Add _precomputable_kwarg_names to BaseExtractor

src/spikeinterface/core/base.py

@@ -47,6 +47,9 @@ class BaseExtractor:
     # these properties are skipped by default in copy_metadata
     _skip_properties = []
 
+    # kwargs which can be precomputed before being used by the extractor
+    _precomputable_kwarg_names = []
+
     installation_mesg = ""
     installed = True
 

src/spikeinterface/preprocessing/whiten.py

@@ -57,6 +57,8 @@ class WhitenRecording(BasePreprocessor):
         The whitened recording extractor
     """
 
+    _precomputable_kwarg_names = ["W", "M"]
+
     def __init__(
         self,
         recording,