gh-253 Shrinkage

[JaimeRZP]

New Features

• dices.io.py: implements functionality to move fields to components programmatically. We still need an if loop to detect if we are doing Cls and covariances. Maybe this could be streamlined as a for loop for each axis in the result but I don't know how.
• dices.shrinkage.py: implements the fields-wise methods to compute a Gaussian estimate of the covariance and shrink to it. The shrinkage routines have been streamlined significantly.
• The jackknife tutorial has been extended to cover shrinkage. See the maths in the markdown for the implementation details.

Known Issues

• Bug in the analytical covariance computation. It seems like the entries are being overwritten later on.

[JaimeRZP]

pre-commit.ci autofix

[Copilot]

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Comments suppressed due to low confidence (3)

heracles/dices/jackknife.py:190

• The parameter name 'dict' shadows the built-in Python type. Consider renaming it to something like 'samples' or 'spectra_dict'.

def jackknife_covariance(dict, nd=1):

heracles/dices/jackknife.py:235

• The variable 'a' is used before it is defined. Please check and replace 'a' with the intended variable (likely the array being processed).

    if (a1, b1) == ("POS", "POS"):
                a = np.expand_dims(a, axis=0)

heracles/dices/jackknife.py:237

• The variable 'a' is used before it is defined. Please verify the correct variable to modify and update accordingly.

    if (a2, b2) == ("POS", "POS"):
                a = np.expand_dims(a, axis=1)

[JaimeRZP]

pre-commit.ci autofix