Adjusting to changes to the GAMMLSS api from mssm

README.md

@@ -2,7 +2,7 @@
 
 ## Description
 
-Plotting functions for the Massive Smooth Models ([mssm](https://github.com/JoKra1/mssm)) toolbox. ``mssm`` is a toolbox to estimate Generalized Additive Mixed Models (GAMMs) and Generalized Additive Mixed Models of Location Scale and Shape (GAMMLSS). In addition, a tutorial for ``mssm`` is provided with this repository.
+Plotting functions for the Mixed Sparse Smooth Models ([mssm](https://github.com/JoKra1/mssm)) toolbox. ``mssm`` is a toolbox to estimate Generalized Additive Mixed Models (GAMMs), Generalized Additive Mixed Models of Location Scale and Shape (GAMMLSS), and even more general smooth models in the sense defined by [Wood, Pya, & Säfken (2016)](https://doi.org/10.1080/01621459.2016.1180986). In addition, a tutorial for ``mssm`` is provided with this repository.
 
 ## Installation
 

src/mssmViz/extract.py

@@ -6,14 +6,14 @@
 ################################## Contains functions to extract useful information from GAMM & GAMMLSS models ##################################
 
 
-def eval_coverage(model:GAMM or GAMLSS,pred_dat,dist_par=0,target:float or [float]=0.0,use:[int]=None,alpha=0.05,whole_function=False,n_ps=10000,seed=None):
+def eval_coverage(model:GAMM or GAMMLSS,pred_dat,dist_par=0,target:float or [float]=0.0,use:[int]=None,alpha=0.05,whole_function=False,n_ps=10000,seed=None):
     """Evaluate CI coverage of ``target`` function over domain defined by ``pred_dat``.
 
-    :param model: ``GAMM`` or ``GAMLSS`` model.
-    :type model: GAMM or GAMLSS
+    :param model: ``GAMM`` or ``GAMMLSS`` model.
+    :type model: GAMM or GAMMLSS
     :param pred_dat: ``pandas.DataFrame`` with data used to compute model predictions to be compared against target as well as CI.
     :type pred_dat: pd.Dataframe
-    :param dist_par: The index corresponding to the parameter for which to make the prediction (e.g., 0 = mean) - only necessary if a GAMLSS model is provided, defaults to 0
+    :param dist_par: The index corresponding to the parameter for which to make the prediction (e.g., 0 = mean) - only necessary if a GAMMLSS model is provided, defaults to 0
     :type dist_par: int, optional
     :param target: Target function. Can either be set to a float (e.g., 0.0 if the target function is believed to be zero everywhere across the domain) or a ``list``/``np.array ``of floats. In the latter case the shape of ``target`` must be a flattened 1D array. defaults to 0.0
     :type target: float or [float], optional
@@ -34,7 +34,7 @@ def eval_coverage(model:GAMM or GAMLSS,pred_dat,dist_par=0,target:float or [floa
     """
 
     # Compute model prediction and CI boundaries
-    if isinstance(model,GAMLSS):
+    if isinstance(model,GAMMLSS):
         pred,_,b = model.predict(dist_par,use,pred_dat,ci=True,whole_interval=whole_function,alpha=alpha,n_ps=n_ps,seed=seed)
     else:
         pred,_,b = model.predict(use,pred_dat,ci=True,whole_interval=whole_function,alpha=alpha,n_ps=n_ps,seed=seed)
@@ -57,7 +57,7 @@ def eval_coverage(model:GAMM or GAMLSS,pred_dat,dist_par=0,target:float or [floa
     return full_coverage,coverage,IN_CI
 
 
-def get_term_coef(model:GAMM or GAMLSS,which:[int],dist_par=0):
+def get_term_coef(model:GAMM or GAMMLSS,which:[int],dist_par=0):
     """Get the coefficients associated with a specific term included in the ``Formula`` of ``model``. Useful to extract for example
     the estimated random intercepts from a random effect model.
 
@@ -66,25 +66,25 @@ def get_term_coef(model:GAMM or GAMLSS,which:[int],dist_par=0):
     the actual factor-level corresponding to the first encoded level can be determined by inspecting the code-book returned from ``formula.get_factor_codings()``.
 
 
-    :param model: ``GAMM`` or ``GAMLSS`` model.
-    :type model: GAMM or GAMLSS
+    :param model: ``GAMM`` or ``GAMMLSS`` model.
+    :type model: GAMM or GAMMLSS
     :param which: Index corresponding to the term in the model's formula for which the coefficients should be extracted.
     :type which: [int]
-    :param dist_par: The index corresponding to the parameter for which to make the prediction (e.g., 0 = mean) - only necessary if a GAMLSS model is provided, defaults to 0
+    :param dist_par: The index corresponding to the parameter for which to make the prediction (e.g., 0 = mean) - only necessary if a GAMMLSS model is provided, defaults to 0
     :type dist_par: int, optional
     """
 
     # Get model-matrix that was used for fitting.
     model_mat = model.get_mmat(use_terms=which)
 
-    if isinstance(model,GAMLSS):
+    if isinstance(model,GAMMLSS):
         model_mat = model_mat[dist_par]
 
     # Find coefficient indices corresponding to indicated terms:
     coef_idx = model_mat.sum(axis=0) != 0
 
     # Return corresponding coefficients.
-    if isinstance(model,GAMLSS):
+    if isinstance(model,GAMMLSS):
         split_coef = np.split(model.coef,model.coef_split_idx)
         return split_coef[dist_par][coef_idx]
     else: