Integrate create_plots with training_wrapper_functions.py

[niekdejonge]

The create plots functions were not yet functional, due to a few bugs.

Functional changes:

• [x] Fix issue with calculate_tanimoto_scores for plotting. The calculations did not work if the df was not symmetric.
• [x] Add and integrate create all plots with training wrapper functions

Internal code changes:

• [x] Added standard file names to SettingsMS2Deepscore
• [x] Make load_negative_train_split (and positive) only return one file at a time, to reduce the risk of assigning the wrong train/test/val variable.
• [x] Add BenchmarkingResultsFileNames for storing the file names.

Added tests:

• [x] Make train_ms2ds_wrapper test work. Now it doesn't work, since there are no positive mode spectra and plotting doesn't work. Check that files are created
• [x] Add tests for plotting
• [x] Add test for calculating tanimoto scores for plotting
• [x] Create create_test_spectra.py to store common util functions used by tests, to make updating and reuse easier.

Folder restructuring

• [x] Created a new folder for all wrapper functions; split plotting function into wrapper functions and tanimoto calculation.
• [x] Move compute_fingerprints_for_training to spectrum_pair_selection
• [x] Store StoreTrainingData in separate file from train_ms2ds_wrapper

To do future PR:

• Update Changelog. A more general message that summarizes that we now have a wrapper that does all training, storing all in between changes would be too detailed.
• Maybe restructure tests in folders as well
• Maybe remove the old data generators
• Store the MS2Deepscore model settings as well (done in #162)

[florian-huber]

@niekdejonge I guess this and a few of the following PRs can be closed because they were part of #162 , right?

[niekdejonge]

@florian-huber Yes they can be closed!