Zugg Software :: View topic - REGEX pattern quandry (can be a zScript trigger if that helps)

Apprentice Joined: 09 Apr 2008 Posts: 168

I'm still learning the in's/out's of REGEX and I honestly don't know if this is possible...

I play on 3kingdoms and have attempted to utilize the G-Script MIP from the library to emulate the Portal GT hpbar/guages in cmud.
for the most part this has been a success... until I get to the gLine1/gLine2 parts.

Sample : <gND:> <yYes><g[><y98%><g]> <g[><yTRS><g|><yAA><g]> <gSI: [><y41.1761%><g]>

The problem that I face is the section "<yTRS><g|><yAA>" as this can be completely empty or have several more, such as... "<yTRS><g|><yAA><yIPP><g|><yNMS><g|><yGE>"

What I want to know is, is it possible to REGEX none to single, multiple "<yXX>'s" sepparated with "<g|>'s" ?

SubAdmin Joined: 18 Nov 2001 Posts: 5182

If I am understanding your question correctly, then I must say that REGEX can not always do this. It is not possible to have an arbitrary number of captures with a given REGEX.

That said you can have a pattern such as:
^<g\a+:> <y\a+>( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?( <g[>.*?<g]>)?<gSI:[><y[^>]*><g]>
You have to make sure you repeat the optional section enough time to cover any and all eventualities. Your script will also have to handle an equal number of situations.

The question mark after each capture makes that entire section optional, and that makes the '.*' inside the captures need to be reduced from greedy to minimal '.*?'. I couldn't see using a narrower interior wildcard after examining your example which contains letters, numbers and the percent sign.

You will have to use script to parse the captured data. You might be able to simplify this a little by using a structure more like '(?: <g[>(.*?)<g]>)?'. That makes the captured portion a little smaller and uses the grouping (?:) to indicate the optional section. You can refine it further to limit what you have to parse in script.

SubAdmin Joined: 18 Nov 2001 Posts: 5182

Alternately you can do the capture this way:
^<g\a+:> <y\a+>((?: <g[>.*?<g]>)*)<gSI:[><y[^>]*><g]>

The result would be a single capture that you can then pass to %subregex to break into to separate entries as needed. Subregex is capable of breaking all the data down and formatting it into whatever structure you would like.

Apprentice Joined: 09 Apr 2008 Posts: 168

I suppose a few sample captures of the actual mud output might help explain things a touch more.

No multiple captures -

SubAdmin Joined: 18 Nov 2001 Posts: 5182

That didn't actually clarify anything since it mostly contradicts what you had before, anyhow here is a regex for that:
ND: (?:Yes|No) [\d+%](?: [([a-zA-Z|]++)])? SI: [[0-9.]+%]

The captured portion would be "", "IFF", and "IFF|GE|NMS|TRS|AA" respectively for your examples.

Apprentice Joined: 09 Apr 2008 Posts: 168

I see what you mean bu contradicts, I tried to simplify some of it from actual mud output as its supposed to be seen once the MIP code and <g <y > are removed.
The un parsed/gag'd/what have you, still looks like the original pattern. The other 3 samples are how it 'should' look. Sorry for the misunderstanding

SubAdmin Joined: 18 Nov 2001 Posts: 5182

I reread your first post and noticed that I had mistaken some of the symbols. Let's try this.

Pattern: ^<g([^>]+):> <y([^>]+)><g[><y([^>]+)><g]>(?: <g[>((?:<y[^>]+>(?:<g\|>)?)++)<g]>)? <gSI: [><y([^>]+)><g]>
Value: #var CurGline1 %concat( %1, ": ", %2, " (",%3, ") [", %subregex(%4, "(?:<(?:y|g)|>)", ""), "] S.I.N.C. : ",%5)

Apprentice Joined: 09 Apr 2008 Posts: 168

There is an unmatched parenthesis (35) Let me see if i can remove a few bits of info that really doesn't need to be captured.

Actual mud output:

Apprentice Joined: 09 Apr 2008 Posts: 168

Update...

New Pattern:

SubAdmin Joined: 18 Nov 2001 Posts: 5182

Glad to help and glad you found the parenthesis mismatch.

You might want to avoid using .* in your pattern because it can lead to a very slow processing and may even cause %3 to be completely lost. I once had a complex pattern that caused a 3 minute lockup of CMud when it was matching specific data.
Use either the [^>]+ construct to match anything that is not > or use .*? to make the regex try to match the fewest characters possible for that wildcard.

Apprentice Joined: 09 Apr 2008 Posts: 168

ah that's very good news to know, was wondering why my client suddenly started to lag about a half second or so constantly.

again thank you