Help converting Notepad++ format to Pythonscript

[Andrew Clark]

I am trying to create a Pythonscript program to convert old CNC programs to a newer format. I am using the editor.rereplace() format to do this and have most of my functions working this way. I am having trouble with one specific format. In Notepad++ I would use Find/Replace with the following conditions to achieve what I want, line by line.

Find: ((.?))|(a-z([-#]?\d(.\d)?))\s
Replace: $1$2

This removes any spaces in the program that are NOT inside of parenthesis and that would otherwise make every other regex more difficult. I cannot seem to get that function working in Pythonscript. I have tried adding enumerable escapes ( \ ) and nothing seems to be working for me. Here are a few examples of before and after code.

BEFORE
O1234
(EXAMPLE PROGRAM)
(THIS IS A CNC COMMENT)
(I WANT THESE LINES TO RETAIN SPACES)
()
G28 G91 Z0
T1 M6 (EXAMPLE TOOL)
M3 S10000
G90 G0 G95 G54 X1.0 Y2.0
G43 Z0.25 H1 D1 M8
G1 Z-1.0 F0.006
X-1.0 Y-2.0
Y2.0
X1.0
G0 Z1.0 M5
G28 G91 Z0 M9
M01 (CHECK PART)

AFTER
O1234
(EXAMPLE PROGRAM)
(THIS IS A CNC COMMENT)
(I WANT THESE LINES TO RETAIN SPACES)
()
G28G91Z0
T1M6(EXAMPLE TOOL)
M3S10000
G90G0G95G54X1.0Y2.0
G43Z0.25H1D1M8
G1Z-1.0F0.006
X-1.0Y-2.0
Y2.0
X1.0
G0Z1.0M5
G28G91Z0M9
M01(CHECK PART)

Notice that the spaces OUTSIDE of the parenthesis have been removed. Can anyone help me with the proper formatting to make this work? Thanks!

[Claudia Frank]

@Andrew-Clark

from the given example this should do the job

lines = editor.getCharacterPointer().splitlines()
new_lines = []
for line in lines:
    if '(' in line:
        pos_round_bracket = line.find('(')
        _line = line[:pos_round_bracket].replace(' ','')
        _line += line[pos_round_bracket:]
        new_lines.append(_line)
    else:
        new_lines.append(line.replace(' ',''))
editor.beginUndoAction()
editor.setText('\r\n'.join(new_lines))
editor.endUndoAction()

Cheers
Claudia

[Scott Sumner]

@Andrew-Clark

Since you are already in Pythonscript, why don’t you forget regex and go with “brute force”, something like this, or something similar? (admittedly this does not do the “ultimate” in error-checking, but it is just supposed to be an idea to get you on an alternate track of thinking–if that’s where you want to go…):

new_text = ''
inside_parens = False
for in_ch in editor.getText():
    if not inside_parens and in_ch == ' ':
        continue
    elif in_ch == '(':
        inside_parens = True
    elif in_ch == ')':
        inside_parens = False
    new_text += in_ch
editor.setText(new_text)

[Andrew Clark]

Well, that works fantastic! I actually set out to use a more programmatical approach to this, but found that I could do almost all of the things I wanted to with simple regex. It wasn’t until I hit the need described above that I got stuck. The regex for it worked, but I couldn’t get the code to work in Pythonscript.

It will take some digging to follow what you did with that code, but it should help me with more advanced alterations in the future…when I get there. Thanks!

[Scott Sumner]

@Andrew-Clark

Not sure if you are directing that response to @Claudia-Frank or to me…but I’ll say this about the regex approach: I never understood how your regex worked live with Notepad++ – maybe the markdown syntax on this site was “stealing” some parts of it.

If I try it as I see it above I get this as the AFTER text, which is clearly not right:

OO11223344
((EEXXAAMMPPLLEE  PPRROOGGRRAAMM))
((TTHHIISS  IISS  AA  CCNNCC  CCOOMMMMEENNTT))
((II  WWAANNTT  TTHHEESSEE  LLIINNEESS  TTOO  RREETTAAIINN  SSPPAACCEESS))
(())
GG2288  GG9911  ZZ00
TT11  MM66  ((EEXXAAMMPPLLEE  TTOOOOLL))
MM33  SS1100000000
GG9900  GG00  GG9955  GG5544  XX11..00  YY22..00
GG4433  ZZ00..2255  HH11  DD11  MM88
GG11  ZZ--11..00  FF00..000066
XX--11..00  YY--22..00
YY22..00
XX11..00
GG00  ZZ11..00  MM55
GG2288  GG9911  ZZ00  MM99
MM0011  ((CCHHEECCKK  PPAARRTT))

What gives? Shouldn’t I see some \( and \) sequences in the regex?

[Andrew Clark]

@Scott-Sumner said:

@Andrew-Clark

Not sure if you are directing that response to @Claudia-Frank or to me…but I’ll say this about the regex approach: I never understood how your regex worked live with Notepad++ – maybe the markdown syntax on this site was “stealing” some parts of it.

