How to get formatted text into arrays

Question

Dear experts and helpers,
For my project I import an RTF file and then read the data from it into 3 arrays. This works fine when just using the string contents of the paragraphs. However, the final script should be able to read and replace formatted text...
Why use the intermediate arrays? Because otherwise I need to switch back and forth between two fm-documents (and one may be a book component).
The imported file starts with a number of lines separated into two items by a TAB (» denotes a TAB, in FM \x08)
[[Garneau, 1990 #12]]    »   [9]
The right item may also be locally formatted text, e.g. [9]
Then follow the same (or smaller) number of paragraphs with formatted text like this:
[9] » D. Garneau, Ed., National Language Support Reference Manual (National language Information Design Guide. Toronto, CDN: IBM National Language Technical Centre, 1990.
Is it possible to replace in the body of the function below the following piece
  while(pgf.ObjectValid()) {
    pgfText = GetText (pgf, newDoc);
    gaBibliography.push(pgfText);
    pgf = pgf.NextPgfInFlow;
with this
  while(pgf.ObjectValid()) {
    gaBibliography.push(pgf);
    pgf = pgf.NextPgfInFlow;
Do I need a special declaration of the array gaBibliography ?
And how to get the right part of the intro lines as formatted thingy into array gaFmtCitsFmt ?
Currently I read into arrays only the 'strings' (function GetText not shown):
var gaFmtCitsRaw  = [];                           // left column in processed RTF
var gaFmtCitsFmt  = [];                           // right column in processed RTF
var gaBibliography= [];                           // bibliography lines from processed RTF
// filename is something like E:\_DDDprojects\FM+EN-escript\FM-testfiles\BibFM-collected-IEEE.rtf
function ReadFileRTF (fileName) {
  var nCits=0, nBib = 0, openParams, openReturnParams, newDoc, pgf, pgfText ;
  var TAB = String.fromCharCode(8);               // FM has wrong ASCI for TAB
  var parts = [];
  openParams = GetOpenDefaultParams();
  openReturnParams =  new PropVals(); 
  newDoc = Open (fileName, openParams, openReturnParams); 
  pgf = newDoc.MainFlowInDoc.FirstTextFrameInFlow.FirstPgf;  // get first pgf in flow
// --- read the temp/formatted citations 
  while(pgf.ObjectValid()) {
    pgfText = GetText (pgf, newDoc);
    if (pgfText.substring (0,2) == "[[") {        // citation lines start with [[
      parts = pgfText.split(TAB);                 // get the two parts of the line
      gaFmtCitsRaw.push (parts[0]);               // Push the result onto the global array
      gaFmtCitsFmt.push (parts[1]);
      pgf = pgf.NextPgfInFlow;
    } else { break }
// --- read the bibliography
  while(pgf.ObjectValid()) {                      // until end of doc
    pgfText = GetText (pgf, newDoc);
    gaBibliography.push(pgfText);
    pgf = pgf.NextPgfInFlow;
  newDoc.Close (Constants.FF_CLOSE_MODIFIED);
} // --- end ReadFileRTF
The next questions then will be how to modify Ian Proudfoot's FindAndReplace script to handle formatted text as replacement. IMHO i will need to use copy/paste ...

Answer 1

Klaus,
Working with text is about the most complicated thing to do within FrameMaker. It seems counter-intuitive, since it is about the easiest thing to do with the GUI. But alas, once you remove the ability to select with a mouse and type with a keyboard, text becomes a wild jungle of complexity.
Text ranges are not too bad, once you get the general idea. It is just that... a range of text, like something you would select with a mouse. Like a mouse selection, it starts before some character in some paragraph and ends after some character in some paragraph. It may be the same paragraph, which is a selection within a paragraph. The character can even be the same, which is then just an insertion point (cursor) somewhere.
So, a text range is a data structure that defines two paragraphs and two characters. In the jargon of scripting, the character is called an "offset." An offset is simply the number of characters past the beginning of said paragraph, where 0 is the beginning.
For example, if you want to capture the first five characters of a paragraph as a text range, you can do this, where 'pgf' is some paragraph object:
var textRange = new TextRange();
textRange.beg.obj = pgf;
textRange.beg.offset = 0;
textRange.end.obj = pgf;
textRange.end.offset = 5;
If you want to capture a whole paragraph, change that last line to the number of characters in the pgf, or you can do this:
textRange.end.offset = Constants.FV_OBJ_END_OFFSET;
...where that constant is just some built-in thing that means "get me to the end of whatever." It's a convenience of the interface.
I'll also note that a text range is actually just an array of two text location structures, one named 'beg' and one named 'end.' If you think of a text location as defined by paragraph and an offset from the first character, maybe that will make more sense.
Text item structures are a whole new mess of complexity. I can't possibly go into an explanation of them here.
I think that many ES developers (definitely myself included) still use the FDK documentation because it is considerably more comprehensive. The two interfaces are largely parallel, but of course somewhat different in the language syntax. Consider that as a potential resource.
Russ