Increase speed of runs

[raphaelshirley]

Problem & Solution Description (including issue #)

We are going to propose some changes to the defaults to optimise for speed for runs on 20 billion objects.

Closes #78

• reduce redshift grid to 0.02
• reduce number of extinction laws
• add a default config for estimating masses

Code Quality

• My code follows the code style of this project
• I have written unit tests or justified all instances of #pragma: no cover; in the case of a bugfix, a new test that breaks as a result of the bug has been added
• My code contains relevant comments and necessary documentation for future maintainers; the change is reflected in applicable demos/tutorials (with output cleared!) and added/updated docstrings use the NumPy docstring format
• Any breaking changes, described above, are accompanied by backwards compatibility and deprecation warnings

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.36%. Comparing base (d570152) to head (3f1eb06).

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #79      +/-   ##
==========================================
+ Coverage   92.25%   92.36%   +0.10%     
==========================================
  Files           2        2              
  Lines         142      144       +2     
==========================================
+ Hits          131      133       +2     
  Misses         11       11              

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[johannct]

Only one bump extinction law perhaps? We need to check that it is ok though. EM_DISPERSION is almost guaranteed to have little effect on 6 bands.

[raphaelshirley]

We may want to add a second physical parameter config set. @olivierilbert can you send me a standard physical parameter config?

[olivierilbert]

Hello,
sorry, I didn't follow this development. Is the assesment that the code is to slow done after our reduction of the emission line dispersion, etc ? And we can still gain a lot of speed by just reducing the number of attenuation curve and E(B-V).

[raphaelshirley]

All I want to do is set a recommended config with a reasonable balance of speed and results. Currently I am just applying the changes we made in the last week. I will merge when we have settled on a recommendation but it would be good to also add the recommended physical parameter settings when we have them.

[olivierilbert]

For the physical parameters, I just did a run on the DP1 by curiosity. Using ZFIX YES and the full BC03 library, I get 20s for 8500 sources with my laptop for one CPU (M3 pro). of course, we need to time all these runs in a much more accurate way. I will send you an example of notebook. But we can also limit the size of the library (e.g. one attenuation law rather than 2).

[olivierilbert]

I can do some tests and decide what to remove. But on my side, I need to see how much I should remove. How far are we of what we need?