Reverse a regex? - regex

Question

I am using AHK to automatically do something but it involves parsing XML. I am aware that it is a bad habit to parse XML with regex, however I pretty much have my regex working. The issue is AHK only has regexreplace as a method and I need something along the lines of regexkeep.
So what happens is the part I want to keep gets deleted and the part I want deleted gets kept.
Here is the code:
RegExReplace(response, "(?<=.dt.\n:)(.*)(?=\n..dt.)")
Is there a way to have everything but the match match? If not is there a better way to go about this?
Edit:
I have no attempted using the inverse regex and regexmatch but neither work in AHK. Both regexs work properly at regex101.com however neither work in AHK. The regexmath returns 0 (meaning it found nothing) and the inverse regex returns nothing as well.
Here is a link to what is being searched by the regex:http://www.dictionaryapi.com/api/v1/references/collegiate/xml/Endoderm?key=17594df4-ff21-4045-88d9-a537fd4bcd61
Here is the entire code:
;responses := RegExReplace(response, "([\w\W])(?<=.dt.\n:)(.*)(?=\n..dt.)([\w\W])")
responses := RegExMatch(response, "(?<=.dt.\n:)(.*)(?=\n..dt.)")
MsgBox %responses%

Answer 1

Here is the "reversed" regex:
s).*dt.\n:|\n..dt.*
The parts in the look-arounds need matching with the * quantifier to match from the start and up to the end. To match the newline with a dot, use singleline mode.
Debuggex Demo (where endings are \r\n)
However, there is a better option with RegExMatch OutputVar:
If any capturing subpatterns are present inside NeedleRegEx, their
matches are stored in a pseudo-array whose base name is OutputVar.
Use
RegExMatch(response, "(?<=.dt.\n:)(?<Val>.*)(?=\n..dt.)")
Then, just refer to this value as MatchVal.

Answer 2

Here's a solution that should work, assuming you want to get whatever's between the <dt> tags. Make sure you're using the latest version of AHK if possible.
xml =
(
<entry_list version="1.0">
<entry id="endoderm">
<ew>endoderm</ew>
<subj>EM#AN</subj>
<hw>en*do*derm</hw>
<sound>
<wav>endode01.wav</wav>
<wpr>!en-du-+durm</wpr>
</sound>
<pr>ˈen-də-ˌdərm</pr>
<fl>noun</fl>
<et>French
<it>endoderme,</it>from
<it>end-</it>+ Greek
<it>derma</it>skin
<ma>derm-</ma>
</et>
<def>
<date>1861</date>
<dt>:the innermost of the three primary germ layers of an embryo that is
the source of the epithelium of the digestive tract and
its derivatives and of the lower respiratory tract</dt>
<sd>also</sd>
<dt>:a tissue derived from this layer</dt>
</def>
<uro>
<ure>en*do*der*mal</ure>
<sound>
<wav>endode02.wav</wav>
<wpr>+en-du-!dur-mul</wpr>
</sound>
<pr>ˌen-də-ˈdər-məl</pr>
<fl>adjective</fl>
</uro>
</entry>
</entry_list>
)
; Remove linebreaks and indentation whitespace
xml := RegExReplace(xml, "\n|\s{2,}|\t", "")
matchArray := []
matchPos := 1
; Keep looping until we're out of matches
while ( matchPos := RegExMatch(xml, "<dt>:([^<]*)", matchVar, matchPos + StrLen(matchVar1)) )
{
; Add matches to array
matchArray.insert(matchVar1)
}
; Show what's in the array
for each, value in matchArray {
; Index = Each, Output = Value
msgBox, Ittr: %each%, Value: %value%
}
Esc::ExitApp
You really shouldn't use RegEx for parsing XML though, it's very simple to read XML in AHK using COM, I know it's outside the scope of your question, but here's a simple example using a COM object to read the same data:
xmlData =
(LTrim
<?xml version="1.0" encoding="utf-8" ?>
<entry_list version="1.0">
<entry id="endoderm"><ew>endoderm</ew><subj>EM#AN</subj><hw>en*do*derm</hw><sound><wav>endode01.wav</wav><wpr>!en-du-+durm</wpr></sound><pr>ˈen-də-ˌdərm</pr><fl>noun</fl><et>French <it>endoderme,</it> from <it>end-</it> + Greek <it>derma</it> skin <ma>derm-</ma></et><def><date>1861</date><dt>:the innermost of the three primary germ layers of an embryo that is the source of the epithelium of the digestive tract and its derivatives and of the lower respiratory tract</dt> <sd>also</sd> <dt>:a tissue derived from this layer</dt></def><uro><ure>en*do*der*mal</ure><sound><wav>endode02.wav</wav><wpr>+en-du-!dur-mul</wpr></sound> <pr>ˌen-də-ˈdər-məl</pr> <fl>adjective</fl></uro></entry>
</entry_list>
)
xmlObj := ComObjCreate("MSXML2.DOMDocument.6.0")
xmlObj.loadXML(xmlData)
nodes := xmlObj.selectSingleNode("/entry_list/entry/def").childNodes
for node in nodes {
if (node.nodeName == "dt")
msgBox % node.text
}
Esc::ExitApp
For more information on how to use this, see this post: http://www.autohotkey.com/board/topic/56987-com-object-reference-autohotkey-v11/?p=367838

Answer 3

If the given phrase only occurs once, you can probably just fetch everything around it, can't you?
RegExReplace(response, "([\w\W]*)(?<=.dt.\n:)(.*)(?=\n..dt.)([\w\W]*)", "$1$5")
looks like the easiest solution to me, but surely not the prettiest
update: in your question update, you quoted responses := RegExReplace(response, "([\w\W])(?<=.dt.\n:)(.*)(?=\n..dt.)([\w\W])"), but it should be responses := RegExReplace(response, "([\w\W]*)(?<=.dt.\n:)(.*)(?=\n..dt.)([\w\W]*)", "$1$5") - meaming keep the first ($1) and the last ($5) key of braces, which include an arbitrary amount of any characters ([\w\W]*) around your initial phrase. seems you copied it wrong. I can't say that it will work for sure tho since I don't have any code to test it on
edit - one thing I don't understand - how does regexMatch help here? it just tells us IF and WHERE there is a substring present, but surely doesn't replace anything?