Remove redundancy between drevalpy package and pipeline - Githubissues

nf-core / drugresponseeval

Pipeline for testing drug response prediction models in a statistically and biologically sound way.

MIT License

4 stars 0 forks source link

Remove redundancy between drevalpy package and pipeline #8

Closed JudithBernett closed 2 months ago

JudithBernett commented 3 months ago

Description of feature

Some functionality is now duplicated - shouldn't be this way

[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:PARAMS_CHECK
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:RUN_CV:LOAD_RESPONSE
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:RUN_CV:CV_SPLIT
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:RUN_CV:HPAM_SPLIT
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:RUN_CV:TRAIN_AND_PREDICT_CV
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:RUN_CV:EVALUATE_FIND_MAX
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:MODEL_TESTING:PREDICT_FULL
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:MODEL_TESTING:RANDOMIZATION_SPLIT
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:MODEL_TESTING:RANDOMIZATION_TEST
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:MODEL_TESTING:ROBUSTNESS_TEST
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:MODEL_TESTING:EVALUATE_FINAL
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:MODEL_TESTING:COLLECT_RESULTS
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:VISUALIZATION:DRAW_VIOLIN
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:VISUALIZATION:DRAW_HEATMAP
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:VISUALIZATION:DRAW_CORR_COMP
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:VISUALIZATION:DRAW_REGRESSION
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:VISUALIZATION:SAVE_TABLES
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:VISUALIZATION:WRITE_HTML
[x] NFCORE_DRUGRESPONSEEVAL:DRUGRESPONSEEVAL:VISUALIZATION:WRITE_INDEX

JudithBernett commented 3 months ago

@PascalIversen : RUN_CV:TRAIN_AND_PREDICT_CV, MODEL_TESTING:RANDOMIZATION_SPLIT, train_and_predict_final.py -> MODEL_TESTING:PREDICT_FULL, MODEL_TESTING:RANDOMIZATION_TEST, MODEL_TESTING:ROBUSTNESS_TEST fix train_and_predict_cv + train_and_predict_final / experiment.py: This:

model_class = MODEL_FACTORY[args.model_name]
model = model_class()

should at least be one function, right --> it also needs to be fixed in randomization_split.py then