Changes to _get_iv to compute covariant errors

[gemmahuai]

Added changes to the function _get_iv in optimize.py

• placed each unit flux model in a corresponding image
• calculated Fisher information matrix F
• found the covariance matrix C by inverting F
• read off the variance from the diagonal values in C.

The algorithm passed the z-score test where as I fixed the primary source at the center and marched a secondary source along the central x-axis (approaching, overlapping, and then leaving the primary source), I repeated the tractor fit 500 times and plotted a histogram of z-score for each position. The z-score histogram follows a Gaussian of sigma~1 distribution except when the two sources completely overlap. So this is fairly stable. For a three-source test, it's also reasonable.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

❌ Patch coverage is 0% with 21 lines in your changes missing coverage. Please review.
✅ Project coverage is 45.88%. Comparing base (48fbef9) to head (e64f1ae).
⚠️ Report is 12 commits behind head on main.

Files with missing lines	Patch %	Lines
tractor/optimize.py	0.00%	21 Missing ⚠️
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #106      +/-   ##
==========================================
- Coverage   45.99%   45.88%   -0.12%     
==========================================
  Files          27       27              
  Lines        7085     7103      +18     
==========================================
  Hits         3259     3259              
- Misses       3826     3844      +18     

☔ 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.

[dstndstn]

don't worry about the code-coverage report :)

[gemmahuai]

Okay, thanks! Please let me know if there is any problem.

[coveralls]

coverage: 45.88% (-0.1%) from 45.994%
when pulling e64f1ae on variance_update
into 48fbef9 on main.

[gemmahuai]

Hi Dan, I updated the Fisher matrix calculation. Now it's faster by 10 times. Hopefully the version is going to pass all tests. :)