Merge pull request #68 from voetberg/demo_nb

Demo notebook

src/client/client.py

@@ -38,15 +38,15 @@ def parser():
     # List of metrics (cannot supply specific kwargs)
     parser.add_argument(
         "--metrics",
-        nargs="+",
+        nargs="?",
         default=list(Defaults["metrics"].keys()),
         choices=Metrics.keys(),
     )
 
     # List of plots
     parser.add_argument(
         "--plots",
-        nargs="+",
+        nargs="?",
         default=list(Defaults["plots"].keys()),
         choices=Plots.keys(),
     )

src/data/data.py

@@ -4,7 +4,7 @@
 import numpy as np
 
 from utils.config import get_item
-
+from utils.register import load_simulator
 
 class Data:
     def __init__(
@@ -19,44 +19,10 @@ def __init__(
             get_item("common", "random_seed", raise_exception=False)
         )
         self.data = self._load(path)
-        self.simulator = self._load_simulator(simulator_name, simulator_kwargs)
+        self.simulator = load_simulator(simulator_name, simulator_kwargs)
         self.prior_dist = self.load_prior(prior, prior_kwargs)
         self.n_dims = self.get_theta_true().shape[1]
 
-    def _load_simulator(self, name, simulator_kwargs):
-        try:
-            sim_location = get_item("common", "sim_location", raise_exception=False)
-            simulator_path = os.environ[f"{sim_location}:{name}"]
-        except KeyError as e:
-            raise RuntimeError(
-                f"Simulator cannot be found using env var {e}. Hint: have you registered your simulation with utils.register_simulator?"
-            )
-
-        new_class = os.path.dirname(simulator_path)
-        sys.path.insert(1, new_class)
-
-        # TODO robust error checks
-        module_name = os.path.basename(simulator_path.rstrip(".py"))
-        m = importlib.import_module(module_name)
-
-        simulator = getattr(m, name)
-
-        simulator_kwargs = simulator_kwargs if simulator_kwargs is not None else get_item("data", "simulator_kwargs", raise_exception=False)
-        simulator_kwargs = {} if simulator_kwargs is None else simulator_kwargs
-        simulator_instance = simulator(**simulator_kwargs)
-
-        if not hasattr(simulator_instance, "generate_context"):
-            raise RuntimeError(
-                "Simulator improperly formed - requires a generate_context method."
-            )
-
-        if not hasattr(simulator_instance, "simulate"):
-            raise RuntimeError(
-                "Simulator improperly formed - requires a simulate method."
-            )
-
-        return simulator_instance
-
     def _load(self, path: str):
         raise NotImplementedError
 

src/metrics/__init__.py

@@ -1,4 +1,9 @@
 from metrics.all_sbc import AllSBC
 from metrics.coverage_fraction import CoverageFraction
+from metrics.local_two_sample import LocalTwoSampleTest
 
-Metrics = {CoverageFraction.__name__: CoverageFraction, AllSBC.__name__: AllSBC}
+Metrics = {
+    CoverageFraction.__name__: CoverageFraction, 
+    AllSBC.__name__: AllSBC, 
+    "LC2ST": LocalTwoSampleTest
+}

src/metrics/all_sbc.py

@@ -1,4 +1,4 @@
-from typing import Any
+from typing import Any, Optional, Sequence
 from torch import tensor
 from sbi.analysis import run_sbc, check_sbc
 
@@ -12,16 +12,19 @@ def __init__(
         model: Any,
         data: Any,
         out_dir: str | None = None,
-        samples_per_inference=None,
+        save: bool=True,
+        use_progress_bar: Optional[bool] = None,
+        samples_per_inference: Optional[int] = None,
+        percentiles: Optional[Sequence[int]] = None,
+        number_simulations: Optional[int] = None,
     ) -> None:
-        super().__init__(model, data, out_dir)
-
-        if samples_per_inference is None:
-            self.samples_per_inference = get_item(
-                "metrics_common", "samples_per_inference", raise_exception=False
-            )
-        else:
-            self.samples_per_inference = samples_per_inference
+
+        super().__init__(model, data, out_dir,
+            save,
+            use_progress_bar,
+            samples_per_inference,
+            percentiles,
+            number_simulations)
 
     def _collect_data_params(self):
         self.thetas = tensor(self.data.get_theta_true())
@@ -41,7 +44,7 @@ def calculate(self):
             dap_samples,
             num_posterior_samples=self.samples_per_inference,
         )
-        self.output = sbc_stats
+        self.output = {key: value.numpy().tolist() for key, value in sbc_stats.items()}
         return sbc_stats
 