If I try it as I see it above I get this as the AFTER text, which is clearly not right:

OO11223344
((EEXXAAMMPPLLEE  PPRROOGGRRAAMM))
((TTHHIISS  IISS  AA  CCNNCC  CCOOMMMMEENNTT))
((II  WWAANNTT  TTHHEESSEE  LLIINNEESS  TTOO  RREETTAAIINN  SSPPAACCEESS))
(())
GG2288  GG9911  ZZ00
TT11  MM66  ((EEXXAAMMPPLLEE  TTOOOOLL))
MM33  SS1100000000
GG9900  GG00  GG9955  GG5544  XX11..00  YY22..00
GG4433  ZZ00..2255  HH11  DD11  MM88
GG11  ZZ--11..00  FF00..000066
XX--11..00  YY--22..00
YY22..00
XX11..00
GG00  ZZ11..00  MM55
GG2288  GG9911  ZZ00  MM99
MM0011  ((CCHHEECCKK  PPAARRTT))

What gives? Shouldn’t I see some \( and \) sequences in the regex?

My response was directed at Claudia. And you’re right, the code I posted above does not result in the proper result in Notepad++. Something must have been lost. I could go back and find it, but at this point my question has been answered so I am going to call this topic closed. Thank you all for the responses!

[Claudia Frank]

@Andrew-Clark

sorry, didn’t understand that you want a python script solution with regular expression replacement.

regex = r'(\(.*?\))|(\w+)\s'  
editor.rereplace(regex, lambda m: '{}{}'.format(*m.groups()))

Cheers
Claudia

[Scott Sumner]

I am most-curious about why the OP (@Andrew-Clark) could get his regex replacement working in interactive N++ and not with Pythonscript, but perhaps that is a question that isn’t going to get answered…my guess would be not using raw string notation (the leading r in r'regex') in the PS…

[Claudia Frank]

or maybe it was about how to get the match groups returned as replacement … ??

Cheers
Claudia

[guy038]

Hello @andrew-clark, @scott-sumner, @claudia-frank and All,

I was away this weekend, to hike in the Cevennes mountains and my feet still remember theses rides ! As you can imagine, just sitting, comfortably, in front of your laptop and discussing Notepad++ matters is a real treat ;-))

So, Andrew, as for me, I would use, with native N++, the regex S/R, below :

SEARCH (?-s)(\(.+?\))|\x20

REPLACE ?1\1

---

Notes :

• As usual , the (?-s) modifier means that any regex dot character ( . ) will match a single standard char., only.

• Then, this regex would match :

  • First, any shortest range of chars, between parentheses, stores it as group1 and rewrites this group as it is. ( Note that the ( and ) characters have to be escaped, with a \ symbol, to be considered as literals ! )

  • Secondly, any space character, just ignored in replacement.

Cheers,

guy038

[Claudia Frank]

@guy038

Hi Guy, nice shortening but one question - it seems that ?1 is, for the posted example data, not needed in the replacement - what is the idea behind it?

Thank you and cheers
Claudia

[Claudia Frank]

Just in case someone wants to add this into a python script it would now look like this

regex = r'(?-s)(\(.+?\))|\x20'  
editor.rereplace(regex, lambda m: '{}'.format(m.groups()[0]))

Cheers
Claudia

[Scott Sumner]

@Claudia-Frank said:

or maybe it was about how to get the match groups returned as replacement … ??

I don’t know…the Pythonscript docs for editor.rereplace() seem to have an adequate example of this…but considering it now maybe it could be made better?

[Claudia Frank]

@Scott-Sumner

there is always room to improve something - so, yeah - go for it. :-D

Cheers
Claudia

[guy038]

Hi, @andrew-clark, @scott-sumner, @claudia-frank and All

Sorry, for my late answer as I was elaborating a regex, for the tricky case, below :

https://notepad-plus-plus.org/community/topic/16053/how-to-ascend-numbers/3

Ah, how silly am I ! Of course, when the regex finds a space character ( the second alternative ), the group1 is not defined. Thus, in replacement, it simply rewrites…nothing, as group1 represents a zero-length string ;-))

Clever deduction, Claudia. So a possible regex is :

SEARCH (?s)(\(.+?\))|\x20

REPLACE \1

Note : I also changed the (?-s) modifier into (?s), which allows possible multi-lines ranges of text, between parentheses !

Best Regards,

guy038

[guy038]

Hi, @andrew-clark, @scott-sumner, @claudia-frank and All

Ah :-(( I should have tested my idea of changing (?-s) into (?s), before posting. Actually, in case of multi-lines range of characters, between parentheses, the suitable regex S/R is , rather :

SEARCH (\([^()]+?\))|\x20

REPLACE \1

Hope it’s the right one, this time :-D

Cheers

guy038