Changes!

pymc_marketing/mmm/delayed_saturated_mmm.py

@@ -107,7 +107,7 @@ def __init__(
         time_varying_intercept : bool, optional
             Whether to consider time-varying intercept, by default False.
             Because the `time-varying` variable is centered around 1 and acts as a multiplier,
-            the variable `base_intercept` now represents the mean of the time-varying intercept.
+            the variable `baseline_intercept` now represents the mean of the time-varying intercept.
         time_varying_media : bool, optional
             Whether to consider time-varying media contributions, by default False.
             The `time-varying-media` creates a time media variable centered around 1,
@@ -364,8 +364,9 @@ def build_model(
                 intercept_distribution = get_distribution(
                     name=self.model_config["intercept"]["dist"]
                 )
-                base_intercept = intercept_distribution(
-                    name="base_intercept", **self.model_config["intercept"]["kwargs"]
+                baseline_intercept = intercept_distribution(
+                    name="baseline_intercept",
+                    **self.model_config["intercept"]["kwargs"],
                 )
 
                 intercept_latent_process = create_time_varying_gp_multiplier(
@@ -378,7 +379,7 @@ def build_model(
                 )
                 intercept = pm.Deterministic(
                     name="intercept",
-                    var=base_intercept * intercept_latent_process,
+                    var=baseline_intercept * intercept_latent_process,
                     dims="date",
                 )
             else:
@@ -387,8 +388,8 @@ def build_model(
                 )
 
             if self.time_varying_media:
-                base_channel_contributions = pm.Deterministic(
-                    name="base_channel_contributions",
+                baseline_channel_contributions = pm.Deterministic(
+                    name="baseline_channel_contributions",
                     var=self.forward_pass(x=channel_data_),
                     dims=("date", "channel"),
                 )
@@ -403,7 +404,7 @@ def build_model(
                 )
                 channel_contributions = pm.Deterministic(
                     name="channel_contributions",
-                    var=base_channel_contributions * media_latent_process[:, None],
+                    var=baseline_channel_contributions * media_latent_process[:, None],
                     dims=("date", "channel"),
                 )
 

pymc_marketing/mmm/tvp.py

@@ -90,31 +90,15 @@
 import pandas as pd
 import pymc as pm
 import pytensor.tensor as pt
+from pymc.distributions.shape_utils import Dims
 
 from pymc_marketing.constants import DAYS_IN_YEAR
 
 
-def _softplus(x: pt.TensorVariable) -> pt.TensorVariable:
-    """
-    Compute the softplus function element-wise on the input tensor.
-
-    Parameters
-    ----------
-    x : pt.TensorVariable
-        Input tensor.
-
-    Returns
-    -------
-    pt.TensorVariable
-        Output tensor after applying the softplus function element-wise.
-    """
-    return pm.math.log(1 + pm.math.exp(x))
-
-
 def time_varying_prior(
     name: str,
     X: pt.sharedvar.TensorSharedVariable,
-    dims: tuple[str, str] | str,
+    dims: Dims,
     X_mid: int | float | None = None,
     m: int = 200,
     L: int | float | None = None,
@@ -190,14 +174,14 @@ def time_varying_prior(
     phi, sqrt_psd = gp.prior_linearized(Xs=X[:, None] - X_mid)
     hsgp_coefs = pm.Normal(f"{name}_hsgp_coefs", dims=hsgp_dims)
     f = phi @ (hsgp_coefs * sqrt_psd).T
-    f = _softplus(f)
+    f = pt.softplus(f)
     centered_f = f - f.mean(axis=0) + 1
     return pm.Deterministic(name, centered_f, dims=dims)
 
 
 def create_time_varying_gp_multiplier(
     name: str,
-    dims: tuple[str, str] | str,
+    dims: Dims,
     time_index: pt.sharedvar.TensorSharedVariable,
     time_index_mid: int,
     time_resolution: int,
@@ -228,19 +212,19 @@ def create_time_varying_gp_multiplier(
         Time-varying Gaussian Process multiplier for a given variable.
     """
 
-    if model_config[f"{name}_tvp_config"]["L"] is None:
-        model_config[f"{name}_tvp_config"]["L"] = (
-            time_index_mid + DAYS_IN_YEAR / time_resolution
-        )
-    if model_config[f"{name}_tvp_config"]["ls_mu"] is None:
-        model_config[f"{name}_tvp_config"]["ls_mu"] = DAYS_IN_YEAR / time_resolution * 2
+    tvp_config = model_config[f"{name}_tvp_config"]
+
+    if tvp_config["L"] is None:
+        tvp_config["L"] = time_index_mid + DAYS_IN_YEAR / time_resolution
+    if tvp_config["ls_mu"] is None:
+        tvp_config["ls_mu"] = DAYS_IN_YEAR / time_resolution * 2
 
     multiplier = time_varying_prior(
         name=f"{name}_temporal_latent_multiplier",
         X=time_index,
         X_mid=time_index_mid,
         dims=dims,
-        **model_config[f"{name}_tvp_config"],
+        **tvp_config,
     )
     return multiplier