Simplify API by using simulation structure

[albangossard]

• Simplification of the API by removing useless arguments and keeping only the simulation parameters
• Fix several inconsistencies in the computation of the average speed
• Use the save_everystep flag to store results of the forward simulation
• Add script to generate jld2 files used for the test of make_thickness_video in Sleipnir, cf Improve visualization tools with a video generation Sleipnir.jl#79

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Please upload report for BASE (main@4ae1c05). Learn more about missing BASE report.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #64   +/-   ##
=======================================
  Coverage        ?   89.07%           
=======================================
  Files           ?        9           
  Lines           ?      412           
  Branches        ?        0           
=======================================
  Hits            ?      367           
  Misses          ?       45           
  Partials        ?        0           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[JordiBolibar]

Nice clean-up. Just a couple of minor comments. Feel free to merge it afterwards!

[JordiBolibar]

This makes me think that @facusapienza21 already created a GitHub action which syncs files among repos. We could potentially use that one to automatically sync these files with Sleipnir, so changes are propagated.

[JordiBolibar]

We can open an issue for that and implement it later on, there's no rush :)

[JordiBolibar]

Nonetheless, as we discussed, this might be overkill. The plotting test should not be affected by the results of the simulation. In the end, it's just a matrix, so even if the simulation is outdated it should work. The only breaking aspects should come from Sleipnir itself (i.e. changing the structs by adding or removing parameters), and that should be already covered and identified within Sleipnir.

[albangossard]

You're right this is completely overkill. The problem with the plotting test is that to instantiate Results we need an iceflow model which is implemented in Huginn. In order to circumvent this problem I replaced results by H in plot_glacier_vid. In the future we should revert that change once Results can be instantiated without an iceflow model, see #66

[facusapienza21]

Nice work! Feel free to merge!