Tabor segment deduplication

[terrorfisch]

Should fix #795

• test with simulator

[github-actions]

Test Results

       6 files         6 suites   5m 20s ⏱️
1 219 tests 1 163 ✔️   56 💤 0 ❌
7 280 runs  6 850 ✔️ 430 💤 0 ❌

Results for commit c8ef798.

♻️ This comment has been updated with latest results.

[terrorfisch]

@Nomos11 Can you test if this works?

[Nomos11]

so far it seems to give me an IndexError at to_upload_size = np.sum(new_segment_lengths[unknown] + 16), which implies that it still tries to find places for all redundant waveforms, I'll investigate that further

[Nomos11]

It seems like the segment_lengths need to be adapted to throw out duplicates, which should be fixable relatively easily.
E.g., something like

unique_segments = list(segments.keys())
return (unique_segments, [segment.num_points for segment in unique_segments]), waveform_to_segment

at the end of _calc_sampled_segments or some more efficient variation

[terrorfisch]

Does it work now?

[Nomos11]

From looking at the code I am a bit confused - does cutting off the array choose the correct values? the segment_lengths can contain subsequent duplicate values, which are not thrown out here, giving the incorrect values to my understanding

[terrorfisch]

You are right. I did not think it through.

[terrorfisch]

@Nomos11 Should work now