Remove pulse support in QPY in 2.0

[eliarbel]

Summary

This PR removes support for saving ScheduleBlock programs in qpy.dump. It also includes support for loading payloads with ScheduleBlocks or pulse gates while ignoring the actual pulse information, returning partially specified circuits.

Part of #13662

Details and comments

Some details to note and give feedback on:

1. Loading a ScheduleBlock program via qpy.load results with a QuantumCircuit object containing only the name and possibly metadata of that ScheduleBlock. Loading a QuantumCircuit payload with pulse gates leads to getting circuit object with undefined custom instructions (representing pulse gates without calibrations). In both cases warnings of type QPYLoadingDeprecatedFeatureWarning are being issued. Is this the type of warning we want to use here? Maybe we should add something like QPYLoadingRemovedFeatureWarning and update the Compatibility section in the QPY doc page accordingly?
2. This PR does not change the QPY format version, i.e. it stays on 13. This means that we still write CALIBRATION header (with 0 cals) so that load continues to work (since it still includes code for reading calibrations, for graceful backward compatibility).
3. When loading payloads with calibrations, the code will issue a message for each calibration, warning the user that if there exists a gate with this name, it will be undefined. Theoretically, one can add calibrations to a circuit without any pulse gates. We could check this and have the warnings accurately relate to existing pulse gates, but it think it'll just impose extra runtime overhead. I'm open to other opinions though.

Open Tasks:

• Complete handling of test/qpy_compat

[coveralls]

Pull Request Test Coverage Report for Build 13522069571

Details

• 19 of 61 (31.15%) changed or added relevant lines in 4 files are covered.
• 96 unchanged lines in 11 files lost coverage.
• Overall coverage decreased (-0.1%) to 87.831%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
qiskit/qpy/interface.py	5	7	71.43%
qiskit/qpy/binary_io/circuits.py	3	10	30.0%
qiskit/qpy/binary_io/schedules.py	9	42	21.43%

Files with Coverage Reduction	New Missed Lines	%
crates/qasm2/src/expr.rs	1	94.23%
qiskit/pulse/builder.py	1	85.31%
qiskit/pulse/instructions/reference.py	1	90.32%
qiskit/qpy/binary_io/circuits.py	2	88.79%
qiskit/qpy/interface.py	2	86.15%
qiskit/qpy/type_keys.py	4	82.16%
crates/qasm2/src/lex.rs	5	92.48%
qiskit/qpy/binary_io/value.py	5	86.26%
qiskit/pulse/library/symbolic_pulses.py	8	93.01%
crates/qasm2/src/parse.rs	18	96.68%

Totals
Change from base Build 13512719904:	-0.1%
Covered Lines:	77084
Relevant Lines:	87764

---

💛 - Coveralls

[qiskit-bot]

One or more of the following people are relevant to this code:

• @Qiskit/terra-core
• @mtreinish
• @nkanazawa1989

[raynelfss]

Just a typo I found :)

[qiskit-bot]

One or more of the following people are relevant to this code:

• @Qiskit/terra-core
• @mtreinish
• @nkanazawa1989

[qiskit-bot]

One or more of the following people are relevant to this code:

• @Qiskit/terra-core
• @mtreinish
• @nkanazawa1989

[mtreinish]

Overall this looks good. I left a few comments and suggestions inline, but nothing major.

I was confused at first by the changes in the binary_io/schedules.py flle because I skipped over the module docs. But it's clear now that it's basically a bunch of seek() methods now. It's obvious now they're needed because the size to read for the pulse parts isn't known ahead of time, so this is needed to skip the pulse portion of a pulse gates circuit payload.

[eliarbel]

Thanks for the review @mtreinish, all comments have been addressed.

[mtreinish]

This LGTM, I had two small inline comments but nothing worth blocking over. Mostly my idle musings and one extra space in a docstring. Thanks for doing this and the quick updates.