for portfolios with shorting.
+        :type weight_bounds: tuple OR tuple list, optional
         """
         super().__init__(n_assets, tickers)
         self.bounds = self._make_valid_bounds(weight_bounds)
@@ -102,21 +103,36 @@ class BaseScipyOptimizer(BaseOptimizer):
         Private method: process input bounds into a form acceptable by scipy.optimize,
         and check the validity of said bounds.
 
-        :param test_bounds: minimum and maximum weight of an asset
+        :param test_bounds: minimum and maximum weight of each asset OR single min/max pair
-        :type test_bounds: tuple
+                              if all identical, defaults to (0, 1).
-        :raises ValueError: if ``test_bounds`` is not a tuple of length two.
+        :type test_bounds: tuple OR list/tuple of tuples.
+        :raises ValueError: if ``test_bounds`` is not a tuple of length two OR a collection
+                            of pairs.
         :raises ValueError: if the lower bound is too high
         :return: a tuple of bounds, e.g ((0, 1), (0, 1), (0, 1) ...)
         :rtype: tuple of tuples
         """
-        if len(test_bounds) != 2 or not isinstance(test_bounds, tuple):
+        # If it is a collection with the right length, assume they are all bounds.
+        print(test_bounds)
+        if len(test_bounds) == self.n_assets and not isinstance(
+            test_bounds[0], (float, int)
+        ):
+            bounds = test_bounds
+        else:
+            if len(test_bounds) != 2 or not isinstance(test_bounds, tuple):
+                raise ValueError(
+                    "test_bounds must be a tuple of (lower bound, upper bound) "
+                    "OR collection of bounds for each asset"
+                )
+            bounds = (test_bounds,) * self.n_assets
+
+        # Ensure lower bound is not too high
+        if sum((0 if b[0] is None else b[0]) for b in bounds) > 1:
             raise ValueError(
-                "test_bounds must be a tuple of (lower bound, upper bound)"
+                "Lower bound is too high. Impossible to construct valid portfolio"
             )
-        if test_bounds[0] is not None:
+
-            if test_bounds[0] * self.n_assets > 1:
+        return bounds
-                raise ValueError("Lower bound is too high")
-        return (test_bounds,) * self.n_assets
 
 
 def portfolio_performance(

pypfopt/cla.py

@@ -67,14 +67,21 @@ class CLA(base_optimizer.BaseOptimizer):
             self.mean[-1, 0] += 1e-5
         self.expected_returns = self.mean.reshape((len(self.mean),))
         self.cov_matrix = np.asarray(cov_matrix)
-        if isinstance(weight_bounds[0], numbers.Real):
+
-            self.lB = np.ones(self.mean.shape) * weight_bounds[0]
+        # Bounds
+        if len(weight_bounds) == len(self.mean):
+            self.lB = np.array([b[0] for b in weight_bounds]).reshape(-1, 1)
+            self.uB = np.array([b[1] for b in weight_bounds]).reshape(-1, 1)
         else:
-            self.lB = np.array(weight_bounds[0]).reshape(self.mean.shape)
+            if isinstance(weight_bounds[0], numbers.Real):
-        if isinstance(weight_bounds[0], numbers.Real):
+                self.lB = np.ones(self.mean.shape) * weight_bounds[0]
-            self.uB = np.ones(self.mean.shape) * weight_bounds[1]
+            else:
-        else:
+                self.lB = np.array(weight_bounds[0]).reshape(self.mean.shape)
-            self.uB = np.array(weight_bounds[1]).reshape(self.mean.shape)
+            if isinstance(weight_bounds[0], numbers.Real):
+                self.uB = np.ones(self.mean.shape) * weight_bounds[1]
+            else:
+                self.uB = np.array(weight_bounds[1]).reshape(self.mean.shape)
+
         self.w = []  # solution
         self.ls = []  # lambdas
         self.g = []  # gammas

pypfopt/efficient_frontier.py

@@ -23,7 +23,7 @@ class EfficientFrontier(base_optimizer.BaseScipyOptimizer):
 
         - ``n_assets`` - int
         - ``tickers`` - str list
-        - ``bounds`` - (float tuple) list
+        - ``bounds`` - float tuple OR (float tuple) list
         - ``cov_matrix`` - pd.DataFrame
         - ``expected_returns`` - pd.Series
 
@@ -52,9 +52,10 @@ class EfficientFrontier(base_optimizer.BaseScipyOptimizer):
         :type expected_returns: pd.Series, list, np.ndarray
         :param cov_matrix: covariance of returns for each asset
         :type cov_matrix: pd.DataFrame or np.array
-        :param weight_bounds: minimum and maximum weight of an asset, defaults to (0, 1).
+        :param weight_bounds: minimum and maximum weight of each asset OR single min/max pair
-                              Must be changed to (-1, 1) for portfolios with shorting.
+                              if all identical, defaults to (0, 1). Must be changed to (-1, 1)
-        :type weight_bounds: tuple, optional
+                              for portfolios with shorting.
+        :type weight_bounds: tuple OR tuple list, optional
         :param gamma: L2 regularisation parameter, defaults to 0. Increase if you want more
                       non-negligible weights
         :type gamma: float, optional

pypfopt/value_at_risk.py

@@ -21,7 +21,7 @@ class CVAROpt(base_optimizer.BaseScipyOptimizer):
 