     def __call__(self, **kwds: Any) -> Any:

src/metrics/coverage_fraction.py

@@ -1,7 +1,7 @@
 import numpy as np
 from torch import tensor
 from tqdm import tqdm
-from typing import Any
+from typing import Any, Optional, Sequence
 
 from metrics.metric import Metric
 from utils.config import get_item
@@ -14,32 +14,22 @@ def __init__(
         self,
         model: Any,
         data: Any,
-        out_dir: str | None = None,
-        samples_per_inference=None,
-        percentiles=None,
-        progress_bar: bool = None,
+        out_dir: Optional[str] = None,
+        save: bool=True,
+        use_progress_bar: Optional[bool] = None,
+        samples_per_inference: Optional[int] = None,
+        percentiles: Optional[Sequence[int]] = None,
+        number_simulations: Optional[int] = None,
     ) -> None:
-        super().__init__(model, data, out_dir)
+
+        super().__init__(model, data, out_dir,
+            save,
+            use_progress_bar,
+            samples_per_inference,
+            percentiles,
+            number_simulations)
         self._collect_data_params()
 
-        self.samples_per_inference = (
-            samples_per_inference
-            if samples_per_inference is not None
-            else get_item(
-                "metrics_common", "samples_per_inference", raise_exception=False
-            )
-        )
-        self.percentiles = (
-            percentiles
-            if percentiles is not None
-            else get_item("metrics_common", "percentiles", raise_exception=False)
-        )
-        self.progress_bar = (
-            progress_bar
-            if progress_bar is not None
-            else get_item("metrics_common", "use_progress_bar", raise_exception=False)
-        )
-
     def _collect_data_params(self):
         self.thetas = self.data.get_theta_true()
         self.context = self.data.true_context()
@@ -54,11 +44,11 @@ def calculate(self):
         )
         count_array = []
         iterator = enumerate(self.context)
-        if self.progress_bar:
+        if self.use_progress_bar:
             iterator = tqdm(
                 iterator,
                 desc="Sampling from the posterior for each observation",
-                unit="observation",
+                unit=" observation",
             )
         for y_sample_index, y_sample in iterator:
             samples = self._run_model_inference(self.samples_per_inference, y_sample)

src/metrics/local_two_sample.py

@@ -1,38 +1,58 @@
-from typing import Any, Optional, Union
+from typing import Any, Optional, Sequence, Union
 import numpy as np 
 
 from sklearn.model_selection import KFold
 from sklearn.neural_network import MLPClassifier
 from sklearn.utils import shuffle
 
 from metrics.metric import Metric
-from utils.config import get_item
 
 class LocalTwoSampleTest(Metric): 
-    def __init__(self, model: Any, data: Any, out_dir: str | None = None, num_simulations: Optional[int] = None) -> None:
-        super().__init__(model, data, out_dir)
-        self.num_simulations = num_simulations if num_simulations is not None else get_item(
-            "metrics_common", "number_simulations", raise_exception=False
+    def __init__(
+            self, 
+            model: Any, 
+            data: Any, 
+            out_dir: Optional[str] = None,         
+            save: bool=True,
+            use_progress_bar: Optional[bool] = None,
+            samples_per_inference: Optional[int] = None,
+            percentiles: Optional[Sequence[int]] = None,
+            number_simulations: Optional[int] = None,
+    ) -> None:
+
+        super().__init__(
+            model, 
+            data, 
+            out_dir,
+            save,
+            use_progress_bar,
+            samples_per_inference,
+            percentiles,
+            number_simulations
         )
+
     def _collect_data_params(self):
 
         # P is the prior and x_P is generated via the simulator from the parameters P.
-        self.p = self.data.sample_prior(self.num_simulations)
+        self.p = self.data.sample_prior(self.number_simulations)
         self.q = np.zeros_like(self.p)
 
-        self.outcome_given_p = np.zeros((self.num_simulations, self.data.simulator.generate_context().shape[-1]))
+        context_size = self.data.true_context().shape[-1]
+        self.outcome_given_p = np.zeros(
+            (self.number_simulations, context_size)
+        )
         self.outcome_given_q = np.zeros_like(self.outcome_given_p)
         self.evaluation_context = np.zeros_like(self.outcome_given_p)
 
         for index, p in enumerate(self.p): 
-            context = self.data.simulator.generate_context()
+            context = self.data.simulator.generate_context(context_size)
             self.outcome_given_p[index] = self.data.simulator.simulate(p, context)
             # Q is the approximate posterior amortized in x
             q =  self.model.sample_posterior(1, context).ravel()
             self.q[index] = q
             self.outcome_given_q[index] = self.data.simulator.simulate(q, context)
 
