Fix NaN values in inverse Box-Cox transform for Lambda > 0

[raashish1601]

The inverse of boxcox_transform returns NaN instead of 0 for dry pixels when Lambda > 0.

The forward transform sets values below the threshold to zerovalue = threshold - 1 in transformed space. For Lambda > 0 that value is usually below -1 / Lambda. For example, with threshold 0.1 and Lambda = 0.5 it is about -2.37, so Lambda * R + 1 is negative and np.log gives NaN. The inverse then runs R[R < threshold] = zerovalue, but NaN < threshold is False, so these pixels stay NaN.

R = np.array([0.0, 0.05, 1.0])
md = {"transform": None, "unit": "mm/h", "threshold": 0.1, "zerovalue": 0.0, "accutime": 5}
Rt, md = boxcox_transform(R.copy(), md, Lambda=0.5)
boxcox_transform(Rt, md, inverse=True)[0]
# array([nan, nan,  1.])   expected [0, 0, 1]

The fix finds the zeros in transformed space, before the back-transformation. The transform is monotonic, so for finite values this gives the same pixels as before. Lambda = 0 and the dB/sqrt transforms are not affected.

I added a round trip test with zeros for Lambda in 0, 0.5 and 1. It fails on master for 0.5 and 1. test_utils_transformation.py and test_utils_conversion.py pass, and black is clean.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.76%. Comparing base (19e5f0d) to head (219a966).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #577      +/-   ##
==========================================
+ Coverage   84.75%   84.76%   +0.01%     
==========================================
  Files         170      170              
  Lines       15104    15114      +10     
==========================================
+ Hits        12801    12811      +10     
  Misses       2303     2303              

Flag	Coverage Δ
unit_tests	84.76% <100.00%> (+0.01%)	⬆️

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

☔ View full report in Codecov by Harness.
📢 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.

[dnerini]

hi @raashish1601, thanks for the fix, looks good to me!