[Date Index] [Thread Index] [Author Index]

Re: Adding labels to a GraphicsGrid object

• To: mathgroup at smc.vnet.net
• Subject: [mg113514] Re: Adding labels to a GraphicsGrid object
• From: Zach Bjornson <bjornson at stanford.edu>
• Date: Sun, 31 Oct 2010 05:37:07 -0500 (EST)
• References: <iaj4ma$n5a$1@smc.vnet.net>

Does using Labeled[(*your graphics object(s)*),(*label*)] (with its 
Alignment and Spacings options) not do what you need? It's a bit 
difficult to give advice on this (and your earlier thread) without 
example code. I'd be a bit surprised if you had to write your own 
intricate labeling function.

Best,
Zach

On 10/31/2010 12:08 AM, kj wrote:
> As I described in an earlier post, I am trying to add labels to a
> complex GraphicsGrid object.
>
> This object is in fact a GraphicsRow of GraphicsGrid objects.
>
> I need to add various labels on each side of the overall graphic.
> These labels must line up with various elements of the overall
> graphic.  Many of these labels are associated with closely-spaced
> objects, therefore they have to be placed within tight tolerances.
> Also, I need to generate many figures like this one, so manual
> adjustment of the label positions is out of the question.
>
> I've tried many ways to do this, but they have all failed.  My most
> successful attempt so far used heuristics to determine the positions
> of the desired labels.  To do this my code extracted the centers
> of the elements in the various grids, and performed the necessary
> scaling and translations of these points to obtain their coordinates
> in the overall graphic (which I call "global" coordinates).
>
> The computed global coordinates are off.  The problem (I think)
> rests with the computed scaling factor, for converting the coordinates
> local to each sub-figure to the global coordinates of the entire
> array.  I have not found a way to compute this factor with sufficient
> accuracy.  I can tell you what doesn't work: taking the ratio of
> the 4th parameter of the Inset objects generated by GraphicsGrid
> to the PlotRange of the inset image.  The reason for this is that
> area corresponding to this PlotRange is substantially smaller (to
> an unpredictable degree) than the area corresponding to the 4th
> parameter to Inset.
>
> If anyone can tell me how I can compute the exact scaling factor
> for converting from the local coordinates of the arrayed subfigures
> to the global coordinates of the entire composite figure I'd greatly
> appreciate it!
>
> TIA!
>
> ~kj
>