minor

- prevent warning in ecg_delineate
- safety check in eventrelated

neurokit2/ecg/ecg_delineate.py

@@ -383,12 +383,11 @@ def _dwt_delineate_tp_peaks(
         srch_idx_start = rpeak_ + srch_bndry
         srch_idx_end = rpeak_ + 2 * int(rt_duration * sampling_rate)
         dwt_local = dwtmatr[degree_tpeak + degree_add, srch_idx_start:srch_idx_end]
-        height = epsilon_T_weight * np.sqrt(np.mean(np.square(dwt_local)))
-
         if len(dwt_local) == 0:
             tpeaks.append(np.nan)
             continue
 
+        height = epsilon_T_weight * np.sqrt(np.mean(np.square(dwt_local)))
         ecg_local = ecg[srch_idx_start:srch_idx_end]
         peaks, __ = scipy.signal.find_peaks(np.abs(dwt_local), height=height)
         peaks = list(
@@ -432,12 +431,11 @@ def _dwt_delineate_tp_peaks(
         srch_idx_start = rpeak - 2 * int(p2r_duration * sampling_rate)
         srch_idx_end = rpeak - srch_bndry
         dwt_local = dwtmatr[degree_ppeak + degree_add, srch_idx_start:srch_idx_end]
-        height = epsilon_P_weight * np.sqrt(np.mean(np.square(dwt_local)))
-
         if len(dwt_local) == 0:
             ppeaks.append(np.nan)
             continue
-
+
+        height = epsilon_P_weight * np.sqrt(np.mean(np.square(dwt_local)))
         ecg_local = ecg[srch_idx_start:srch_idx_end]
         peaks, __ = scipy.signal.find_peaks(np.abs(dwt_local), height=height)
         peaks = list(filter(lambda p: np.abs(dwt_local[p]) > 0.025 * max(dwt_local), peaks))

neurokit2/epochs/eventrelated_utils.py

@@ -12,7 +12,14 @@
 def _eventrelated_sanitizeinput(epochs, what="ecg", silent=False):
     # Sanity checks
     if isinstance(epochs, pd.DataFrame):
-        epochs = _df_to_epochs(epochs)  # Convert df to dict
+        if "Time" in epochs.columns.values:
+            # Assumpe it is a dataframe of epochs created by `epochs_to_df`
+            epochs = _df_to_epochs(epochs)  # Convert df back to dict
+        else:
+            raise ValueError(
+                "It seems that you are trying to pass a single epoch to an eventrelated function."
+                + 'If this is what you want, please wrap it in a dict: `{"0": epoch}` '
+            )
 
     if not isinstance(epochs, dict):
         raise ValueError(
@@ -37,7 +44,6 @@ def _eventrelated_sanitizeinput(epochs, what="ecg", silent=False):
 
 
 def _eventrelated_addinfo(epoch, output={}):
-
     # Add label
     if "Index" in epoch.columns:
         output["Event_Onset"] = epoch.loc[np.min(np.abs(epoch.index))]["Index"]
@@ -58,7 +64,6 @@ def _eventrelated_addinfo(epoch, output={}):
 
 
 def _eventrelated_sanitizeoutput(data):
-
     df = pd.DataFrame.from_dict(data, orient="index")  # Convert to a dataframe
 
     colnames = df.columns.values
@@ -76,12 +81,12 @@ def _eventrelated_sanitizeoutput(data):
 
 
 def _eventrelated_rate(epoch, output={}, var="ECG_Rate"):
-
     # Sanitize input
     colnames = epoch.columns.values
     if len([i for i in colnames if var in i]) == 0:
         warn(
-            "Input does not have an `" + var + "` column." " Will skip all rate-related features.",
+            "Input does not have an `" + var + "` column."
+            " Will skip all rate-related features.",
             category=NeuroKitWarning,
         )
         return output