[FIXED] ValueError: could not convert string to float -- how to convert list of numpy array to float?

Issue

I am using protein-ligand affinity value file as input that return its values to list. Code is

def sep_input_file(input_file):
    lines = preprocessing(read_file(open(input_file,"r")))
    
    if lines.shape[1] == 2:
        return list(lines[:,0]), list(lines[:,1]), list(np.ones(len(lines)))
    elif lines.shape[1] == 3:
        return list(lines[:,0]), list(lines[:,1]), list(np.array(lines[:,2], dtype = np.float32))

Input file is:

protein	ligand	affinity
./data/complexes/2hyyProtein.mol2	./data/complexes/2hyyLigand.mol2	10.2

Please suggest me how to resolve it?

Getting the error

error:

return list(lines[:,0]), list(lines[:,1]), list(np.array(lines[:,2], dtype = np.float32))
ValueError: could not convert string to float:

Solution

There are most likely strings that can't be parsed into floats in the affinity column. Try replacing

list(np.array(lines[:,2], dtype = np.float32))

with

result = []
for data in lines[:,2]:
    try:
        value = float(data.replace(',', '.').strip())
    except (ValueError, TypeError):
        value = None
    result.append(value)

There will be None values in the result where the parsing fails.

Answered By - Michal Racko