         - ``tickers`` - str list
         - ``returns`` - pd.DataFrame
-        - ``bounds`` - (double tuple) list
+        - ``bounds`` - float tuple OR (float tuple) list
 
     - Optimisation parameters:
 
@@ -40,10 +40,10 @@ class CVAROpt(base_optimizer.BaseScipyOptimizer):
         """
         :param returns: asset historical returns
         :type returns: pd.DataFrame
-        :param weight_bounds: minimum and maximum weight of an asset, defaults to (0, 1).
+        :param weight_bounds: minimum and maximum weight of each asset OR single min/max pair
-                              Must be changed to (-1, 1) for portfolios with shorting.
+                              if all identical, defaults to (0, 1). Must be changed to (-1, 1)
-                              For CVaR opt, this is not a hard boundary.
+                              for portfolios with shorting.
-        :type weight_bounds: tuple, optional
+        :type weight_bounds: tuple OR tuple list, optional
         :raises TypeError: if ``returns`` is not a dataframe
         """
         if not isinstance(returns, pd.DataFrame):

tests/test_base_optimizer.py

@@ -23,7 +23,7 @@ def test_custom_lower_bound():
     np.testing.assert_almost_equal(ef.weights.sum(), 1)
 
 
-def test_custom_bounds():
+def test_custom_bounds_same():
     ef = EfficientFrontier(
         *setup_efficient_frontier(data_only=True), weight_bounds=(0.03, 0.13)
     )
@@ -33,11 +33,28 @@ def test_custom_bounds():
     np.testing.assert_almost_equal(ef.weights.sum(), 1)
 
 
+def test_custom_bounds_different():
+    bounds = [(0.01, 0.13), (0.02, 0.11)] * 10
+    ef = EfficientFrontier(
+        *setup_efficient_frontier(data_only=True), weight_bounds=bounds
+    )
+    ef.max_sharpe()
+    assert (0.01 <= ef.weights[::2]).all() and (ef.weights[::2] <= 0.13).all()
+    assert (0.02 <= ef.weights[1::2]).all() and (ef.weights[1::2] <= 0.11).all()
+    np.testing.assert_almost_equal(ef.weights.sum(), 1)
+
+    bounds = ((0.01, 0.13), (0.02, 0.11)) * 10
+    assert EfficientFrontier(
+        *setup_efficient_frontier(data_only=True), weight_bounds=bounds
+    )
+
+
 def test_bounds_errors():
     with pytest.raises(ValueError):
         EfficientFrontier(
             *setup_efficient_frontier(data_only=True), weight_bounds=(0.06, 1)
         )
+
     assert EfficientFrontier(
         *setup_efficient_frontier(data_only=True), weight_bounds=(0, 1)
     )
@@ -47,6 +64,12 @@ def test_bounds_errors():
             *setup_efficient_frontier(data_only=True), weight_bounds=(0.06, 1, 3)
         )
 
+    with pytest.raises(ValueError):
+        bounds = [(0.01, 0.13), (0.02, 0.11)] * 5
+        EfficientFrontier(
+            *setup_efficient_frontier(data_only=True), weight_bounds=bounds
+        )
+
 
 def test_clean_weights():
     ef = setup_efficient_frontier()
@@ -92,7 +115,9 @@ def test_clean_weights_no_rounding():
     # in previous commits, this call would raise a ValueError
     cleaned = ef.clean_weights(rounding=None, cutoff=0)
     assert cleaned
-    np.testing.assert_array_almost_equal(np.sort(ef.weights), np.sort(list(cleaned.values())))
+    np.testing.assert_array_almost_equal(
+        np.sort(ef.weights), np.sort(list(cleaned.values()))
+    )
 
 
 def test_efficient_frontier_init_errors():

tests/test_cla.py

@@ -51,6 +51,19 @@ def test_cla_max_sharpe_short():
     assert sharpe > long_only_sharpe
 
 
+def test_cla_custom_bounds():
+    bounds = [(0.01, 0.13), (0.02, 0.11)] * 10
+    cla = CLA(*setup_cla(data_only=True), weight_bounds=bounds)
+    df = get_data()
+    cla.cov_matrix = risk_models.exp_cov(df).values
+    w = cla.min_volatility()
+    assert isinstance(w, dict)
+    assert set(w.keys()) == set(cla.tickers)
+    np.testing.assert_almost_equal(cla.weights.sum(), 1)
+    assert (0.01 <= cla.weights[::2]).all() and (cla.weights[::2] <= 0.13).all()
+    assert (0.02 <= cla.weights[1::2]).all() and (cla.weights[1::2] <= 0.11).all()
+
+
 def test_cla_min_volatility():
     cla = setup_cla()
     w = cla.min_volatility()