physbo 2.0.0__cp310-cp310-macosx_12_0_arm64.whl

physbo/misc/set_config.py

@@ -0,0 +1,299 @@
+# SPDX-License-Identifier: MPL-2.0
+# Copyright (C) 2020- The University of Tokyo
+#
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at https://mozilla.org/MPL/2.0/.
+
+import numpy as np
+import configparser
+
+
+class set_config:
+    def __init__(self, search_config=None, learning_config=None):
+        """
+        Setting configuration for search and learning.
+
+        Parameters
+        ----------
+        search_config: physbo.misc.search object
+        learning_config: physbo.misc.learning object
+
+        """
+        if search_config is None:
+            search_config = search()
+        self.search = search_config
+
+        if learning_config is None:
+            learning_config = adam()
+        self.learning = learning_config
+
+    def show(self):
+        """
+        Showing information of search and learning objects.
+
+        Returns
+        -------
+
+        """
+        self.search.show()
+        self.learning.show()
+
+    def load(self, file_name="config.ini"):
+        """
+        Loading information of configuration.
+
+        Parameters
+        ----------
+        file_name: str
+        An input file name of configuration.
+        Returns
+        -------
+
+        """
+        config = configparser.SafeConfigParser()
+        config.read(file_name)
+
+        search_config = search()
+        self.search = search_config
+        self.search.load(config)
+
+        temp_dict = config._sections["learning"]
+        method = temp_dict.get("method", "adam")
+
+        if method == "adam":
+            learning_config = adam()
+            self.learning = learning_config
+            self.learning.load(config)
+
+        if method in ("bfgs", "batch"):
+            learning_config = batch()
+            self.learning = learning_config
+            self.learning.load(config)
+
+
+class search:
+    def __init__(self):
+        self.multi_probe_num_sampling = 20
+        self.alpha = 1.0
+
+    def load(self, config):
+        """
+        Loading information of configuration from config._sectoins['search'].
+
+        Parameters
+        ----------
+        config: physbo.misc.set_config object
+
+        Returns
+        -------
+
+        """
+        temp_dict = config._sections["search"]
+        self.multi_probe_num_sampling = int(
+            temp_dict.get("multi_probe_num_sampling", 20)
+        )
+        self.alpha = np.float64(temp_dict.get("alpha", 1.0))
+
+    def show(self):
+        """
+        Showing information about search object.
+
+        Returns
+        -------
+
+        """
+        print("(search)")
+        print("multi_probe_num_sampling: ", self.multi_probe_num_sampling)
+        print("alpha: ", self.alpha)
+        print("\n")
+
+
+class learning(object):
+    def __init__(self):
+        self.is_disp = True
+        self.num_disp = 10
+        self.num_init_params_search = 20
+        self.method = "adam"
+
+    def show(self):
+        """
+        Showing information about learning object.
+
+        Returns
+        -------
+
+        """
+        print("( learning )")
+        print("method : ", self.method)
+        print("is_disp: ", self.is_disp)
+        print("num_disp: ", self.num_disp)
+        print("num_init_params_search: ", self.num_init_params_search)
+
+    def load(self, config):
+        """
+        Loading information of configuration from config._sectoins['learning'].
+
+        Parameters
+        ----------
+        config: physbo.misc.set_config object
+
+
+        Returns
+        -------
+
+        """
+        temp_dict = config._sections["learning"]
+        self.method = temp_dict.get("method", "adam")
+        self.is_disp = boolean(temp_dict.get("is_disp", True))
+        self.num_disp = int(temp_dict.get("num_disp", 10))
+        self.num_init_params_search = int(temp_dict.get("num_init_params_search", 20))
+
+
+class batch(learning):
+    def __init__(self):
+        super(batch, self).__init__()
+        self.method = "bfgs"
+        self.max_iter = 200
+        self.max_iter_init_params_search = 20
+        self.batch_size = 5000
+
+    def show(self):
+        """
+        Showing information about configuration about batch object.
+
+        Returns
+        -------
+
+        """
+        super(batch, self).show()
+        print("max_iter: ", self.max_iter)
+        print("max_iter_init_params_search: ", self.max_iter_init_params_search)
+        print("batch_size: ", self.batch_size)
+
+    def load(self, config):
+        """
+        Loading information of configuration from config._sectoins['batch'].
+
+        Parameters
+        ----------
+        config: physbo.misc.set_config object
+
+        Returns
+        -------
+
+        """
+        super(batch, self).load(config)
+        temp_dict = config._sections["batch"]
+        self.max_iter = int(temp_dict.get("max_iter", 200))
+        self.max_iter_init_params_search = int(
+            temp_dict.get("max_iter_init_params_search", 20)
+        )
+        self.batch_size = int(temp_dict.get("batch_size", 5000))
+
+
+class online(learning):
+    def __init__(self):
+        super(online, self).__init__()
+        self.max_epoch = 500
+        self.max_epoch_init_params_search = 50
+        self.batch_size = 64
+        self.eval_size = 5000
+
+    def show(self):
+        """
+        Showing information about configuration about online object.
+
+        Returns
+        -------
+
+        """
+        super(online, self).show()
+        print("max_epoch: ", self.max_epoch)
+        print("max_epoch_init_params_search: ", self.max_epoch_init_params_search)
+        print("batch_size: ", self.batch_size)
+        print("eval_size: ", self.eval_size)
+
+    def load(self, config):
+        """
+        Loading information of configuration from config._sectoins['online'].
+
+        Parameters
+        ----------
+        config: physbo.misc.set_config object
+
+
+        Returns
+        -------
+
+        """
+        super(online, self).load(config)
+        temp_dict = config._sections["online"]
+        self.max_epoch = int(temp_dict.get("max_epoch", 1000))
+        self.max_epoch_init_params_search = int(
+            temp_dict.get("max_epoch_init_params_search", 50)
+        )
+        self.batch_size = int(temp_dict.get("batch_size", 64))
+        self.eval_size = int(temp_dict.get("eval_size", 5000))
+
+
+class adam(online):
+    def __init__(self):
+        super(adam, self).__init__()
+        self.method = "adam"
+        self.alpha = 0.001
+        self.beta = 0.9
+        self.gamma = 0.999
+        self.epsilon = 1e-6
+
+    def show(self):
+        """
+        Showing information about configuration about adam object.
+
+        Returns
+        -------
+
+        """
+        super(adam, self).show()
+        print("alpha = ", self.alpha)
+        print("beta = ", self.beta)
+        print("gamma = ", self.gamma)
+        print("epsilon = ", self.epsilon)
+        print("\n")
+
+    def load(self, config):
+        """
+        Loading information of configuration from config._sectoins['adam'].
+
+        Parameters
+        ----------
+        config: physbo.misc.set_config object
+
+        Returns
+        -------
+
+        """
+        super(adam, self).load(config)
+        temp_dict = config._sections["adam"]
+        self.alpha = np.float64(temp_dict.get("alpha", 0.001))
+        self.beta = np.float64(temp_dict.get("beta", 0.9))
+        self.gamma = np.float64(temp_dict.get("gamma", 0.9999))
+        self.epsilon = np.float64(temp_dict.get("epsilon", 1e-6))
+
+
+def boolean(str):
+    """
+    Return boolean.
+
+    Parameters
+    ----------
+    str: str or boolean
+
+    Returns
+    -------
+    True or False
+    """
+    if str == "True" or str is True:
+        return True
+    else:
+        return False

