rename diagnostics plots (#265)

* rename diagnostics plots

* Slight name change

---------

Co-authored-by: stefanradev93 <[email protected]>

Author: paul-buerkner

bayesflow/diagnostics/__init__.py

@@ -1,9 +1,11 @@
-from .plot_losses import plot_losses
-from .plot_recovery import plot_recovery
-from .plot_sbc_ecdf import plot_sbc_ecdf
-from .plot_sbc_histograms import plot_sbc_histograms
-from .plot_samples_2d import plot_samples_2d
-from .plot_z_score_contraction import plot_z_score_contraction
-from .plot_prior_2d import plot_prior_2d
-from .plot_posterior_2d import plot_posterior_2d
-from .plot_calibration_curves import plot_calibration_curves
+from .plots import calibration_ecdf
+from .plots import calibration_histogram
+from .plots import loss
+from .plots import mc_calibration
+from .plots import mc_confusion_matrix
+from .plots import mmd_hypothesis_test
+from .plots import pairs_posterior
+from .plots import pairs_prior
+from .plots import pairs_samples
+from .plots import recovery
+from .plots import z_score_contraction

bayesflow/diagnostics/plots/__init__.py

@@ -0,0 +1,11 @@
+from .calibration_ecdf import calibration_ecdf
+from .calibration_histogram import calibration_histogram
+from .loss import loss
+from .mc_calibration import mc_calibration
+from .mc_confusion_matrix import mc_confusion_matrix
+from .mmd_hypothesis_test import mmd_hypothesis_test
+from .pairs_posterior import pairs_posterior
+from .pairs_prior import pairs_prior
+from .pairs_samples import pairs_samples
+from .recovery import recovery
+from .z_score_contraction import z_score_contraction

bayesflow/diagnostics/plot_sbc_ecdf.py renamed to bayesflow/diagnostics/plots/calibration_ecdf.py

@@ -2,12 +2,12 @@
 import matplotlib.pyplot as plt
 
 from typing import Sequence
-from ..utils.plot_utils import preprocess, add_titles_and_labels, prettify_subplots
-from ..utils.ecdf import simultaneous_ecdf_bands
-from ..utils.ecdf.ranks import fractional_ranks, distance_ranks
+from ...utils.plot_utils import preprocess, add_titles_and_labels, prettify_subplots
+from ...utils.ecdf import simultaneous_ecdf_bands
+from ...utils.ecdf.ranks import fractional_ranks, distance_ranks
 
 
-def plot_sbc_ecdf(
+def calibration_ecdf(
     post_samples: dict[str, np.ndarray] | np.ndarray,
     prior_samples: dict[str, np.ndarray] | np.ndarray,
     filter_keys: Sequence[str] = None,
@@ -61,12 +61,15 @@ def plot_sbc_ecdf(
     stacked           : bool, optional, default: False
         If `True`, all ECDFs will be plotted on the same plot.
         If `False`, each ECDF will have its own subplot,
-        similar to the behavior of `plot_sbc_histograms`.
+        similar to the behavior of `calibration_histogram`.
     rank_type   : str, optional, default: 'fractional'
-        If `fractional` (default), the ranks are computed as the fraction of posterior samples that are smaller than
-        the prior. If `distance`, the ranks are computed as the fraction of posterior samples that are closer to
-        a reference points (default here is the origin). You can pass a reference array in the same shape as the
-        `prior_samples` array by setting `references` in the ``ranks_kwargs``. This is motivated by [2].
+        If `fractional` (default), the ranks are computed as the fraction
+        of posterior samples that are smaller than the prior.
+        If `distance`, the ranks are computed as the fraction of posterior
+        samples that are closer to a reference points (default here is the origin).
+        You can pass a reference array in the same shape as the
+        `prior_samples` array by setting `references` in the ``ranks_kwargs``.
+        This is motivated by [2].
     variable_names    : list or None, optional, default: None
         The parameter names for nice plot titles.
         Inferred if None. Only relevant if `stacked=False`.

bayesflow/diagnostics/plot_sbc_histograms.py renamed to bayesflow/diagnostics/plots/calibration_histogram.py

@@ -9,7 +9,7 @@
 from bayesflow.utils import preprocess, add_titles_and_labels, prettify_subplots
 
 
-def plot_sbc_histograms(
+def calibration_histogram(
     post_samples: dict[str, np.ndarray] | np.ndarray,
     prior_samples: dict[str, np.ndarray] | np.ndarray,
     filter_keys: Sequence[str] = None,

bayesflow/diagnostics/plot_losses.py renamed to bayesflow/diagnostics/plots/loss.py

@@ -6,10 +6,10 @@
 import matplotlib.pyplot as plt
 
 
-from ..utils.plot_utils import make_figure, add_titles_and_labels
+from ...utils.plot_utils import make_figure, add_titles_and_labels
 
