Best approach to tweak ascii files directly in python

#! C:/Program Files/Nuke8.0v5/nuke-8.0.v5.dll -nx
version 8.0 v5
define_window_layout_xml {<?xml version="1.0" encoding="UTF-8"?>
<layout version="1.0">
    <window x="0" y="0" w="1936" h="1160" screen="0">
        <splitter orientation="1">
            <split size="1369"/>
            <splitter orientation="1">
                <split size="51"/>
                <dock id="" hideTitles="1" activePageId="Toolbar.1">
                    <page id="Toolbar.1"/>
                </dock>
            <split size="547"/>
            <dock id="" activePageId="Properties.1">
                <page id="Properties.1"/>
            </dock>
        </splitter>
    </window>
</layout>
}
Root {
 inputs 0
 name C:/Users/Admin/Desktop/test.nk
 format "1280 720 0 0 1280 720 1 hd_720"
 proxy_type scale
 proxy_format "1024 778 0 0 1024 778 1 1K_Super_35(full-ap)"
}
Viewer {
 inputs 0
 frame 1
 input_process false
 name Viewer1
 xpos -40
 ypos -9
}
Camera {
 inputs 0
 translate {{curve x1 -5 x50 5} {curve x1 0.3 x50 2} {curve x1 17 x50 30}}
 rotate {{curve} {curve x1 -20 x50 10} {curve}}
 name Camera1
 xpos 570
 ypos -6
}
CheckerBoard2 {
 inputs 0
 centerlinewidth 0
 name CheckerBoard1
 xpos 90
 ypos -6
}
 Cube {
 translate {-2 0.5 0}
 name Cube1
 xpos 250
 ypos 154
}
Cylinder {
 translate {2 1 0}
 name Cylinder1
 xpos 250
 ypos 74
}
Sphere {
 translate {0 0.5 0}
 name Sphere1
 xpos 250
 ypos -6
}

Light {
 inputs 0
 translate {-5 3 10}
 name Point1
 xpos 410
 ypos -6
}

Scene {
 inputs 7
 name Scene1
 xpos 410
 ypos 154
}

ScanlineRender {
 inputs 3
 name ScanlineRender1
 xpos 560
 ypos 175
}
Write {
 file C:/Users/Admin/Desktop/testing/super_####.png
 file_type png
 checkHashOnRead false
 name Write1
 xpos 560
 ypos 219
}

Above is sample ascii file. its simple nuke script. I wanted to edit this file via python currently my immediate need is tweaking Write node file output path. immediate problem is need to verify that am I going in right direction or possibly totally wrong

My psudo approach is something like

1. read file


2. search for string pattern "Write{*}"


3. once found, read details (line by line) etc


4. change information and re-write the file on disk again (new)

QUESTIONS:

1. can I read this file (parse) with some xml way as well, and directly get write node settings? or RegEx is best way to go?


2. currently sample file is small but real file can go to 500Kb easily that is approximately 3000 lines of data. So can we directly tweak a line in ascii files or do we have to read data, then do change in data and re-write all data from scratch on disk?


3. can someone please give some hints, from re module, how to reach chunk blocks of data (node settings) with format name{} format. so I can filter if multiple write nodes exists.

Huge thanks in advance for suggestions , regards,lala