PythonScript to replace TextFX Rewrap function?

Reiner Bühl

I have until recently still used the long abandoned TextFx plugin just for one function: The Rewrap function. Unfortunately the Plugin is now completely unusable in the last version of Notepad++. According to the transition FAQ, this function can be replaced by a PytonScript function. Unfortunately I don’t know anything about PythonScript. Is there already such a script available somewhere that implements the rewraping of text?
I would greatly appreciate pointers to such a script or something close functionality wise.

Alan Kilborn

@Reiner-Bühl

I haven’t used or attempted to use TextFX in years. What exactly does “Rewrap” do?

Ekopalypse

@Reiner-Bühl

I’m using something like this

from textwrap import wrap
from Npp import editor, notepad

def rewrap(text, pos):
	return wrap(text, pos, expand_tabs=False, replace_whitespace=False, break_on_hyphens=False)

def main():
	pos = int(notepad.prompt('Wrap at position:', 'ReWrap', '72'))
	if pos < 8 or pos > 2048:
		pos = 72

	start, end = editor.getUserLineSelection()
	start_pos = editor.positionFromLine(start)
	end_pos = editor.getLineEndPosition(end)
	
	rewrapped = rewrap(editor.getRangePointer(start_pos, end_pos-start_pos), pos)
	eol = {0:'\r\n', 1:'\r', 2:'\n'}[editor.getEOLMode()]
	editor.setTarget(start_pos, end_pos)
	editor.beginUndoAction()
	editor.replaceTarget(eol.join(rewrapped))
	editor.endUndoAction()

main()

Instead of copying the position to the clipboard, a prompt is opened to get the position for the wrap.

Note, if nothing is selected the whole document is wrapped.

PeterJones

@Ekopalypse ,

Thanks for posting that. I will now link the FAQ’s “rewrap” entry here – that’s exactly what I was hoping would happen when I created that FAQ: as people publish implementations of the various features, I will link the FAQ to that implementation (making the FAQ something of a “table of contents”).

Ekopalypse

@PeterJones

maybe we can wait and see if this is really what is assumed to be the replacement? I’m not 100% sure that it does exactly what TextFX does. And the code is a bit rough (a one line function?? what the hell) :-D
I just copied the appropriate parts from my slightly longer code.

Ekopalypse

@Ekopalypse

oopsss - to late :-D

PeterJones

@Ekopalypse said in PythonScript to replace TextFX Rewrap function?:

oopsss - to late :-D

Sorry. ;-)

I think it’s a “good enough” first implementation; if it gets refined, I will just update the link.

PeterJones

@Ekopalypse ,

So I spun up an 8.3.3-32bit with TextFX to see what it did.

With input

this is a group 
of short lines 
that will be 
merged into a 
line that's 
around 72 char 
long

the TextFX “rewrap” will turn it into

this is a group of short lines that will be merged into a line that's 
around 72 char long

but your script will do

this is a group
of short lines
that will be
merged into a
line
that's
around 72 char
long

When I looked at the wrap function, I thought maybe changing to replace_whitespace=True … but that appears to do 1 space for each newline character, so CRLF becomes two spaces:

123456789x123456789x123456789x123456789x123456789x123456789x123456789x123456789x
this is a group  of short lines  that will be  merged into a  line
that's  around 72 char  long

(number line added to make the double space obvious)

Is there an option that will collapse the \h*\R to a single space rather than n spaces for each horizontal or vertical character?

Ekopalypse

@PeterJones said in PythonScript to replace TextFX Rewrap function?:

this is a group of short lines that will be merged into a line that’s
around 72 char long

ahh - you see, I always “split” the text and never actually “join” lines.
There is a fill method that does the “joining” - let’s see if that needs to be called in advance.

Alan Kilborn

It may be worth pointing out that if you always want to hard-wrap lines at a the same column, that THIS THREAD treats the topic fairly exhaustively, and requires no external scripting.

PeterJones

@Ekopalypse said in PythonScript to replace TextFX Rewrap function?:

There is a fill method that does the “joining” - let’s see if that needs to be called in advance.

It might not have been the way you thought of, but I used an re.sub() in the rewrap function which first merges the equivalent of \h*\R into a single space. With that, and using the replace_whitespace=True, the behavior of the script seems to match the essential nature of the Rewrap command from TextFX:

# encoding=utf-8
"""
PythonScript replacement of TextFX>Edit>Rewrap
https://community.notepad-plus-plus.org/post/78161
Author: @Ekopalypse , with input from @PeterJones
"""
from Npp import editor, notepad
from textwrap import wrap
import re

def rewrap(text, pos):
	joined = re.sub(r'\h*(\r\n|\r|\n)', " ", text)
	return wrap(joined, pos, expand_tabs=False, replace_whitespace=True, break_on_hyphens=False)

def main():
	pos = int(notepad.prompt('Wrap at position:', 'ReWrap', '72'))
	if pos < 8 or pos > 2048:
		pos = 72

	start, end = editor.getUserLineSelection()
	start_pos = editor.positionFromLine(start)
	end_pos = editor.getLineEndPosition(end)

	rewrapped = rewrap(editor.getRangePointer(start_pos, end_pos-start_pos), pos)
	eol = {0:'\r\n', 1:'\r', 2:'\n'}[editor.getEOLMode()]
	editor.setTarget(start_pos, end_pos)
	editor.beginUndoAction()
	editor.replaceTarget(eol.join(rewrapped))
	editor.endUndoAction()

main()

Reiner Bühl

@Ekopalypse My use case is between the two. I have texts that consist of long and short lines, sometimes with excessive white space, that I want to normalize to 78 character text.

