TextConverterPattern.jsx

// InDesign script (DRAFT.)
//----------------------------------
// Apply some text conversion routine on consistent
// style ranges *while preserving special characters*
// (anchored objects, tables, footnotes...)


function myConverter(/*str*/text)
//----------------------------------
// Change a string into something else...
// Put your  ascii-to-unicode routine here
{
   return 'x' + text.split('').reverse().join('');
}

(function(  doc,pp,ptn,src,rng,a,i,tx,t,x,b,j,z)
//----------------------------------
{
   if( !(doc=app.properties.activeDocument) ) { alert("No document."); return; }

   // Special characters that MUST be skipped.
   // Footnote refs, tables, anchored objects...
   // ---
   const RE_SKIP = /[\u0004\u0016\uFFFC]/;

   // YOUR SETTINGS
   // ---
   const TARGET_FAMILY_LOWER = ("Minion Pro").toLowerCase();  // E.g  "nudi";
   const DEST_FONT_NAME = "Myriad Pro\tItalic"                // E.g  "Tunga";

   // Check dest font and set properties.
   // ---
   if( !(t=app.fonts.item(DEST_FONT_NAME)).isValid )
   { alert("The font " + DEST_FONT_NAME + " is missing."); return; }
   
   pp =
   {
      contents:        '',
      appliedFont:     t.getElements()[0],
      composer:        "$ID/HL Composer Optyca", // Adobe World-Ready Composer
      // etc
   };

   // Pattern formatter.
   // ---
   const __ = $.global.localize;

   // Array of consistently styled Text items.
   // ---
   a = doc.stories.everyItem().textStyleRanges.everyItem().getElements().slice();
   for( i=a.length ; i-- ; )
   {
      // [REM] Forcing resolve is important!
      // ---
      tx = a[i].getElements()[0];                                             // tx :: Text ; consistent text range

      // Filter based on font family.
      // ---
      t = (tx.properties.appliedFont||0).fontFamily||'';
      if( -1 == t.toLowerCase().indexOf(TARGET_FAMILY_LOWER) ) continue;

      // Get the last character index (x) and create a specifier
      // pattern for accessing character ranges from %1 to %2.
      // ---
      ptn = tx.toSpecifier().split('%').join('%%');                           // Safer (in case the spec contains `%`, change into `%%` for the pattern.)
      ptn = ptn.split('/character[');                                         // E.g `(/document[@id=4]//story[@id=222]`, `12] to /document[@id=4]//story[@id=222]`, `34])`
      if( 3 != ptn.length || isNaN(x=parseInt(ptn[2],10)) ) continue;         // Shouldn't happen.
      ptn = ptn[0]+'/character[%1] to '+ptn[0].slice(1)+'/character[%2])';    // Pattern for characterRange(%1,%2)

      // Split text contents w.r.t SPLIT_CHARS (U+0016, U+FFFC, etc)
      // in order to preserve anchored objects and so.
      // ---
      t = tx.texts[0].contents + '\x01';                                      // [HACK] Help prevent the split(RegExp) bug (last char issue.)
      b = t.split(RE_SKIP);
      ( t=b.length-1, b[t]=b[t].slice(0,-1) );                                // [HACK] Remove \x01 suffix.

      for( j=b.length ; j-- ; x-=(1+z) )                                      // b :: [src1, src2, src3...] ; tx.contents == `src1•src2•src3•...` where • is any INDD mark.
      {
         z = (src=b[j]).length;                                               // src :: original substring  ;  z :: length
         if( !z ) continue;                                                   // Empty -> simply jump over the mark.
         rng = resolve( __(ptn, ''+(x-z+1), ''+x) );                          // rng :: character range of the source (Text object.)

         pp.contents = myConverter(src)||'';                                  // Converted substring.
         rng.properties = pp;                                                 // Replace.
      }
   }
})();