Vectorize differenced process

[dylanhmorris]

This PR:

• refactors DifferencedProcess to be compatible with vectorized input
• Moves integration of discrete timeseries to its own helper function integrate_discrete(), in pyrenew.math, which DifferencedProcess calls under the hood.
• Adds tests for multidimensional input, including checks that appropriate and informative errors are raised.
• Changes from using numpyro.contrib.control_flow.scan to jax.lax.scan in the integrator, since it's fully deterministic.

Out of scope but planned

• Helper functions for broadcasting / shape checking / etc for multi-dim array dimensions after the first, where the first dimension is time and is scanned along

[codecov]

Codecov Report

Attention: Patch coverage is 96.96970% with 1 line in your changes missing coverage. Please review.

Project coverage is 94.40%. Comparing base (afbb340) to head (53740d8).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
pyrenew/process/differencedprocess.py	92.30%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #446      +/-   ##
==========================================
+ Coverage   94.26%   94.40%   +0.14%     
==========================================
  Files          42       42              
  Lines         994     1001       +7     
==========================================
+ Hits          937      945       +8     
+ Misses         57       56       -1     

Flag	Coverage Δ
unittests	94.40% <96.96%> (+0.14%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[damonbayer]

Out of scope but planned

• Helper functions for broadcasting / shape checking / etc for multi-dim array dimensions after the first, where the first dimension is time and is scanned along

Is there an issue for this?

[dylanhmorris]

Created #451

[damonbayer]

Could be good to have a test that makes use of both the vectorized differenced process and the vectorized AR process to make sure they work together. Perhaps as a test for rtperiodicdiffar. Could be done here or as part of #419

[sbidari]

Changes here look good to me!