HPF 80 Characters per line to LCR 76 Characters per line

[Mary Kobis]

Before I start inventing a new text editor, is anyone willing to share a tcl that will reformat a transcribed report from HPF to LCR. The new twist is that I need to keep the paragraphing. Today we have a tcl in place that will place a tilde before the overage word if the line is over 76 characters.

When displayed in Siemens LCR the overage word is on a line by its self. It may occur in a middle of a sentence. When this happens it makes the entire document look very unprofessional.

My first pass on the document would be to identify the paragraphs by a few rules we found (I know more will show up) and then a second pass to take the paragraph and realign to 76 characters. This may create a new obx or extend an existing obx.  Should be fun!… Any help is appreciated. Thanks, Mary.

Topic

[James Cobane]

Mary,

Not sure what platform you are running, but in AIX there is the “fmt” command that allows you to reformat text to a specified width per line; i.e.

echo “This is a line of text that I am specifying to be only 20 characters per line” | fmt -20

This is a line of

text that I am

specifying to be

only 20 characters

per line

You could potientially run your text through this command via an exec statement within a proc.  There may be a native tcl command to do this, but I’m not aware of it.

Hope this helps.

Jim Cobane

Henry Ford Health

Jim Cobane

[Mary Kobis]

Thanks Jim! I think it will be very useful. We are on AIX 6 and CL 5.8. Mary

[Robert Kersemakers]

Hi Mary,

I have had a similar situation: we have medical results in a Word-document (a letter from the specialist to the family doctor), that needs to be transformed into EDIFact messages, which is a flat-file format where all the text lines are limited to 72 characters. As in your case, new paragraphs and tabs did not show well. I did a lot of ‘parsing’ to get a reasonable result, but it will never be perfect.

One of the first things I made was a proc to reformat the text to a specified number of characters per line, without breaking words:

Code:

proc orbis_split_tekst { tekst {lengte 80}} {

 # This procedure checks whether (a part of) an additional LAB-text is longer than $lengte characters
 # If so, this text is split into separate parts; the list of parts is returned.

set returnlist {};

 if {[string length $tekst] > $lengte} {

   # String is te lang: opsplitsen in woorden.
   set woordlist [split $tekst ” “];

   # Tevens wordt er een lege ‘nieuwe tekst’ aangelegd
   set newtekst “”;

   # Alle woorden worden nu doorlopen
   foreach woord $woordlist {

     if { [clength $newtekst] == 0} {
       # Eerste woord: gewoon toevoegen (zonder spatie) aan ‘newtekst’
       set newtekst $woord;

     } else {
       # Indien er een x0e (andere representatie van .br) volgt: nieuwe regel
       if {[cequal $woord x0e]} then {
         set returnlist [lappend returnlist $newtekst];
         set newtekst “”;

       } else {
         # Nu wordt de ‘nieuwe’ lengte bepaald: lengte ‘woord’ + lengte ‘nieuwe tekst’ + 1 (spatie)
         set nl [expr [clength $woord] + [clength $newtekst] + 1];
         
         # Als ‘nieuwe’ lengte groter dan $lengte, dan wordt de ‘newtekst’ aan de ‘returnlist’ toegevoegd
         if {$nl > $lengte} {
           set returnlist [lappend returnlist $newtekst];
           set newtekst $woord;
         } else {
           # Anders wordt ‘woord’ aan ‘newtekst’ (met een spatie) toegevoegd
           set newtekst [string trim [cconcat $newtekst ” ” $woord]]
         };
       };
     };
   };
   # De laatste ‘newtekst’ wordt nog aan de ‘returnlist’ toegevoegd.
   lappend returnlist $newtekst;
 } else {
   # Hele tekst kan aan returnlist worden toegevoegd
   if {[string length $tekst] > 0} {
     lappend returnlist $tekst;
   }
 };

 return $returnlist;

}

Documentation in Dutch though…

Use the proc like this:

Code:

set lpostekst [orbis_split_tekst  $postekst 132]

As with most things, if I look at this proc now, I’m wondering what I’m doing at some lines. Especially the line ‘set returnlist [lappend returnlist $newtekst]’…

Hope you can do something with this.

Zuyderland Medisch Centrum; Heerlen/Sittard; The Netherlands

[Mary Kobis]

Thank you Robert and James. I have coded the first proc to identify the paragraphs in the messages (OBX5). I replaced OBX4 (not being used) with a counter and sub-counter so I can track when a paragraph starts and ends. It is not perfect but it seems as though it will get most of them. Now I can evaluate each paragraph and use logic (which we have and Robert and James have supplied) to create the finished report.

Thank you both for some great ideas! Mary

[Author]

Replies