Fixed plapt_cli.py to apply score_candidates and predict_affinity

[csecker]

Previously, plapt_cli.py only called the predict_affinity function, which seems to be for scoring protein-ligand pairs only (according to README.md). It was also passing the first target sequence only (args.target[0]) in predict_affinity, which raises a ValueError when the function checks for the length of the input targets and input molecules.

plapt_cli.py now uses predict_affinity for scoring protein-ligand pairs, if there are the same number of targets and molecules. If there are more or less molecules than targets, score_candidates is used (scoring all molecules on the first target provided).