fix(l2g_predictions): annotate based on list of features + filter out missing annotation

[ireneisdoomed]

✨ Context

The feature matrix contains all features we have developed for L2G.
However, that doesn't mean that we want to use all of them during training. Right now, this is true for isProteinCoding.

🛠 What does this PR implement

• add_locus_to_gene_features didn't take into account the list of features used for training. Now it does
• We were filtering out features with nulls, instead of features equal to 0. Now each prediction only include features that have annotation

🙈 Missing

🚦 Before submitting

• Do these changes cover one single feature (one change at a time)?
• Did you read the contributor guideline?
• Did you make sure to update the documentation with your changes?
• Did you make sure there is no commented out code in this PR?
• Did you follow conventional commits standards in PR title and commit messages?
• Did you make sure the branch is up-to-date with the dev branch?
• Did you write any new necessary tests?
• Did you make sure the changes pass local tests (make test)?
• Did you make sure the changes pass pre-commit rules (e.g poetry run pre-commit run --all-files)?

[project-defiant]

Out of curiosity, how many features did we add additionally to the predictions with last run that we do not care about?

[project-defiant]

It extracts the features based on features_list from L2GFeatureMatrix and reannotates the locusToGeneFeatures column with the map constructed from extracted features if the column exists or
creates the column when it is missing from the schema.

[project-defiant]

List of features to extract from feature matrix.

[ireneisdoomed]

@project-defiant One: isProteinCoding, which is true for everyone right now.

Thanks for the review!