Refactor media to simplify line creation

[mhuser]

This PR is the second attempt to fix #651

Ready for review.

Todo

• Refactor all media for z0_port, z0 scheme
• Refactor waveguides media for z0_override, z0 scheme
• Make all tests pass and refine
• Make all notebooks build
• Update notebooks (check contents)
• Update documentation
• Extend documentation (e.g. by passing z0 to Media.line())

[mhuser]

A deprecation warning is now emmited if z0 is used in Media initialization. This helped to find some more z0 remaining.
The z0 fallback on z0_port, resp. z0_override if these are None.

[mhuser]

After having reflected a bit on it, I think there is some tradeoff here. If we keep the PR to merge within 1.0, it probably will not benefits from the extensive testing by many users on Master, with the risk to miss key issues that we later have to fix in 1.0.
On the other hand, breaking user code can act as a repellent. Here they will always get a media line with a warning in the cases were the behaviour is subject to have changed.

[jhillairet]

Updated documentation for reviewers:
https://scikit-rf--727.org.readthedocs.build/en/727/

• doc/source/tutorials/Introduction.ipynb
• doc/source/tutorials/Media.ipynb
• doc/source/examples/circuit/Wilkinson Power Splitter.ipynb
• doc/source/examples/interactive/Interact Transmission Lines.ipynb
• doc/source/examples/interactive/Interactive Coaxial Properties.ipynb
• doc/source/examples/interactive/Interactive Mismatched Line.ipynb
• doc/source/examples/metrology/Calibration With Three Receivers.ipynb
• doc/source/examples/metrology/LRRM.ipynb
• doc/source/examples/metrology/Multi-port Calibration.ipynb
• doc/source/examples/metrology/Multiline TRL.ipynb
• doc/source/examples/metrology/NanoVNA_V2_4port-splitter.ipynb
• doc/source/examples/metrology/SOLT Calibration Standards Creation.ipynb
• doc/source/examples/metrology/TwoPortOnePath, EnhancedResponse, and FakeFlip.ipynb
• doc/source/examples/networksets/Sorting Network Sets.ipynb
• doc/source/examples/networktheory/Balanced Network De-embedding.ipynb
• doc/source/examples/networktheory/CPW media example.ipynb
• ...ce/examples/networktheory/Correlating microstripline model to measurement.ipynb
• doc/source/examples/networktheory/DC Extrapolation for Time Domain .ipynb
• doc/source/examples/networktheory/DefinedAEpTandZ0 media example.ipynb
• doc/source/examples/networktheory/IEEEP370 Deembedding.ipynb
• doc/source/examples/networktheory/LNA Example.ipynb
• doc/source/examples/networktheory/Properties of Rectangular Waveguides.ipynb
• ...amples/networktheory/Time domain reflectometry, measurement vs simulation.ipynb
• ...rce/examples/networktheory/Transmission Line Properties and Manipulations.ipynb
• doc/source/examples/taper/tapered_transmission_lines.ipynb

[jhillairet]

I think additional explanation should be given about what is returned here (None or something)

[mhuser]

Now that z0_override is a property of the abstract class Media, the switch between z0_override and abstract method z0_characteristic can be done here. The media classes will need to define z0_characteristic, which sounds quite clear what it is 😉

    @property
    @abstractmethod
    def z0_characteristic(self):
        """
        Characteristic Impedance.
        This abstract method has to be defined in the Media Class.
        
        Returns
        -------
        z0_characteristic: npy.ndarray
            Characteristic Impedance in units of ohms
        """
        return None
    
    @property
    def z0(self):
        """
        Return Characteristic Impedance `z0_characteristic`.
        If `z0_override` is not None, it is returned instead.
        
        Returns
        -------
        z0 : npy.ndarray
            z0_characteristic or z0_override in units of ohms
        """
        if self.z0_override is None:
            return self.z0_characteristic
        else:
            return self.z0_override

[mhuser]

Done.

[jhillairet]

@mhuser do you think it is ok for the next release now ? (or the one after the next one?)

[mhuser]

@mhuser do you think it is ok for the next release now ? (or the one after the next one?)

I think yes. Let's merge branch master inside. 😅

[jhillairet]

Great (and huge) work @mhuser . Let's see if anyone complain after the next release ;)

[drsvenkuehn]

This PR is the second attempt to fix #651

Ready for review.

Todo

• Refactor all media for z0_port, z0 scheme
• Refactor waveguides media for z0_override, z0 scheme
• Make all tests pass and refine
• Make all notebooks build
• Update notebooks (check contents)
• Update documentation
• Extend documentation (e.g. by passing z0 to Media.line())

I believe this breaks the "media coaxial" example in the tutorial.

[jhillairet]

I believe this breaks the "media coaxial" example in the tutorial.

@drsvenkuehn, which one ?
https://scikit-rf.readthedocs.io/en/latest/tutorials/Introduction.html#Coax ?
or
https://scikit-rf.readthedocs.io/en/latest/tutorials/Media.html#RG58C/U-coaxial-cable ?

[drsvenkuehn]

I believe this breaks the "media coaxial" example in the tutorial.

@drsvenkuehn, which one ? https://scikit-rf.readthedocs.io/en/latest/tutorials/Introduction.html#Coax ? or https://scikit-rf.readthedocs.io/en/latest/tutorials/Media.html#RG58C/U-coaxial-cable ?

@jhillairet
the RG58 one. print(f'z0_port = {rg58.z0_port[0]}') gives an error as z0_port[0] is gone. Also, the S-parameter results are not the same as before: rg58.line(100, unit = 'mm', z0 = 25, name = '100 mm, 25 Ohm') appears always perfectly matched independent of z0

[mhuser]

Hello @drsvenkuehn ,
Thank you for your interest in Media.
I am a bit confused because after looking at https://scikit-rf.readthedocs.io/en/latest/tutorials/Media.html#RG58C/U-coaxial-cable, I did not notice the behaviour that you describe.

print(f'z0_port = {rg58.z0_port[0]}')

gives z0_port = 50.0

and

rg58.line(100, unit = 'mm', z0 = 25, name = '100 mm, 25 Ohm')

Gives mismatch

Is it possible you are speaking of another version ?

[drsvenkuehn]

Hi @mhuser ...sorry, my bad. I didn't use the right virtual environment. I used an old version. I get the same results with 0.30.0

[jhillairet]

Ok, thanks for letting us know.