Python >2.7 function to replace a string in a file using mmap.

replace_string_in_file.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
from __future__ import (unicode_literals,
                        with_statement)
import mmap
import os.path


def replace_placeholder_in_file(filepath, placeholder, text):
    """Replace `placeholder` in file `filename` with `text` if found.
    
    Do nothing, if the file given by `filepath` does not exist. Replaces are
    done in-place. If `text` is longer than `placeholder` the file is expanded.
    If `text` is shorter the file is truncated by the amount of excessive bytes.
    Otherwise there is just a replacement.
    
    Examples:
        >>> replace_placeholder_in_file("somefile.txt", "Hallo", "Hello")

    Args:
        string filepath - The path to the file, where the placeholder shall be
                          replaced.
        string placeholder - The placeholder string, for which to look inside
                             the file given by `filepath`. This string will get
                             replaced by `text` if found inside the file.
        string text - The replacement text.
    """
    if not os.path.isfile(filepath):
      return

    placeholder_size = len(placeholder.encode("utf-8"))
    text_size = len(text.encode("utf-8"))
    new_bytes = text_size - placeholder_size
    filesize_before_modification = os.path.getsize(filepath)

    if new_bytes > 0:
        # Pad file to right size
        with open(filepath, 'a') as f:
            f.write(new_bytes * b'\0')

    with open(filepath, "r+b") as f:
        mm = mmap.mmap(f.fileno(), 0, access=mmap.ACCESS_WRITE)
        begin = mm.find(placeholder.encode("utf-8"))
        if begin == -1:
            # Nothing to replace here.
            return

        # mm.resize(filesize + NEW_BYTES) does not work on macOS
        end_of_placeholder = begin + placeholder_size
        following_bytes_length = filesize_before_modification -\
            end_of_placeholder
        mm.move(end_of_placeholder + new_bytes,
                end_of_placeholder,
                following_bytes_length)
        mm.seek(begin)
        mm.write(text.encode("utf-8"))
        mm.flush()
        mm.close()
        if new_bytes < 0:
            f.seek(new_bytes, os.SEEK_END)
            f.truncate()