Consistent norm computation

[Yash-10]

From observations in #1433.

Thanks!

[Yash-10]

Trying to investigate the docs error...

[codecov]

Codecov Report

Merging #1442 (b489ec6) into main (9797eb0) will decrease coverage by 0.09%.
The diff coverage is 100.00%.

❗ Current head b489ec6 differs from pull request most recent head d01ad0b. Consider uploading reports for the commit d01ad0b to get more accurate results

@@            Coverage Diff             @@
##             main    #1442      +/-   ##
==========================================
- Coverage   91.93%   91.83%   -0.10%     
==========================================
  Files          93       95       +2     
  Lines        4400     4448      +48     
  Branches      422      430       +8     
==========================================
+ Hits         4045     4085      +40     
- Misses        267      273       +6     
- Partials       88       90       +2     

Impacted Files	Coverage Δ
src/poliastro/_math/linalg.py	100.00% <100.00%> (ø)
src/poliastro/core/czml_utils.py	100.00% <100.00%> (ø)
src/poliastro/core/elements.py	89.62% <100.00%> (ø)
src/poliastro/core/events.py	100.00% <100.00%> (ø)
src/poliastro/core/flybys.py	100.00% <100.00%> (ø)
src/poliastro/core/iod.py	90.00% <100.00%> (ø)
src/poliastro/core/maneuver.py	100.00% <100.00%> (ø)
src/poliastro/core/perturbations.py	100.00% <100.00%> (ø)
src/poliastro/core/propagation/vallado.py	87.50% <100.00%> (+0.40%)	⬆️
src/poliastro/core/spheroid_location.py	100.00% <100.00%> (ø)
... and 10 more

---

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 9797eb0...d01ad0b. Read the comment docs.

[Yash-10]

It seems if the input array contains extra unwanted dimensions, it fails. np.squeeze() is not yet an option in numba, so maybe something else is needed.

(Maybe support to different types of norms, i.e. L1, L2, etc could also be kept just as in the case of np.linalg.norm)

[astrojuanlu]

I see that norm(Ephem.from_body(Venus, flyby_1_time).rv()[0]) fails, because the resulting vector has ndim == 2. On the other hand, both norm(Ephem.from_body(Venus, flyby_1_time).rv()[0][0]) and norm(Ephem.from_body(Venus, flyby_1_time).rv(flyby_1_time)[0]) succeed.

We want norms of vectors, not norm of matrices, so I think this failure is correct. It forces the user to put the vector in the correct shape though. Addmittedly, Ephem.from_body(Venus, flyby_1_time).rv(flyby_1_time)[0] is not a very convenient way of computing the position of a planet at a given time, we are probably overusing Ephem there. But I suggest we make the small change in the notebook and we continue.

[astrojuanlu]

Looks perfect to me! When we fix the notebook that is failing, we can merge.

[astrojuanlu]

Thanks a lot @Yash-10 !