physbo/opt/__init__.py

@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: MPL-2.0
+# Copyright (C) 2020- The University of Tokyo
+#
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at https://mozilla.org/MPL/2.0/.
+
+from .adam import adam

physbo/opt/adam.py

@@ -0,0 +1,107 @@
+# SPDX-License-Identifier: MPL-2.0
+# Copyright (C) 2020- The University of Tokyo
+#
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at https://mozilla.org/MPL/2.0/.
+
+import numpy as np
+
+
+class adam:
+    """
+    Optimizer of f(x) with the adam method
+
+    Attributes
+    ==========
+    params: numpy.ndarray
+        current input, x
+    nparams: int
+        dimension
+    grad: function
+        gradient function, g(x) = f'(x)
+    m: numpy.ndarray
+    v: numpy.ndarray
+    epoch: int
+        the number of update already done
+    max_epoch: int
+        the maximum number of update
+    alpha: float
+    beta: float
+    gamma: float
+    epsilon: float
+    """
+
+    def __init__(self, params, grad, options={}):
+        """
+
+        Parameters
+        ==========
+        params:
+        grad:
+        options: dict
+            Hyperparameters for the adam method
+
+                - "alpha" (default: 0.001)
+                - "beta" (default: 0.9)
+                - "gamma" (default: 0.9999)
+                - "epsilon" (default: 1e-8)
+                - "max_epoch" (default: 4000)
+        """
+        self.grad = grad
+        self.params = params
+        self.nparams = params.shape[0]
+        self._set_options(options)
+        self.m = np.zeros(self.nparams)
+        self.v = np.zeros(self.nparams)
+        self.epoch = 0
+
+    def set_params(self, params):
+        self.params = params
+
+    def update(self, params, *args, **kwargs):
+        """
+        calculates the updates of params
+
+        Parameters
+        ==========
+        params: numpy.ndarray
+            input
+        args:
+            will be passed to self.grad
+        kwargs:
+            will be passed to self.grad
+
+        Returns
+        =======
+        numpy.ndarray
+            update of params
+        """
+        g = self.grad(params, *args, **kwargs)
+        self.m = self.m * self.beta + g * (1 - self.beta)
+        self.v = self.v * self.gamma + g**2 * (1 - self.gamma)
+        hat_m = self.m / (1 - self.beta ** (self.epoch + 1))
+        hat_v = self.v / (1 - self.gamma ** (self.epoch + 1))
+        self.epoch += 1
+        return -self.alpha * hat_m / (np.sqrt(hat_v) + self.epsilon)
+
+    def run(self, *args, **kwargs):
+        params = self.params
+        for epoch in range(self.max_epoch):
+            update = self.update(params, *args, **kwargs)
+            params += update
+
+    def _set_options(self, options):
+        """
+        set hyperparameters for the method
+
+        Parameters
+        ==========
+        options: dict
+
+        """
+        self.alpha = options.get("alpha", 0.001)
+        self.beta = options.get("beta", 0.9)
+        self.gamma = options.get("gamma", 0.9999)
+        self.epsilon = options.get("epsilon", 1e-8)
+        self.max_epoch = options.get("max_epoch", 4000)

