SRS tests

[manny118]

Tests for the SRS models.

[manny118]

Extends the previous request here.

[juraul1]

All tests are failing because the fibre_spectral_attenuation list is not in edfa_params.py in this PR.
Might be worth adding it to see if tests are passing

[lantz]

Yes, that should have been included in this PR instead. However we can merge #133 and see if it helps.

[lantz]

@juraul1 @manny118 I enabled the (somewhat recently added) repo setting enabling admins to update (i.e. git rebase) pull requests. For an out-of-date branch that merges cleanly, there should now be an update button (visible to admins at least.)

It's a very useful feature, and I just used it to update/rebase this PR after merging #133.

[lantz]

@manny118 Now the SRS tests look like they are legitimately failing:

======================================================================
FAIL: test_SRS (__main__.TestSRS)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/runner/work/mininet-optical/mininet-optical/tests/SRS_Test2.py", line 124, in test_SRS
    self.assertEqual(actual_result, expected_result)
AssertionError: 0.62 != 1.01
----------------------------------------------------------------------

see line 744 of: https://github.com/Mininet-Optical/mininet-optical/runs/7729810659?check_suite_focus=true#step:7:744

So, the SRS tests are doing their thing, which is good. Now we just have to get them passing.

[manny118]

The first script passed whereas others failed, and that is because each script calls a different SRS function. Only one function can be enabled each time in the code.

[lantz]

Could you fix it so that all SRS functions are tested and the tests pass?

We can't merge the branch unless the tests pass.

[manny118]

An option would be to use 4 different pull requests, although the same link.py script would be used for each. I am not sure of how that would work since it would mean having all four functions enabled in the code when they are merged into the repo which is not ideal.

[lantz]

1. Shouldn't srs_effect=True be the default?
2. Could you explicitly add srs_effect=True as a Link parameter in your test code?
3. How is the SRS model specified? Could you do something like this in link.py?:

...
class Link:
    srs_models = [MyCoolSRSModel, SomeOtherSRSModel, ..., None]
    srs_model = MyCoolSRSModel
    def __init__(self, ... **params):
        ...
       self.srs_model = params.get('srs_model', self.srs_model)

...
    def propagate(....):
        if self.srs_model: 
            self.srs_model(...)
        

That would also replace srs_effect=False with srs_model=None which makes sense to me.