Merge pull request #79 from thomaswmorris/logging

Add logging

Author: jennmald

src/blop/__init__.py

@@ -1,5 +1,15 @@
+import logging
+
 from . import utils  # noqa F401
 from ._version import __version__, __version_tuple__  # noqa: F401
 from .agent import Agent  # noqa F401
 from .dofs import DOF  # noqa F401
 from .objectives import Objective  # noqa F401
+
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s.%(msecs)03d %(levelname)s: %(message)s",
+    datefmt="%Y-%m-%d %H:%M:%S",
+)
+
+logger = logging.getLogger("maria")

src/blop/agent.py

@@ -35,6 +35,8 @@
 from .objectives import Objective, ObjectiveList
 from .plans import default_acquisition_plan
 
+logger = logging.getLogger("maria")
+
 warnings.filterwarnings("ignore", category=botorch.exceptions.warnings.InputDataWarning)
 
 mpl.rc("image", cmap="coolwarm")
@@ -382,7 +384,7 @@ def tell(
                         t0 = ttime.monotonic()
                         train_model(obj.model)
                         if self.verbose:
-                            print(f"trained model '{obj.name}' in {1e3 * (ttime.monotonic() - t0):.00f} ms")
+                            logger.debug(f"trained model '{obj.name}' in {1e3 * (ttime.monotonic() - t0):.00f} ms")
 
                     else:
                         train_model(obj.model, hypers=cached_hypers)
@@ -432,7 +434,7 @@ def learn(
 
         for i in range(iterations):
             if self.verbose:
-                print(f"running iteration {i + 1} / {iterations}")
+                logger.info(f"running iteration {i + 1} / {iterations}")
             for single_acqf in np.atleast_1d(acqf):
                 res = self.ask(n=n, acqf=single_acqf, upsample=upsample, route=route, **acqf_kwargs)
                 new_table = yield from self.acquire(res["points"])
@@ -761,7 +763,7 @@ def _train_all_models(self, **kwargs):
                 train_model(obj.validity_conjugate_model)
 
         if self.verbose:
-            print(f"trained models in {ttime.monotonic() - t0:.01f} seconds")
+            logger.info(f"trained models in {ttime.monotonic() - t0:.01f} seconds")
 
         self.n_last_trained = len(self._table)
 
@@ -861,15 +863,11 @@ def _set_hypers(self, hypers):
         self.validity_constraint.load_state_dict(hypers["validity_constraint"])
 
     def constraint(self, x):
-        p = torch.ones(x.shape[:-1])
+        log_p = torch.zeros(x.shape[:-1])
         for obj in self.objectives(active=True):
-            # if the constraint is non-trivial
-            if obj.constraint is not None:
-                p *= obj.constraint_probability(x)
-            # if the validity constaint is non-trivial
-            if obj.validity_conjugate_model is not None:
-                p *= obj.validity_constraint(x)
-        return p  # + 1e-6 * normalize(x, self.sample_domain).square().sum(axis=-1)
+            log_p += obj.log_total_constraint(x)
+
+        return log_p.exp()  # + 1e-6 * normalize(x, self.sample_domain).square().sum(axis=-1)
 
     @property
     def hypers(self) -> dict:

src/blop/bayesian/models.py

@@ -6,13 +6,22 @@
 from . import kernels
 
 
-def train_model(model, hypers=None, **kwargs):
+def train_model(model, hypers=None, max_fails=4, **kwargs):
     """Fit all of the agent's models. All kwargs are passed to `botorch.fit.fit_gpytorch_mll`."""
-    if hypers is not None:
-        model.load_state_dict(hypers)
-    else:
-        botorch.fit.fit_gpytorch_mll(gpytorch.mlls.ExactMarginalLogLikelihood(model.likelihood, model), **kwargs)
-    model.trained = True
+    fails = 0
+    while True:
+        try:
+            if hypers is not None:
+                model.load_state_dict(hypers)
+            else:
+                botorch.fit.fit_gpytorch_mll(gpytorch.mlls.ExactMarginalLogLikelihood(model.likelihood, model), **kwargs)
+            model.trained = True
+            return
+        except Exception as e:
+            if fails < max_fails:
+                fails += 1
+            else:
+                raise e
 
 
 def construct_single_task_model(X, y, skew_dims=None, min_noise=1e-6, max_noise=1e0):

src/blop/objectives.py

@@ -155,6 +155,19 @@ def constrain(self, y):
         else:
             return np.array([value in self.constraint for value in np.atleast_1d(y)])
 
+    def log_total_constraint(self, x):
+
+        log_p = 0
+        # if you have a constraint
+        if self.constraint is not None:
+            log_p += self.constraint_probability(x).log()
+
+        # if the validity constaint is non-trivial
+        if self.validity_conjugate_model is not None:
+            log_p += self.validity_constraint(x).log()
+
+        return log_p
+
     @property
     def _trust_domain(self):
         if self.trust_domain is None:

src/blop/plotting.py

@@ -33,7 +33,7 @@ def _plot_fitness_objs_one_dof(agent, size=16, lw=1e0):
     test_model_inputs = agent.dofs(active=True).transform(test_inputs)
 
     for obj_index, obj in enumerate(fitness_objs):
-        obj_values = agent.train_targets()(obj.name).numpy()
+        obj_values = agent.train_targets()[obj.name].numpy()
 
         color = DEFAULT_COLOR_LIST[obj_index]