GenBank to FASTA format using regular expressions without Biopython

Question

This is a Biopython alternative with pretty straightforward code. How can I make this more concise?

def genbank_to_fasta():
    file = input(r'Input the path to your file: ')
    with open(f'{file}') as f:
        gb = f.readlines()
        locus = re.search('NC_\d+\.\d+', gb[3]).group()
        region = re.search('(\d+)?\.+(\d+)', gb[2])
        definition = re.search('\w.+', gb[1][10:]).group()
        definition = definition.replace(definition[-1], "")
        tag = locus + ":" + region.group(1) + "-" + region.group(2) + " " + definition
        sequence = ""
        for line in (gb):
            pattern = re.compile('[a,t,g,c]{10}')
            matches = pattern.finditer(line)
            for match in matches:
                sequence += match.group().upper()
        end_pattern = re.search('[a,t,g,c]{1,9}', gb[-3])
        sequence += end_pattern.group().upper()
        print(len(sequence))
        return sequence, tag

Answer 1

Line iteration

    gb = f.readlines()
    locus = re.search('NC_\d+\.\d+', gb[3]).group()
    region = re.search('(\d+)?\.+(\d+)', gb[2])
    definition = re.search('\w.+', gb[1][10:]).group()

    for line in (gb):
        # ...
    end_pattern = re.search('[a,t,g,c]{1,9}', gb[-3])

can be

next(f)
definition = re.search('\w.+', next(f)[10:]).group()
region = re.search('(\d+)?\.+(\d+)', next(f))
locus = re.search('NC_\d+\.\d+', next(f)).group()

gb = tuple(f)

for line in gb:

Since you need gb[-3], you can't get away with a purely "streamed" iteration. You could be clever and keep a small queue of the last few entries you've read, if you're deeply worried about memory consumption.

Debugging statements

Remove this:

    print(len(sequence))

or convert it to a logging call at the debug level.

Compilation

Do not do this:

        pattern = re.compile('[a,t,g,c]{10}')

in an inner loop. The whole point of compile is that it can be done once to save the cost of re-compilation; so a safe option is to do this at the global level instead.