 
-def plot_losses(
+def loss(
     train_losses: pd.DataFrame | np.ndarray,
     val_losses: pd.DataFrame | np.ndarray = None,
     moving_average: bool = False,

bayesflow/diagnostics/plot_calibration_curves.py renamed to bayesflow/diagnostics/plots/mc_calibration.py

@@ -7,9 +7,9 @@
 from bayesflow.utils import expected_calibration_error, preprocess, add_titles_and_labels, add_metric, prettify_subplots
 
 
-def plot_calibration_curves(
-    post_model_samples: dict[str, np.ndarray] | np.ndarray,
-    true_model_samples: dict[str, np.ndarray] | np.ndarray,
+def mc_calibration(
+    pred_models: dict[str, np.ndarray] | np.ndarray,
+    true_models: dict[str, np.ndarray] | np.ndarray,
     names: Sequence[str] = None,
     num_bins: int = 10,
     label_fontsize: int = 16,
@@ -28,11 +28,11 @@ def plot_calibration_curves(
 
     Parameters
     ----------
-    true_model_samples       : np.ndarray of shape (num_data_sets, num_models)
+    true_models       : np.ndarray of shape (num_data_sets, num_models)
         The one-hot-encoded true model indices per data set.
-    post_model_samples      : np.ndarray of shape (num_data_sets, num_models)
+    pred_models       : np.ndarray of shape (num_data_sets, num_models)
         The predicted posterior model probabilities (PMPs) per data set.
-    names       : list or None, optional, default: None
+    names             : list or None, optional, default: None
         The model names for nice plot titles. Inferred if None.
     num_bins          : int, optional, default: 10
         The number of bins to use for the calibration curves (and marginal histograms).
@@ -60,7 +60,7 @@ def plot_calibration_curves(
     fig : plt.Figure - the figure instance for optional saving
     """
 
-    plot_data = preprocess(post_model_samples, true_model_samples, names, num_col, num_row, figsize, context="M")
+    plot_data = preprocess(pred_models, true_models, names, num_col, num_row, figsize, context="M")
 
     # Compute calibration
     cal_errors, true_probs, pred_probs = expected_calibration_error(

bayesflow/diagnostics/plot_confusion_matrix.py renamed to bayesflow/diagnostics/plots/mc_confusion_matrix.py

@@ -12,7 +12,7 @@
 from bayesflow.utils.plot_utils import make_figure
 
 
-def plot_confusion_matrix(
+def mc_confusion_matrix(
     true_models: dict[str, np.ndarray] | np.ndarray,
     pred_models: dict[str, np.ndarray] | np.ndarray,
     model_names: Sequence[str] = None,

bayesflow/diagnostics/plot_mmd_hypothesis_test.py renamed to bayesflow/diagnostics/plots/mmd_hypothesis_test.py

@@ -5,7 +5,7 @@
 from keras import ops
 
 
-def plot_mmd_hypothesis_test(
+def mmd_hypothesis_test(
     mmd_null: np.ndarray,
     mmd_observed: float = None,
     alpha_level: float = 0.05,

bayesflow/diagnostics/plot_posterior_2d.py renamed to bayesflow/diagnostics/plots/pairs_posterior.py

@@ -6,10 +6,10 @@
 
 from matplotlib.lines import Line2D
 
-from .plot_samples_2d import plot_samples_2d
+from .pairs_samples import pairs_samples
 
 
-def plot_posterior_2d(
+def pairs_posterior(
     post_samples: np.ndarray,
     prior_samples: np.ndarray = None,
     prior=None,
@@ -70,7 +70,7 @@ def plot_posterior_2d(
 
     # Plot posterior first
     context = ""
-    g = plot_samples_2d(
+    g = pairs_samples(
         post_samples, context=context, variable_names=variable_names, render=False, height=height, **kwargs
     )
 

bayesflow/diagnostics/plot_prior_2d.py renamed to bayesflow/diagnostics/plots/pairs_prior.py

@@ -3,10 +3,10 @@
 import seaborn as sns
 
 from bayesflow.simulators import Simulator
-from .plot_samples_2d import plot_samples_2d
+from .pairs_samples import pairs_samples
 
 
-def plot_prior_2d(
+def pairs_prior(
     simulator: Simulator,
     variable_names: Sequence[str] | str = None,
     num_samples: int = 2000,
@@ -43,6 +43,6 @@ def plot_prior_2d(
     if isinstance(samples, dict):
         samples = samples["theta"]
 
-    return plot_samples_2d(
+    return pairs_samples(
         samples, context="Prior", height=height, color=color, param_names=variable_names, render=True, **kwargs
     )