Local parameters and introducing new levels

rowlesmr

Hi all

I'm doing a bunch of stuff with local parameters and macros, and am coming across a few issues with not finding local parameters once I've defined them, and I think it is because of the different macro expansions having objects in them that introduce a new level, and so the order in which I have my macros is important.

What introduces a new level (or scope)?

From techref: "The local keyword is used for defining named parameters as local to the top, xdd or phase level."

I take this as meaning I can have global local parameters (??) if they are declared at the beginning of an inp file, local prms that belong to an `xdd` or `yobs_eqn`, and local prms that belong to a `str`, `hkl_Is`, `xo_Is`, or `d_Is`.

Are there any other keywords that introduce a new level? Is there anyway of manually defining (or just ending) a level?

How I think it works:

'this is top level. everything here globally accessable.
...
xdd ...
   'local prms in here are only visible to things inside the xdd
   ...
   str
      'local prms in here are only visible to things inside the str
      ...
      ...
      ...
      ...
      'this still belongs to the str
yobs_eqn ...
   'local prms in here are only visible to things inside the yobs_eqn
   ...
   d_Is
      'local prms in here are only visible to things inside the str
      ...
      ...
      ...
      'this still belongs to the d_Is. There is no way to declare a local prm of yobs_eqn scope down here

.

Also, most of this is in v5, as that is what I have at work...

alancoelho

How you have described it is how it works.

Parameters defined using 'prm' have global scope

Parameters defined using 'local' can have scope that can be either:

- global if defined at before any xdds
- local to an xdd if defined at the xdd level
- local at the phase level if defined with a phase

Example:

   local a ... ' global scope
   xdd
      local b ... ' local to this xdd
      prm c ...   ' global scope
      str
         local d ... ' local to this str
         prm e ...   ' global scope
         prm = a; ' OK
         prm = b; ' OK
         prm = c; ' OK
         prm = d; ' OK
         prm = e; ' OK
         prm = f; ' NOT OK
      xo_Is
         local f ... ' local to this str
         prm = a; ' OK
         prm = b; ' OK
         prm = c; ' OK
         prm = d; ' NOT OK
         prm = e; ' OK
         prm = f; ' NOT OK
   xdd
      local f ... ' local to this xdd
      prm = a; ' OK
      prm = b; ' NOT OK
      prm = c; ' OK
      prm = d; ' NOT OK
      prm = e; ' OK
      prm = f; ' NOT

rowlesmr

Ta a lot Alan

I've been doing some more investigating (in TA7), and have found that my issue comes from fit_objs.

I'm using macros to add background bumps and the like, and they use fit objects to describe the shapes. I'm using local prms, as I've got multiple xdds per inpfile, and want to keep the same names for reporting them to file.

Using your example:

local a ... ' global scope
xdd
    local b ... ' local to this xdd

    fit_obj = bb SomeEquation; 

    local bb ... 'this is local to the fit_obj? I can't get to it from the str below, but I can use it in the fit_obj
   
    prm c ...   ' global scope
    str
       local d ... ' local to this str
       prm e ...   ' global scope
       prm = a; ' OK
       prm = b; ' OK
       prm = bb; 'NOT OK
       prm = c; ' OK

alancoelho

Apologies, I forgot to mention that fit_obj's also captures local parameters.

To solve your problem below then place the 'local bb...' above the fit_obj keyword.

rowlesmr

That explains quite a bit! Also, kind of explains why you put all your prms below fit_objs in topas.inc (even though they're global..., and not counting Bkg_Diffuse :))

Where do fit_objs fit in the hierarchy? I know they're below xdd, but are they above str? I'm also assuming that dummy_str has the same precedence as str.

It's just that in your example of plotting fit_objs in the techref, you wrote:

xdd ...
    fit_obj !f1 = ...
        dummy_str
            phase_name name
            scale = p;

implying that the dummy_str belongs to the fit_obj.

.

Also, is there any chance of local_with_error?

alancoelho

fit_obj is on the same level as str/phase.

dummy_str is on the same level as phase.

>It's just that in your example of plotting fit_objs in the techref, you wrote:

I couldn't find that reference in the V7 manual. I found :

fit_obj !f1 = ...
Plot_Fit_Obj(f1, “Fit Obj”)

This expand such that there is an indentation. Anyway, fit_obj and phase are in the same level.

rowlesmr

Ta a lot.

I managed to expand the macro in my head and misremember the indentation.

.

Anyhoo, this makes everything make sense, and I'll rewrite my fit_obj macros to have their prms after the fit_obj.

Matthew