Reiner Bühl

@PeterJones said in PythonScript to replace TextFX Rewrap function?:

encoding=utf-8

“”"
PythonScript replacement of TextFX>Edit>Rewrap
https://community.notepad-plus-plus.org/post/78161
Author: @Ekopalypse , with input from @PeterJones
“”"
from Npp import editor, notepad
from textwrap import wrap
import re

def rewrap(text, pos):
joined = re.sub(r’\h*(\r\n|\r|\n)', " ", text)
return wrap(joined, pos, expand_tabs=False, replace_whitespace=True, break_on_hyphens=False)

def main():
pos = int(notepad.prompt(‘Wrap at position:’, ‘ReWrap’, ‘72’))
if pos < 8 or pos > 2048:
pos = 72

start, end = editor.getUserLineSelection()
start_pos = editor.positionFromLine(start)
end_pos = editor.getLineEndPosition(end)

rewrapped = rewrap(editor.getRangePointer(start_pos, end_pos-start_pos), pos)
eol = {0:‘\r\n’, 1:‘\r’, 2:‘\n’}[editor.getEOLMode()]
editor.setTarget(start_pos, end_pos)
editor.beginUndoAction()
editor.replaceTarget(eol.join(rewrapped))
editor.endUndoAction()

main()

I tried that code but it does not keep the paragraphs separated. In TextFX, the rewrap did only join if there was no blank line between two lines. One or more blank lines where treated as a paragraph break and converted to one blank line to keep the paragraphs intact. Can the code be changed to do that?

Ekopalypse

@Reiner-Bühl

I will try to understand what the TextFX code does and port to Python accordingly. I will post an updated code later today or tomorrow at the latest.

PeterJones

@Reiner-Bühl ,

I think this accomplishes your goal:

# encoding=utf-8
"""
PythonScript replacement of TextFX>Edit>Rewrap
https://community.notepad-plus-plus.org/post/78177
Author: @Ekopalypse , with input from @PeterJones
"""
from Npp import editor, notepad
from textwrap import wrap
import re

def rewrap(text, pos, eol):
    paragraphed = re.sub(eol+eol, u'\u00B6', text)
    joined = re.sub(r'\h*(\r\n|\r|\n)', " ", paragraphed)
    unparagraphed = re.sub(u'\u00B6', eol+eol, joined)
    retlist = []
    for linetext in unparagraphed.splitlines():
        if linetext == '':
            retlist.append('')

        for partial in wrap(linetext, pos, expand_tabs=False, replace_whitespace=False, break_on_hyphens=False):
            retlist.append(partial)

    return retlist

def main():
    pos = int(notepad.prompt('Wrap at position:', 'ReWrap', '72'))
    if pos < 8 or pos > 2048:
        pos = 72

    start, end = editor.getUserLineSelection()
    start_pos = editor.positionFromLine(start)
    end_pos = editor.getLineEndPosition(end)

    eol = {0:'\r\n', 1:'\r', 2:'\n'}[editor.getEOLMode()]
    rewrapped = rewrap(editor.getRangePointer(start_pos, end_pos-start_pos), pos, eol)
    editor.setTarget(start_pos, end_pos)
    editor.beginUndoAction()
    editor.replaceTarget(eol.join(rewrapped))
    editor.endUndoAction()

main()

The input text selection

these are two really long paragraphs that have lots and lots and lots
and lots and lots and lots and lots and lots of words

this is the second of the really long paragraphs that have lots and lots
and lots and lots and lots and lots and lots and lots of words

when run with a value of 16 will end up like:

these are two
really long
paragraphs that
have lots and
lots and lots
and lots and
lots and lots
and lots and
lots of words

this is the
second of the
really long
paragraphs that
have lots and
lots and lots
and lots and
lots and lots
and lots and
lots of words

and that output, when selected and run again with 72, will end up back as the original.

Reiner Bühl

@PeterJones the second version of the code does not work if I mark a whole text. If I mark only one paraghraph, then it does work but only for english texts. As soon as I run it against a German text with Umlauts (ä,ö,ü, etc.) (UTF-8 or UTF-8-BOM) it converts the Umlauts to strange characters and paragraph breaks.

