python numpy convert string to complex number

Python natively supports complex numbers, but it uses the engineering convention of j for the imaginary unit instead of the mathematician's i.

lr = ['3+4i', '3.123+2.43i', '0.23454-1.2134i', '0.6543+0.76543i', '1.12324+0.6543i']
for s in lr:
    s = s.replace('i', 'j')
    v = complex(s)
    print(v, abs(v))

output

(3+4j) 5.0
(3.123+2.43j) 3.9570227444380457
(0.23454-1.2134j) 1.2358594465391282
(0.6543+0.76543j) 1.006971486637035
(1.12324+0.6543j) 1.2999140693138143

We can create a list of the moduli using a list comprehension:

moduli = [abs(complex(s.replace('i', 'j'))) for s in lr]
print(moduli)

output

[5.0, 3.9570227444380457, 1.2358594465391282, 1.006971486637035, 1.2999140693138143]

Here's a variation on the 'replace i with j' solution:

In [182]: lr=['3.123+2.43i','0.23454-1.2134i','0.6543+0.76543i','1.12324+0.6543i']

Can make an array of strings:

In [183]: np.array(lr)
Out[183]: 
array(['3.123+2.43i', '0.23454-1.2134i', '0.6543+0.76543i',
       '1.12324+0.6543i'], 
      dtype='<U15')

but not complex - at least not directly

In [184]: np.array(lr,np.complex)
...
TypeError: a float is required

but there is a function that performs string replace on array elements:

In [185]: a=np.array(lr)
In [186]: np.char.replace(a,'i','j')
Out[186]: 
array(['3.123+2.43j', '0.23454-1.2134j', '0.6543+0.76543j',
       '1.12324+0.6543j'], 
      dtype='<U15')

Now that array of strings can be cast as complex

In [187]: np.char.replace(a,'i','j').astype(np.complex)
Out[187]: 
array([ 3.12300+2.43j   ,  0.23454-1.2134j ,  0.65430+0.76543j,
        1.12324+0.6543j ])