Optional final production model fitting procedure after CV

[PascalIversen]

closes #217
I decided for a proper retuning on the full dataset instead of any hacks. (will approx. take as long as one fold)

also sneaking in tissue mapping tests which I forgot to add.

TODO:
testing

[Copilot]

Pull Request Overview

This PR introduces an optional final production model fitting procedure after cross‐validation and enhances model persistence by adding save/load methods across multiple model implementations. Key changes include:

• Adding a new command-line flag and corresponding logic in the experiment pipeline to train a final model on the full dataset.
• Implementing and testing save/load functionality in various model modules.
• Introducing tissue mapping tests and minor type hint/documentation improvements.

Reviewed Changes

Copilot reviewed 15 out of 15 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/test_tissue_mapping.py	Added tests for tissue mapping functionality with dummy data.
tests/models/test_single_drug_models.py	Added save/load tests for single drug models to verify persistence.
tests/models/test_global_models.py	Extended global model tests with save/load functionality.
tests/models/test_baselines.py	Added persistence tests for baseline models.
drevalpy/utils.py	Introduced a new flag (--final_model_on_full_data) for final model training.
drevalpy/models/drp_model.py	Added abstract save/load methods in the base model class.
drevalpy/models/baselines/singledrug_elastic_net.py	Implemented save/load functionality for the ElasticNet baseline.
drevalpy/models/SimpleNeuralNetwork/*.py	Added save/load methods and updated hyperparameter defaults in neural network models.
drevalpy/models/SRMF/srmf.py	Added methods to save and load the SRMF model parameters/configurations.
drevalpy/models/DIPK/dipk.py	Updated DIPK training to use hyperparameters and added save/load support.
drevalpy/experiment.py	Extended experiment pipeline to include a final model training step and added a new splitting utility.
drevalpy/datasets/map_tissues.py	Refined type hints and ensured proper conversion of file path objects.

Comments suppressed due to low confidence (1)

drevalpy/models/baselines/singledrug_elastic_net.py:1

• Typo in docstring: 'seperately' should be corrected to 'separately'.

"""SingleDrugElasticNet and SingleDrugProteomicsElasticNet classes. Fit an Elastic net for each drug seperately."""

[Copilot]

[nitpick] Consider using a logging mechanism instead of print statements in tests to reduce console clutter during automated runs.

Suggested change

	print("\n=== tissue_mapping.csv ===")
	print(df_out)
	assert "tissue" in df_out.columns
	print("\n=== tissue for CVCL_TEST1 ===")
	print(df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST1", "tissue"])
	print("\n=== tissue for CVCL_TEST2 ===")
	print(df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST2", "tissue"])
	assert df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST1", "tissue"].values[0] == "Lung"
	assert df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST2", "tissue"].values[0] == "Blood"
	# Check the updated dataset has a tissue column
	updated = pd.read_csv(root / ds_name / f"{ds_name}.csv")
	print("\n=== Updated dataset ===")
	print(updated)
	logging.info("\n=== tissue_mapping.csv ===")
	logging.info(df_out)
	assert "tissue" in df_out.columns
	logging.info("\n=== tissue for CVCL_TEST1 ===")
	logging.info(df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST1", "tissue"])
	logging.info("\n=== tissue for CVCL_TEST2 ===")
	logging.info(df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST2", "tissue"])
	assert df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST1", "tissue"].values[0] == "Lung"
	assert df_out.loc[df_out["cellosaurus_id"] == "CVCL_TEST2", "tissue"].values[0] == "Blood"
	# Check the updated dataset has a tissue column
	updated = pd.read_csv(root / ds_name / f"{ds_name}.csv")
	logging.info("\n=== Updated dataset ===")
	logging.info(updated)

[JudithBernett]

lgtm