Async with real detector

[uellue]

Edit: Now closes #29 thanks to fix by @sk1p!

This PR is just a failing test case since I couldn't get it to work properly. It is based on #29

The reason that it fails is that we have to keep the connection until the computation is complete, i.e. until the result is awaited or the async generator consumed. The sync case is fixed in #29.

@sk1p could you have a look? I am not sure how live processing in an async context
should be implemented, since we have to read the data in the proper
sequence from one socket.

Use case for this API would probably be integration in an async GUI application? Or should we not support async at all and throw a NotImplementedError if sync is False?

Contributor Checklist:

• I have added or updated my entry in the creators.json file
• I have added a changelog entry for my contribution
• I have added/updated documentation for all user-facing changes
• I have added/updated test cases

Reviewer Checklist:

• /azp run libertem.libertem-live-data passed

[uellue]

/azp run libertem.libertem-live-data

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[codecov]

Codecov Report

Merging #30 (4c9c1d4) into master (c203202) will decrease coverage by 0.03%.
The diff coverage is 80.00%.

@@            Coverage Diff             @@
##           master      #30      +/-   ##
==========================================
- Coverage   74.00%   73.97%   -0.04%     
==========================================
  Files          14       15       +1     
  Lines        1208     1241      +33     
  Branches      142      145       +3     
==========================================
+ Hits          894      918      +24     
- Misses        276      284       +8     
- Partials       38       39       +1     

Impacted Files	Coverage Δ
src/libertem_live/detectors/merlin/data.py	71.30% <72.72%> (-0.35%)	⬇️
src/libertem_live/detectors/common.py	76.00% <76.00%> (ø)
src/libertem_live/api.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c203202...4c9c1d4. Read the comment docs.

[sk1p]

Use case for this API would probably be integration in an async GUI application? Or should we not support async at all and throw a NotImplementedError if sync is False?

I think it does make sense to support the async case - especially if lots of things need to be orchestrated, lots of network connections need to be handled concurrently etc. it would be useful to write the control script or notebook using async.

I'll have a look at the PR changes next.

[sk1p]

/azp run libertem.libertem-live-data

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[sk1p]

I pushed my changes into your branch, as discussed offline. This should now be a general solution, for all cases, and should thus supersede #29.

[sk1p]

The py36 tests are failing again - this should be unrelated to the current PR, The py36 tests are failing again - this should be unrelated to the current PR, but I'll have a look at maybe stress testing py36 a bit tomorrow.

[uellue]

Nice, thank you! 👍

[uellue]

/azp run libertem.libertem-live-data

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[sk1p]

/azp run libertem.libertem-live-data

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[uellue]

Regarding the intermittent test failures: @sk1p now has better error handling that will hopefully give better insight on what happens there. We'll merge for now and see what happens!