Question

I have a text file in this format:

abc? cdfde" nhj.cde' dfwe-df$sde.....  

How can i ignore all the special characters, blanks, numbers, end of the lines, etc and write only the characters in another file?For example, the above file becomes

abccdfdenhjcdedfwedfsde.....  

And from this output file,

• Should able to read single character by character till the end of file.
• Should be able to read two characters at a time, like ab,bc,cc,cd,df,… from above file
• Should be able to read three characters at a time, like abc,bcc,ccd,cdf,… from the above file

First of all, how can i read only characters and write to external file?
I can read single character by character by using f.read(1) till end of file.How can i apply this to read 2,3 chars at a time, that too skipping only one character(that is, if i have abcd, i should read ab,bc,cd but not ab,cd(this, i think can be done by f.read(2))). Thanks. I am doing this for cryptanalysis work to analyze ciphertexts by frequency.

Answer 1

If you need to peek ahead (read a few extra characters at a time), you need a buffered file object. The following class does just that:

import io

class AlphaPeekReader(io.BufferedReader):
    def readalpha(self, count):
        "Read one character, and peek ahead (count - 1) *extra* characters"
        val = [self.read1(1)]

        # Find first alpha character
        while not val[0].isalpha():
            if val == ['']:
                return ''  # EOF
            val = [self.read1(1)]


        require = count - len(val)
        peek = self.peek(require * 3)  # Account for a lot of garbage
        if peek == '':  # EOF
               return val[0]

        for c in peek:
            if c.isalpha():
               require -= 1
               val.append(c)
               if not require:
                   break

        # There is a chance here that there were not 'require' alpha chars in peek
        # Return anyway.
        return ''.join(val)

This attempts to find extra characters beyond the one character you are reading, but doesn’t make a guarantee it’ll be able to satisfy your requirements. It could read fewer if we are at the end of the file or if there is a lot of non-alphabetic text in the next block.

Usage:

with AlphaPeekReader(io.open(filename, 'rb')) as alphafile:
    alphafile.readalpha(3)

Demo, using a file with your example input:

>>> f = io.open('/tmp/test.txt', 'rb')
>>> alphafile = AlphaPeekReader(f)
>>> alphafile.readalpha(3)
'abc'
>>> alphafile.readalpha(3)
'bcc'
>>> alphafile.readalpha(3)
'ccd'
>>> alphafile.readalpha(10)
'cdfdenhjcd'
>>> alphafile.readalpha(10)
'dfdenhjcde'

To use the readalpha() calls in a loop, where you get each and every character separately plus the two next 2 bytes, use the iter() with a sentinel:

for alpha_with_extra in iter(lambda: alphafile.readalpha(3), ''):
    # Do something with alpha_with_extra