physbo/predictor.py

@@ -0,0 +1,261 @@
+# SPDX-License-Identifier: MPL-2.0
+# Copyright (C) 2020- The University of Tokyo
+#
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at https://mozilla.org/MPL/2.0/.
+
+import pickle as pickle
+import numpy as np
+from physbo import gp
+
+
+class base_predictor(object):
+    """
+    Base predictor is defined in this class.
+
+    """
+
+    def __init__(self, config, model=None):
+        """
+
+        Parameters
+        ----------
+        config: set_config object (physbo.misc.set_config)
+        model: model object
+            A default model is set as gp.core.model
+        """
+
+        self.config = config
+        self.model = model
+        if self.model is None:
+            self.model = gp.core.model(
+                cov=gp.cov.gauss(num_dim=None, ard=False),
+                mean=gp.mean.const(),
+                lik=gp.lik.gauss(),
+            )
+
+    def fit(self, *args, **kwds):
+        """
+
+        Default fit function.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def prepare(self, *args, **kwds):
+        """
+
+        Default prepare function.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def delete_stats(self, *args, **kwds):
+        """
+
+        Default function to delete status.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def get_basis(self, *args, **kwds):
+        """
+
+        Default function to get basis
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def get_post_fmean(self, *args, **kwds):
+        """
+
+        Default function to get a mean value of the score.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def get_post_fcov(self, *args, **kwds):
+        """
+
+        Default function to get a covariance of the score.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def get_post_params(self, *args, **kwds):
+        """
+
+        Default function to get parameters.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def get_post_samples(self, *args, **kwds):
+        """
+
+        Default function to get samples.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def get_predict_samples(self, *args, **kwds):
+        """
+
+        Default function to get prediction variables of samples.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def get_post_params_samples(self, *args, **kwds):
+        """
+
+        Default function to get parameters of samples.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def update(self, *args, **kwds):
+        """
+
+        Default function to update variables.
+        This function must be overwritten in each model.
+
+        Parameters
+        ----------
+        args
+        kwds
+
+        Returns
+        -------
+
+        """
+        raise NotImplementedError
+
+    def save(self, file_name):
+        """
+
+        Default function to save information by using pickle.dump function.
+        The protocol version is set as 3.
+
+        Parameters
+        ----------
+        file_name: str
+            A file name to save self.__dict__ object.
+
+        Returns
+        -------
+
+        """
+        with open(file_name, "wb") as f:
+            pickle.dump(self.__dict__, f, 4)
+
+    def load(self, file_name):
+        """
+
+        Default function to load variables.
+        The information is updated using self.update function.
+
+        Parameters
+        ----------
+        file_name: str
+            A file name to load variables from the file.
+
+        Returns
+        -------
+
+        """
+        with open(file_name, "rb") as f:
+            tmp_dict = pickle.load(f)
+            self.config = tmp_dict["config"]
+            self.model = tmp_dict["model"]

physbo/search/__init__.py

@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: MPL-2.0
+# Copyright (C) 2020- The University of Tokyo
+#
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at https://mozilla.org/MPL/2.0/.
+
+from . import score
+from . import discrete
+from . import score_multi
+from . import discrete_multi

physbo/search/discrete/__init__.py

@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: MPL-2.0
+# Copyright (C) 2020- The University of Tokyo
+#
+# This Source Code Form is subject to the terms of the Mozilla Public
+# License, v. 2.0. If a copy of the MPL was not distributed with this
+# file, You can obtain one at https://mozilla.org/MPL/2.0/.
+
+from .policy import policy
+from .results import history
+
+# __all__ = ["policy", "history"]