-        self.evaluation_context = np.array([self.data.simulator.generate_context() for _ in range(self.num_simulations)])
+        self.evaluation_context = np.array([self.data.simulator.generate_context(context_size) for _ in range(self.number_simulations)])
 
     def train_linear_classifier(self, p, q, x_p, x_q, classifier:str, classifier_kwargs:dict={}): 
         classifier_map = {
@@ -162,8 +182,8 @@ def calculate(
 
         null =  np.array(null_hypothesis_probabilities)
         self.output = {
-            "lc2st_probabilities": probabilities, 
-            "lc2st_null_hypothesis_probabilities": null
+            "lc2st_probabilities": probabilities.tolist(), 
+            "lc2st_null_hypothesis_probabilities": null.tolist()
         }
         return probabilities, null
 

src/metrics/metric.py

@@ -1,19 +1,37 @@
-from typing import Any, Optional
+from typing import Any, Optional, Sequence
 import json
 import os
 
 from data import data
 from models import model
+from utils.config import get_item
 
 
 class Metric:
-    def __init__(self, model: model, data: data, out_dir: Optional[str] = None) -> None:
+    def __init__(
+            self, 
+            model: model,
+            data: data, 
+            out_dir: Optional[str] = None, 
+            save: bool=True,
+            use_progress_bar: Optional[bool] = None,
+            samples_per_inference: Optional[int] = None,
+            percentiles: Optional[Sequence[int]] = None,
+            number_simulations: Optional[int] = None,
+              ) -> None:
         self.model = model
         self.data = data
 
-        self.out_dir = out_dir
+        if save: 
+            self.out_dir = out_dir if out_dir is not None else get_item("common", "out_dir", raise_exception=False)
+
         self.output = None
 
+        self.use_progress_bar = use_progress_bar if use_progress_bar is not None else get_item("metrics_common", "use_progress_bar", raise_exception=False)
+        self.samples_per_inference = samples_per_inference if samples_per_inference is not None else get_item("metrics_common", "samples_per_inference", raise_exception=False)
+        self.percentiles = percentiles if percentiles is not None else get_item("metrics_common", "percentiles", raise_exception=False)
+        self.number_simulations = number_simulations if number_simulations is not None else get_item("metrics_common", "number_simulations", raise_exception=False)
+
     def _collect_data_params():
         raise NotImplementedError
 
@@ -29,17 +47,29 @@ def _finish(self):
         ), "Calculation has not been completed, have you run Metric.calculate?"
 
         if self.out_dir is not None:
-            if not os.path.exists(os.path.dirname(self.out_dir)):
-                os.makedirs(os.path.dirname(self.out_dir))
+            if not os.path.exists(self.out_dir):
+                os.makedirs(self.out_dir)
+
+            with open(f"{self.out_dir.rstrip('/')}/diagnostic_metrics.json", "w+") as f:
+                try: 
+                    data = json.load(f)
+                except json.decoder.JSONDecodeError: 
+                    data = {}
 
-            with open(self.out_dir) as f:
-                data = json.load(f)
                 data.update(self.output)
                 json.dump(data, f, ensure_ascii=True)
                 f.close()
 
     def __call__(self, **kwds: Any) -> Any:
-        self._collect_data_params()
-        self._run_model_inference()
+
+        try: 
+            self._collect_data_params()
+        except NotImplementedError: 
+            pass 
+        try: 
+            self._run_model_inference() 
+        except NotImplementedError: 
+            pass 
+
         self.calculate(kwds)
         self._finish()

src/plots/__init__.py

@@ -2,10 +2,12 @@
 from plots.coverage_fraction import CoverageFraction
 from plots.ranks import Ranks
 from plots.tarp import TARP
+from plots.local_two_sample import LocalTwoSampleTest
 
 Plots = {
     CDFRanks.__name__: CDFRanks,
     CoverageFraction.__name__: CoverageFraction,
     Ranks.__name__: Ranks,
     TARP.__name__: TARP,
+    "LC2ST": LocalTwoSampleTest
 }