Add Wifi 802.11 physical simulation

[eSoares]

This adds the following changes:

• Adds an helper class that wraps the basics of an 802.11 simulation channel (the generator matrix according to the standard, coding and puncturing when needed, depuncturing and decoding);
• Adds the collection of more parameters of the simulation;
• Add chunk aggregation possibility. This looks like it is the same as sending a bigger chunk, but for accounting number of some parameters it makes a difference (i.e. number of chunk loss);
• [Breaking change] Changes from number of bits to simulate sending, to number of runs to simulate. The parameter always had that meaning and in the end I think it makes more sense. For a quick migration path, any user is advice to simple divide the number number of bits to simulate by the send chunk (and perform a rounding/ceiling);

This last change can be reverted, but makes some code of collecting more metrics a bit worse.

[coveralls]

Coverage increased (+1.2%) to 79.19% when pulling 78c6e25bae9a37ab2975bd6414ecc93fba72e185 on eSoares:wifi802.11 into 8687174a85275e677fd0703da24e639a8b11cea2 on veeresht:master.

[BastienTr]

We will also think about updating the readme and documentation build after these additions. I can handle this once the code has been stabilized.

[eSoares]

Thanks for the comments. :smile:

Sorry for the error on the documentation of code_rate, I was confusing the python typings syntax and found weird the (0,1] syntax. Ended up documenting the puncturing and depuncturing methods with np.ndarray, while is not strictly necessary to be ndarrays (lists work there), since all the documentation indicates they should be np.ndarrays I believe to be better if the typings reinforce this. The suggested Iterable doesn't work in some cases, since they don't support len() and indexing.

link_performance I agree that is better not break any code or functionality. While my change is straightforward to follow, and documentation could be added, is better and simpler if no change is needed. I have reintroduced the link_performance exactly as it was and renamed my proposed method to link_performance_full_metrics and changed the return to have all the calculated metrics.

I have also hidden some functions as suggested under wifi80211.py and documented the public ones. If you find that more documentation is needed feel free to comment. (Sometimes we are so deeply working on something, that we don't even notice is not obvious for others 😅 )

[BastienTr]

Great! Thanks again for this new features. It must be the most important additions for v0.6 by someone that is not a official collaborator. :+1: