Use StringRegExpReplace to replace outside comments and quotes?

[brycetech]

Hi

I've mentioned before that I implement LIS systems for hospitals.

I have been writing a syntax checker for one of the parts of this implementation. It helps write what could be considered "macros" that take complex tasks that would normally take a body to do and has the computer do it.

I've ran up on a part in the autoformatting portion of the checker where I want it to do some formatting outside of quotes or outside of comments, but not within them.

For example, I want it to add spaces in front of and behind general mathmatic functions (+,-,*,/, etc).

the current method I'm using is to scan the script and replace each quoted item or commented item with a variable, then adding the spaces for the math functions and then going back and replacing the variables above with the original quoted or commented item. While this works, I'm wondering if a straight regex can be written that would do it without all the swapping and replacing.

In the snippet script example below there are items which need the spaces (uamicmin-curinemin) but others that do not ("5-10" for example)

RULE

(

UAMIC.FINDTEST AND

CONTRIBCHANGED AND CONTRIBPRESENT AND

NOT UAMIC.ANYTESTCOM("CURINE REFLEXED BY RULE CURINE_S") AND

NOT UAMIC.ANYTESTCOM("CURINE REFLEXED BY RULE UACI_REFL_S") AND

(

NOT CURINE.FINDTEST OR

(CURINE.FINDTEST AND ABS(UAMICMIN-CURINEMIN) >= TIMELIMIT)

) AND

(

(

MATCH(SITE_CODE,"LR") AND MATCH(TESTREQ.PTCODE,"LRER") AND

(

MATCH(UAWBC.VALUE,"5-10","10-20","20-30","30-40","40-50","50-100","100-180","AB180","TNTC")

) AND

UAMIC.ANYTESTCOM("MICROSCOPIC REFLEXED FROM UA")

) OR

(

MATCH(SITE_CODE,"MR") AND MATCH(TESTREQ.PTCODE,"MRCOUR","MRSERI") AND

(

MATCH(UAWBC.VALUE,"5-10","10-20","20-30","30-40","40-50","50-100","100-180","AB180","TNTC")

) AND

UAMIC.ANYTESTCOM("MICROSCOPIC REFLEXED FROM UACI")

) OR

(

MATCH(SITE_CODE,"CH") AND

(

MATCH(UAWBC.VALUE,"5-10","10-20","20-30","30-40","40-50","50-100","100-180","AB180","TNTC") OR

(NOT UABAC.MISSING AND UABAC.DATACHANGED("VALUE") AND MATCH(UABAC.VALUE, "2+", "3+", "4+"))

) AND

UAMIC.ANYTESTCOM("MICROSCOPIC REFLEXED FROM UACI")

)

)

)

{

ENTORDER.ADDTEST("CURINE");

UAMIC.TESTCOMMENT("N","CURINE REFLEXED BY RULE CURINE_S");

REFTEST.TESTCOMMENT("N","TEST REFLEXED FROM ACCN: " & UAMIC.ACCESSION);

}

What I'm using now is:

;~strip // comments and quoted items

local $arr = StringRegExp($_cont, "(?://.+rn)|((?:x22[^x22]*x22)+)", 3)

;~ unique items

local $newarr = _ArrayUnique($arr)

;~ _ArrayDisplay($newarr)

;~ replace with variables

for $tX = 1 to $newarr[0]

$_cont = stringreplace($_cont,$newarr[$tX],"#newarr" & $tX & "#")

next

;~ add spaces

$_cont = StringRegExpReplace($_cont, "()|(|d|w|x22)(<|>|=|+|-|*)", "$1" & " " & "$2")

;~ put it back together

for $tX = 1 to $newarr[0]

$_cont = stringreplace($_cont,"#newarr" & $tX & "#", $newarr[$tX])

next

can all of that be done with a single regex?

tia

BT

Edited March 23, 2012 by brycetech

[Xenobiologist]

Can you provide text before, text after "the regex" so that I can test it on my own.

[jchd]

I don't have a clear idea about the grammar behind, but something I'd look at is the fact that >= will now get a space between > and =, possibly changing the meaning of the test by making them now two separate tokens for the interpretor. That could be of importance in this context.

As to do all this in one shot, I'm not sure. Mainly because the use of double quote in the "formatting" regep is unclear to me.

[brycetech]

Hi

thanks for your replies

I'll attach an example of a full rul, before formatting and after. This is an extreme case of what could occur, sometimes there are a correct spaces where they should be.

I can get the tabs to work, so feel free to ignore them in the "after" file.

Notice that only the items outside quotes or comments will be changed to have the addional spaces around the mathmatic operators.

If I could just get a guide or some pointers on how to start making this work, I think I could go from there.

and yes, the comment above about >= or <= or <> or == is valid, there should not be spaces between them if they appear together in the order indicated.

=> or =< is not a valid operator in this language nor is +- or some other that other languages permit

I'm already scanning for these before this part cause they can crash the server

While my program adds the spaces, I just think there 's gotta be a better way than all the swapping I'm doing.

any help is appreciated.

tia

BT

before.txt

after.txt