Add empty metadata to template_metrics on load if doesn't exist

src/spikeinterface/metrics/template/template_metrics.py

@@ -169,6 +169,24 @@ def _handle_backward_compatibility_on_load(self):
             if "waveform_ratios" not in self.params["metric_names"]:
                 self.params["metric_names"].append("waveform_ratios")
 
+        if self.data.get("peaks_data") is None:
+            import pandas as pd
+
+            self.data["peaks_data"] = pd.DataFrame(
+                columns=get_peaks_data_columns(), index=self.sorting_analyzer.unit_ids
+            )
+
+        if self.data.get("main_channel_templates") is None:
+            num_units = self.sorting_analyzer.get_num_units()
+            if self.sorting_analyzer.has_extension("templates"):
+                templates_ext = self.sorting_analyzer.get_extension("templates")
+                template_samples = templates_ext.nbefore + templates_ext.nafter
+                upsampling_factor = self.params["upsampling_factor"]
+                upsampled_template_samples = template_samples * upsampling_factor
+                self.data["main_channel_templates"] = np.zeros((num_units, upsampled_template_samples))
+            else:
+                warnings.warn("Cannot set all `template_metrics` metadata as `templates` extension is not available.")
+
     def _set_params(
         self,
         metric_names: list[str] | None = None,
@@ -315,16 +333,8 @@ def _prepare_data(self, sorting_analyzer, unit_ids):
             tmp_data["depth_direction"] = self.params["depth_direction"]
 
         # Add peaks_info and preprocessed templates to self.data for storage in extension
-        columns = []
-        for k in ("trough", "peak_before", "peak_after"):
-            for suffix in (
-                "index",
-                "width_left",
-                "width_right",
-                "half_width_left",
-                "half_width_right",
-            ):
-                columns.append(f"{k}_{suffix}")
+        columns = get_peaks_data_columns()
+
         tmp_data["peaks_data"] = pd.DataFrame(
             index=unit_ids,
             data=peaks_info,
@@ -335,6 +345,21 @@ def _prepare_data(self, sorting_analyzer, unit_ids):
         return tmp_data
 
 
+def get_peaks_data_columns():
+    """Generates the column names of the `peaks_data` DataFrame."""
+    columns = []
+    for k in ("trough", "peak_before", "peak_after"):
+        for suffix in (
+            "index",
+            "width_left",
+            "width_right",
+            "half_width_left",
+            "half_width_right",
+        ):
+            columns.append(f"{k}_{suffix}")
+    return columns
+
+
 register_result_extension(ComputeTemplateMetrics)
 compute_template_metrics = ComputeTemplateMetrics.function_factory()