createCSVfromSeparatedPheno.py

import csv
import sys
import argparse
import os

def parseArgs():
    parser = argparse.ArgumentParser(description='Parse ADSP files that were previously separated by phenotype.')
    parser.add_argument('-i', '--input_dir', default='separated_by_phenotype',type=str, required=False, help='(input) Input directory with 3 files (more allowed, but will not be read): cases.csv, controls.csv, and mci.csv')
    parser.add_argument('-o', '--output', type=str, default="ml_data.csv",required=False,help='(output) Path to output CSV with formatted ADNI data')
    args = parser.parse_args()

    if not os.path.isdir(args.input_dir):
        print("Must input a valid directory path with the -i option")
        sys.exit()
    return args

def main():
    args=parseArgs()
    headerWritten=False
    for file_name in ['cases.csv','controls.csv','mci.csv']:
        input_file = args.input_dir +"/" + file_name
        #input_file = sys.argv[1]
        #output_file = sys.argv[2]
        selected_headers = ['SUBJID','Sex', 'Age', 'APOE_reported', 'Race'] # Replace with your desired headers
        
        with open(input_file, 'r', newline='') as infile, \
             open(args.output, 'a', newline='') as outfile:
        
            reader = csv.DictReader(infile)
            
            # Filter the fieldnames to include only the selected headers
            writer_fieldnames = [header for header in selected_headers if header in reader.fieldnames]
            writer_fieldnames2 = writer_fieldnames[:]
            writer_fieldnames2.append("Class")
            writer = csv.DictWriter(outfile, fieldnames=writer_fieldnames2)
            
            if not headerWritten:
                writer.writeheader() # Write the header row to the new file
                headerWritten=True
        
            for row in reader:
                # Create a new dictionary with only the selected columns
                new_row = {header: row[header] for header in writer_fieldnames}
                if file_name == 'cases.csv':
                    new_row['Class'] ='AD' #AD, Control, PSP, etc.
                elif file_name == 'mci.csv':
                    new_row['Class'] ='MCI' 
                elif file_name == 'controls.csv':
                    new_row['Class'] ='Control' 
                if new_row['Age']=='90+':
                    new_row['Age']='90'
                if not any(value == 'NA' for value in new_row.values()):
                    writer.writerow(new_row)

if __name__ == '__main__':
    main()

#Make sure to get rid of lines with NA values afterwards.