Sequence building

[alecandido]

Just missing tests.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 51.47%. Comparing base (40b09e4) to head (18ec55d).
Report is 11 commits behind head on 0.2.

Additional details and impacted files

@@            Coverage Diff             @@
##              0.2    #1026      +/-   ##
==========================================
+ Coverage   51.13%   51.47%   +0.33%     
==========================================
  Files          56       56              
  Lines        2726     2745      +19     
==========================================
+ Hits         1394     1413      +19     
  Misses       1332     1332              

Flag	Coverage Δ
unittests	51.47% <100.00%> (+0.33%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[alecandido]

The tests/integration/ folder is a mini showcase of what I have in mind for #1021. I did not really want to enter that chapter, but I wanted a comprehensive test to showcase the interface as I believe it could be used in Qibocal, and this was not a unit test neither for native nor for sequence, but spanning both.
That's why I created the additional folder, but I don't want to face the more general issue right now.

[stavros11]

Why do we accept any iterable here? Isn't PulseSequence sufficient? Same for __or__ and __ior__.

[alecandido]

qibolab/tests/integration/test_sequence.py

Lines 22 to 29 in 746fe8f

	seq = (
	q1.RX()
	| p12.CZ()
	| [(ch1.drive, Delay(duration=6.5))]
	| q2.RX()
	| q0.RX12()
	| p02.CZ()
	)

The explicit pulse-like here (the Delay) is contained in a list, which is perfectly homogeneous to a PulseSequence (since PulseSequence is a list subclass), but it's not a PulseSequence.

In general, it is very cheap to support, and quite user-friendly (e.g. even an iterator over _Element would work, so you can use literals such as list comprehension and generators, that otherwise should be manually wrapped in a PulseSequence).

[alecandido]

On my side, I'm ok to merge. But there are still two open discussions.