The sorting object has spikes exceeding the recording duration

[llobetv]

I suspect that is cause by realignment or spike close to end/begining where lussac try to extract after the end or before the begining?

One line could be add to remove those spikes from sorting object? (or if it's to calculate mean wf exclude them for this step?

---

** merge_units ***

---

Running category 'CS':

• Sorting clip_size_20_freq_max_2000_add_end_clip_45 (Done in 21.2 s)
• Sorting clip_size_40_freq_min_600_freq_max_6000_freq_wid_1000 (Done in 18.7 s)
• Sorting clip_size_40_freq_min_800_freq_max_10000_freq_wid_1000_detect_threshold_2.5 (Done in 14.8 s)
• Sorting clip_size_50_freq_min_150_freq_max_2000_freq_wid_1000 (Done in 18.1 s)
• Sorting clip_size_60_freq_min_300_freq_max_6000_freq_wid_1000 (Done in 16.4 s)
• Sorting clip_size_120_freq_min_60_freq_max_1000_freq_wid_300 (Done in 17.0 s)

• Sorting ks2_CS Traceback (most recent call last):

  File "/mnt/data/llobet/lussac/src/lussac/main.py", line 67, in main
  pipeline.launch() File "/mnt/data/llobet/lussac/src/lussac/core/pipeline.py", line 60, in launch self._run_mono_sorting_module(module, module_key, category, params) File "/mnt/data/llobet/lussac/src/lussac/core/pipeline.py", line 99, in _run_mono_sorting_module sub_sorting = module_instance.run(params0) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/mnt/data/llobet/lussac/src/lussac/modules/merge_units.py", line 55, in run wvf_extractor = self.extract_waveforms(params['wvf_extraction'])
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/mnt/data/llobet/lussac/src/lussac/core/module.py", line 186, in extract_waveforms return si.extract_waveforms(recording, sorting, folder_path, allow_unfiltered=True, params) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/mnt/data/llobet/miniconda3/envs/lussac/lib/python3.11/site-packages/spikeinterface/core/waveform_extractor.py", line 1567, in extract_waveforms we = WaveformExtractor.create( ^^^^^^^^^^^^^^^^^^^^^^^^^ File "/mnt/data/llobet/miniconda3/envs/lussac/lib/python3.11/site-packages/spikeinterface/core/waveform_extractor.py", line 251, in create raise ValueError( ValueError: The sorting object has spikes exceeding the recording duration. You have to remove those spikes with the spikeinterface.curation.remove_excess_spikes() function /mnt/data/llobet/miniconda3/envs/lussac/lib/python3.11/site-packages/numba/np/ufunc/parallel.py:371: NumbaWarning:

[DradeAW]

I added a patch in the align_units module to take care of the spikes going over the edge of the recording. I also added the corresponding unit tests.

Can you pull the latest commit from the dev branch and confirm whether it solves your issue? You will have to remove the logs for align_units and all modules after it.

Thanks for the report!

[llobetv]

Fix works thanks !