Give users easier access to container logs

[BSchilperoort]

Closes #469, #231

Usage example:

[codecov]

Codecov Report

Attention: Patch coverage is 16.66667% with 10 lines in your changes missing coverage. Please review.

Project coverage is 78.49%. Comparing base (cf5f3c9) to head (09e8d3b).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/ewatercycle/base/model.py	16.66%	10 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #470      +/-   ##
==========================================
- Coverage   78.89%   78.49%   -0.40%     
==========================================
  Files          28       28              
  Lines        1876     1888      +12     
  Branches      159      162       +3     
==========================================
+ Hits         1480     1482       +2     
- Misses        337      347      +10     
  Partials       59       59              

Files with missing lines	Coverage Δ
src/ewatercycle/base/model.py	92.42% <16.66%> (-4.89%)	⬇️

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[BSchilperoort]

@RolfHut to review you could just test the functionality in a notebook you are already using somewhere else.

Install this in-dev version of ewatercycle in your terminal with:

python -m pip install --no-deps https://github.com/eWaterCycle/ewatercycle/archive/refs/heads/main.zip

This should not mess up your environment (due to the no-deps flag only this package will be updated).

---

I know that the code coverage check is failing, but it's not trivial to write a test for this case and time is sadly limited.

[BSchilperoort]

@ahmedfarhatx22 this should give you easy access to the model container logs :)

[RolfHut]

machines already running do not have remote BMI and now fall over...

[RolfHut]

in testing on HBV I get this error:

'HBV' object has no attribute 'logs'

screenshot:

so maybe something goed wrong in error handling if a model does not support this?

[BSchilperoort]

in testing on HBV I get this error:

'HBV' object has no attribute 'logs'

screenshot:

so maybe something goed wrong in error handling if a model does not support this?

It's a property of the ContainerizedModel class: any (grpc) containerized model should work.

I did add a check if the model is set up yet in 388f748, and improved the error messages.

[RolfHut]

also a no at leaky bucket

[RolfHut]

in testing on HBV I get this error:
'HBV' object has no attribute 'logs'
screenshot:
so maybe something goed wrong in error handling if a model does not support this?

It's a property of the ContainerizedModel class: any (grpc) containerized model should work.

I did add a check if the model is set up yet in 388f748, and improved the error messages.

how do I pull your commit on my SRC machine? than I can check if this fixes it.

[BSchilperoort]

I realize now I gave you the wrong command, sorry.

It should have been
python -m pip install --no-deps https://github.com/eWaterCycle/ewatercycle/archive/refs/heads/access-logs.zip

the other one was for main

[RolfHut]

tried that and same error on leaky bucket: no attribute 'logs'

[sverhoeven]

Ran the notebook with model_instance.logs calls successfully.

However it would be nice to have some unit tests that take less work to setup.

[BSchilperoort]

However it would be nice to have some unit tests that take less work to setup.

I considered this, but the logs are part of the containerized model, where testing relies on mocking a lot. As we don't actually start a container, I would have to mock the starting of the container, and mock the log return. At that point not much would be left to actually test as it's only mocks. The classes which actually implement the logs are in grpc4bmi and remotebmi. It would be good for this package to actually have a containerized model integrated which is used in unit tests, but that's not in the scope of this PR.