Regular expression size limited 2048 chars How to make it longer?

hcchen

@Alan-Kilborn thank you very much. We plan to develop a plugin instead. I wish someone would like to be payed to develop that for us. User use Notepad++ Style Token to draw colors onto keywords and the plugin do the search according to the colored keywords.

e.g.

‘windows 10’, ‘win10’, and ‘window10’ are colored with blue
‘blue screen’ and ‘BSOD’ are colored with red
‘blue tooth’ and ‘bt’ are colored with green

Then the plugin search the document with the logic shown below and get all lines that match the logic to an output window or jump to the next line that match the logic.

 ('windows 10' OR 'win10' OR 'window10') AND ('blue screen' OR 'BSOD') AND ('blue tooth' OR 'bt')

Would you like do this? Or introduce us one who can do it?

陳厚成 H.C. Chen
緯創資通 Wistron Corporation
新北市22181汐止區新台五路一段88號21樓
21F., No.88, Sec.1, Xintai 5th Rd., Xizhi Dist.,
New Taipei City 22181, Taiwan
eMail: h.c._chen@wistron.com
Phone: +886-2-6612-27887, Cell: +886-922-417-555
Line: hcchen5600, WeChat: hcchen5600

Alan Kilborn

@hcchen

This would not be terribly difficult to accomplish; if I were doing it I would do it with PythonScript, not a dedicated plugin.

Your boolean logic could be accomplished with help from THIS information.

The hardest part would be “presentation of results” since writing data to the Search results area is not really practical, either with a script or a plugin.

Alan Kilborn

OP jumping this to new thread for some reason: https://community.notepad-plus-plus.org/topic/22012

hcchen

@Alan-Kilborn thank you so much! PythonScript looks good, I wish it’s good enough …

hcchen

@Alan-Kilborn I am so glad to have spend a day today on the Notepad++ PythonScript, it’s fun. But still too much to learn in prior being able to make something decent. We need to know:

how to get ‘Style token’ list,
use the logic to find matched lines.
If output them to a view is not easy as you mentioned then to a new buffer is also okay.

So users draw colors to keywords, run the python script and get a ‘new 1’ file with all matched lines.

I am afraid you probably very busy? We can accept only a draft sample code so that we can proceed to finish it.

Best regards,
H.C. Chen
Wistron Corp.

Alan Kilborn

@hcchen said in Regular expression size limited 2048 chars How to make it longer?:

how to get ‘Style token’ list

Here’s a demo script for that:

# -*- coding: utf-8 -*-
from __future__ import print_function

from Npp import *

#-----------------------------------------------------------------------------

def highlight_indicator_range_tups_generator(indicator_number):
    '''
    the following logic depends upon behavior that isn't exactly documented;
    it was noticed that calling editor.indicatorEnd() will yield the "edge"
    (either leading or trailing) of the specified indicator greater than the position
    specified by the caller
    this is definitely different than the scintilla documentation:
    "Find the start or end of a range with one value from a position within the range"
    '''
    if editor.indicatorEnd(indicator_number, 0) == 0:
        return
    indicator_end_pos = 0  # set special value to key a check the first time thru the while loop
    while True:
        if indicator_end_pos == 0 and editor.indicatorValueAt(indicator_number, 0) == 1:
            # we have an indicator starting at position 0!
            # when an indicator highlight starts at position 0, editor.indicatorEnd()
            #  gives us the END of the marking rather than the beginning;
            #  have to compensate for that:
            indicator_start_pos = 0
        else:
            indicator_start_pos = editor.indicatorEnd(indicator_number, indicator_end_pos)
        indicator_end_pos = editor.indicatorEnd(indicator_number, indicator_start_pos)
        if indicator_start_pos == indicator_end_pos: break  # no more matches
        yield (indicator_start_pos, indicator_end_pos)

#-----------------------------------------------------------------------------

def INDICATORS_TEST__main():

    if 1:  # show currently ACTIVE indicators

        at_least_one = False

        num_to_english_dict = {
            25 : 'cyan;style1',
            24 : 'orange;style2',
            23 : 'yellow;style3',
            22 : 'purple;style4',
            21 : 'dark-green;style5',
            #31 : 'red;mark',
            }

        for i in list(num_to_english_dict.keys()):

            english = 'indic#{0:02}'.format(i)
            try:
                english += '[{}]'.format(num_to_english_dict[i])
            except KeyError:
                pass

            for tup in highlight_indicator_range_tups_generator(i):
                (start, end) = tup
                print('pos{s}-{e}      {eng}      {text}'.format(
                    s=start,
                    e=end,
                    eng=english,
                    text=editor.getTextRange(start, end),
                    )
                )
                at_least_one = True

        if not at_least_one:
            print("no indicators active")

#-----------------------------------------------------------------------------

INDICATORS_TEST__main()

Style some text, then run the script and observe something like this in the PythonScript console window:

pos25-29      indic#24[orange;style2]      from
pos127-131      indic#24[orange;style2]      from
pos696-700      indic#24[orange;style2]      from
pos1855-1874      indic#21[dark-green;style5]      num_to_english_dict
pos2127-2146      indic#21[dark-green;style5]      num_to_english_dict
pos2266-2285      indic#21[dark-green;style5]      num_to_english_dict

Alan Kilborn

@hcchen said in Regular expression size limited 2048 chars How to make it longer?:

use the logic to find matched lines.

Well, you could follow the advice in the thread I pointed you to before, when I said:

Your boolean logic could be accomplished with help from THIS information.

But since you said earlier that you want to do ANDs and ORs, you might benefit from thinking about the following search regular expression:

(?-s)^(?=.*?and1)(?=.*?and2)(?=.*?and3)(?=.*?(?:or1|or2|or3)).+

which will match lines containing all of the AND expressions and one (or more) of the OR expressions.

Using that base expression you can alter it as needed to fit the scenario you need. For example, if at the moment you don’t need AND conditions, you just remove those parts from the regular expression.

hcchen

@Alan-Kilborn First thing I do this morning is to have confirmed this sample code works very well! Wow, it’s very useful. I wish that I can send your kid a Disney land ticket to indicate how much happiness your helps have brought to us.

hcchen

@Alan-Kilborn is it possible to call the ‘find’ dialog box? So we can easily get all matched lines to the ‘Search results’ window.

Alan Kilborn

@hcchen said in Regular expression size limited 2048 chars How to make it longer?:

I wish that I can send your kid a Disney land ticket

My kid wishes that too. :-)
It wasn’t much effort, though, as I already had the sample code I posted, and the logical search regex expressions have been discussed before.

is it possible to call the ‘find’ dialog box?

It is possible but it is not easy.

So we can easily get all matched lines to the ‘Search results’ window.

Did you end up not liking the idea of writing your results to a “new” tab?

hcchen

@Alan-Kilborn Thank you very much for the valuable answers. After all, I pop up a message box that informs user that the RegEx will be copy to clipboard, OK or Cancel? that’s all. Users can easily use the ‘Find’ dialog then. I am so glad to have finished this magic so well.

Alan Kilborn

@hcchen said in Regular expression size limited 2048 chars How to make it longer?:

After all, I pop up a message box that informs user that the RegEx will be copy to clipboard, OK or Cancel? that’s all. Users can easily use the ‘Find’ dialog then

That’s a nice approach to solving your problem.
Sometimes it just takes some discussion and background thinking to get to the right conclusion.