Missing file from hparams.py

[pailozka]

Hi,
I decided to run predict.py to familiarize myself with the model, and the following error appeared:

FileNotFoundError: [Errno 2] No such file or directory: '/path/to/the/SHEPHERD/shepherd/shepherd_data/patients/simulated_patients/disease_split_all_sim_patients_8.9.21_kg.txt'

Here, the test file for test_data mentioned in hparams.py is missing. Could you please add it to the archive? Or I can combine the train and val text files into a single one?

def get_patient_data_args(args, hparams):
    if args.patient_data == "disease_simulated":
        hparams.update({'train_data': f'simulated_patients/disease_split_train_sim_patients_{project_config.CURR_KG}.txt',
                        'validation_data': f'simulated_patients/disease_split_val_sim_patients_{project_config.CURR_KG}.txt', 
                        'test_data': f'simulated_patients/disease_split_all_sim_patients_{project_config.CURR_KG}.txt',
                        'spl': f'simulated_patients/disease_split_all_sim_patients_{project_config.CURR_KG}_spl_matrix.npy',
                        'spl_index': f'simulated_patients/disease_split_all_sim_patients_{project_config.CURR_KG}_spl_index_dict.pkl'
                        })
 

[HelloWorldLTY]

Hi, did you resolve this problem? Do you know what is the minimial file requirement for using this model for inference? Thanks

[pailozka]

Hi, I didn't find test file and used the concatenated file of validation_data and train_data. Here, you can basically use anything, even just a couple of samples.
The input file to run the model should contain a list of candidate genes, HPO terms and ids. It is enough to run prediction tasks with this information, although the data provided by SHEPHERD creators contains much more details.