Use ProbeGroup object instead of contact_vector property and set_probe/probegroup from select_channels_with_probe/probegroup

src/spikeinterface/core/base.py

@@ -213,6 +213,14 @@ def id_to_index(self, id) -> int:
         return ind
 
     def annotate(self, **new_annotations) -> None:
+        """Adds annotations.
+
+        Parameters
+        ----------
+        **new_annotations : dict
+            Key-value pairs of annotations to add. If an annotation key already exists,
+            it will be overwritten.
+        """
         self._annotations.update(new_annotations)
 
     def set_annotation(self, annotation_key: str, value: Any, overwrite=False) -> None:
@@ -236,6 +244,24 @@ def set_annotation(self, annotation_key: str, value: Any, overwrite=False) -> No
             else:
                 raise ValueError(f"{annotation_key} is already an annotation key. Use 'overwrite=True' to overwrite it")
 
+    def delete_annotation(self, annotation_key: str) -> None:
+        """Deletes existing annotation.
+
+        Parameters
+        ----------
+        annotation_key : str
+            The annotation key to delete
+
+        Raises
+        ------
+        ValueError
+            If the annotation key does not exist
+        """
+        if annotation_key in self._annotations.keys():
+            del self._annotations[annotation_key]
+        else:
+            raise ValueError(f"{annotation_key} is not an annotation key")
+
     def get_preferred_mp_context(self):
         """
         Get the preferred context for multiprocessing.
@@ -434,6 +460,15 @@ def copy_metadata(
         if self._preferred_mp_context is not None:
             other._preferred_mp_context = self._preferred_mp_context
 
+        if not only_main:
+            self._extra_metadata_copy(other)
+
+    def _extra_metadata_copy(self, other: BaseExtractor):
+        """
+        This is a hook to copy extra metadata that is not in the annotations/properties dict.
+        """
+        pass
+
     def to_dict(
         self,
         include_annotations: bool = False,
@@ -567,6 +602,8 @@ def to_dict(
                 folder_metadata = Path(folder_metadata).resolve().absolute().relative_to(relative_to)
             dump_dict["folder_metadata"] = str(folder_metadata)
 
+        self._extra_metadata_to_dict(dump_dict)
+
         return dump_dict
 
     @staticmethod
@@ -855,6 +892,14 @@ def _extra_metadata_to_folder(self, folder):
         # This implemented in BaseRecording for probe
         pass
 
+    def _extra_metadata_from_dict(self, dump_dict):
+        # This implemented in BaseRecording for probe
+        pass
+
+    def _extra_metadata_to_dict(self, dump_dict):
+        # This implemented in BaseRecording for probe
+        pass
+
     def save(self, **kwargs) -> BaseExtractor:
         """
         Save a SpikeInterface object.
@@ -1154,6 +1199,8 @@ def _load_extractor_from_dict(dic) -> BaseExtractor:
     for k, v in dic["properties"].items():
         extractor.set_property(k, v)
 
+    extractor._extra_metadata_from_dict(dic)
+
     return extractor
 
 

src/spikeinterface/core/baserecording.py

@@ -21,7 +21,6 @@ class BaseRecording(BaseRecordingSnippets):
     _main_annotations = BaseRecordingSnippets._main_annotations + ["is_filtered"]
     _main_properties = [
         "group",
-        "location",
         "gain_to_uV",
         "offset_to_uV",
         "gain_to_physical_unit",
@@ -591,15 +590,16 @@ def _save(self, format="binary", verbose: bool = False, **save_kwargs):
         kwargs, job_kwargs = split_job_kwargs(save_kwargs)
 
         if format == "binary":
+            from .binaryfolder import BinaryFolderRecording
+            from .binaryrecordingextractor import BinaryRecordingExtractor
+
             folder = kwargs["folder"]
             file_paths = [folder / f"traces_cached_seg{i}.raw" for i in range(self.get_num_segments())]
             dtype = kwargs.get("dtype", None) or self.get_dtype()
             t_starts = self._get_t_starts()
 
             write_binary_recording(self, file_paths=file_paths, dtype=dtype, verbose=verbose, **job_kwargs)
 
-            from .binaryrecordingextractor import BinaryRecordingExtractor
-
             # This is created so it can be saved as json because the `BinaryFolderRecording` requires it loading
             # See the __init__ of `BinaryFolderRecording`
             binary_rec = BinaryRecordingExtractor(
@@ -616,8 +616,9 @@ def _save(self, format="binary", verbose: bool = False, **save_kwargs):
                 offset_to_uV=self.get_channel_offsets(),
             )
             binary_rec.dump(folder / "binary.json", relative_to=folder)
-
-            from .binaryfolder import BinaryFolderRecording
+            if self.has_probe():
+                probegroup = self.get_probegroup()
+                write_probeinterface(folder / "probe.json", probegroup)
 
             cached = BinaryFolderRecording(folder_path=folder)
 
@@ -648,10 +649,7 @@ def _save(self, format="binary", verbose: bool = False, **save_kwargs):
         else:
             raise ValueError(f"format {format} not supported")
 
-        if self.get_property("contact_vector") is not None:
-            probegroup = self.get_probegroup()
-            cached.set_probegroup(probegroup)
-
+        # TODO: write binary should save timestamps too
         for segment_index in range(self.get_num_segments()):
             if self.has_time_vector(segment_index):
                 # the use of get_times is preferred since timestamps are converted to array
@@ -676,7 +674,7 @@ def _extra_metadata_from_folder(self, folder):
 
     def _extra_metadata_to_folder(self, folder):
         # save probe
-        if self.get_property("contact_vector") is not None:
+        if self.has_probe():
             probegroup = self.get_probegroup()
             write_probeinterface(folder / "probe.json", probegroup)