I tried it with this example:

Um die sechste Morgenstunde des 3. Juli dieses 
Jahres war ich gerade, nichts Böses ahnend, dabei, meine Petunien zu begießen, als ich einen großen, bartlosen, blonden jungen Mann bei mir eintreten sah, geschmückt mit einer goldenen Brille, das Haupt bedeckt mit einer 
deutschen Schirmmütze. Trübselig, wie ein Segel längs des Mastes, wenn der Wind sich gelegt hat, baumelte ein weiter Überzieher aus einem sehr dauerhaften englischen Stoff um seine Person. Handschuhe trug er nicht; seine rohledernen Schuhe hatten derartig mächtige, breite Sohlen, daß deren Rand den Fuß mit einer Art Trottoir umgab. In einer 
Seitentasche, ungefähr über dem Herzen, machte sich, unter dem glänzenden Stoff vage ihre Form abzeichnend, eine große Porzellanpfeife bemerkbar. Nicht einmal im Traume wäre ich darauf verfallen, den Unbekannten zu fragen, ob 
er an einer der deutschen Universitäten studiert habe. Ich setzte meine Gießkanne hin und begrüßte ihn sofort auf deutsch mit einem schönen »Guten Morgen!«.

»Mein Herr«, erwiderte er französisch, 
wenn auch mit einem erbärmlichen Akzent, 
»ich heiße Hermann Schultz; ich habe gerade 
einige Monate in Griechenland verbracht, wo 
übrigens Ihr Buch mein ständiger Reisebegleiter war.«

Ekopalypse

@Reiner-Bühl

an updated version, tested with utf8 and ansi

def rewrap(text, pos, eol):
    paragraphed = re.sub(eol+eol, '\0xB6', text)
    joined = re.sub(r'\s*(\r\n|\r|\n)', " ", paragraphed)

    unparagraphed = re.sub('\0xB6', eol+eol, joined)
    retlist = []
    for linetext in unparagraphed.splitlines():
        if linetext == '':
            retlist.append('')

        for partial in wrap(linetext, pos, expand_tabs=False, replace_whitespace=False, break_on_hyphens=False):
            retlist.append(partial.lstrip())


    return retlist

def main():
    _prompt = "Wrap at position:"
    _title = 'ReWrap'
    default_pos = 78
    if editor.getSelectionEmpty():
        _prompt = "ATTENTION!! - Since nothing is selected, the WHOLE document is rewrapped.\n" + _prompt
        _title = "ATTENTION!! - " + _title
    pos = notepad.prompt(_prompt, _title, default_pos)
    if pos is None:
        return  # cancelled
    else:
        pos = int(pos)

    start, end = editor.getUserLineSelection()
    start_pos = editor.positionFromLine(start)
    end_pos = editor.getLineEndPosition(end)

    eol = {0:'\r\n', 1:'\r', 2:'\n'}[editor.getEOLMode()]
    rewrapped = rewrap(editor.getRangePointer(start_pos, end_pos-start_pos), pos, eol)
    editor.setTarget(start_pos, end_pos)
    editor.beginUndoAction()
    editor.replaceTarget(eol.join(rewrapped))
    editor.endUndoAction()

main()

The use of the Unicode B6 seemed to confuse the python2 re engine, and furthermore python re does not know \h, \s had to be used instead.

Ekopalypse

Sorry - the complete script

# encoding=utf-8
"""
PythonScript replacement of TextFX>Edit>Rewrap
https://community.notepad-plus-plus.org/post/78262
Author: @Ekopalypse, @PeterJones
"""
from Npp import editor, notepad
from textwrap import wrap
import re

def rewrap(text, pos, eol):
    paragraphed = re.sub(eol+eol, '\0xB6', text)
    joined = re.sub(r'\s*(\r\n|\r|\n)', " ", paragraphed)

    unparagraphed = re.sub('\0xB6', eol+eol, joined)
    retlist = []
    for linetext in unparagraphed.splitlines():
        if linetext == '':
            retlist.append('')

        for partial in wrap(linetext, pos, expand_tabs=False, replace_whitespace=False, break_on_hyphens=False):
            retlist.append(partial.lstrip())


    return retlist

def main():
    _prompt = "Wrap at position:"
    _title = 'ReWrap'
    default_pos = 78
    if editor.getSelectionEmpty():
        _prompt = "ATTENTION!! - Since nothing is selected, the WHOLE document is rewrapped.\n" + _prompt
        _title = "ATTENTION!! - " + _title
    pos = notepad.prompt(_prompt, _title, default_pos)
    if pos is None:
        return  # cancelled
    else:
        pos = int(pos)

    start, end = editor.getUserLineSelection()
    start_pos = editor.positionFromLine(start)
    end_pos = editor.getLineEndPosition(end)

    eol = {0:'\r\n', 1:'\r', 2:'\n'}[editor.getEOLMode()]
    rewrapped = rewrap(editor.getRangePointer(start_pos, end_pos-start_pos), pos, eol)
    editor.setTarget(start_pos, end_pos)
    editor.beginUndoAction()
    editor.replaceTarget(eol.join(rewrapped))
    editor.endUndoAction()

main()

Reiner Bühl

@Ekopalypse Yes, this version seems to work fine! Many thanks!