SortingAnalyzer recording / sorting segments match

[pas-calc]

In earlier versions of SpikeInterface, the SortingAnalyzer's recording's and sorting's number of segments did not need to match.
For example, we were recording for several trials and having several recording files. We give the merged (concatenated cat) file to our favorite spike sorter. We get a sorting result out in global times.

Creating a SortingAnalyzer
calls

spikeinterface/src/spikeinterface/core/sortinganalyzer.py

Line 376 in 25940b4

if has_exceeding_spikes(sorting=sorting, recording=recording):

which does

spikeinterface/src/spikeinterface/curation/remove_excess_spikes.py

Lines 30 to 32 in 25940b4

	assert (
	sorting.get_num_segments() == recording.get_num_segments()
	), "The sorting and recording objects must have the same number of samples!"

and complains for out multi-segment recordings with single-segment sorting.

I think we should treat the recording as a single recording.

The error message should mean "segments" instead of "samples".