| Line-Mouse widget tool [message #43237] |
Thu, 24 March 2005 07:30  |
![Howie[1] is currently offline](theme/default/images/xoffline.png.pagespeed.ic.XRkd1fkXye.png "Howie[1] is currently offline")
Howie[1]
Messages: 3
Registered: May 2004
|
Junior Member |
|
|
I would like to build or use a simple pre-developed tool which
wouuld draw a plot with data points in a certain set of x-y
(possibly completely different) axes with specific units set by
the user and is able to interactively draw a "moveable" lines
on that plot with mouse clicks and notify the user of the positions
in the data-space of the endpoints of the lines. This widget tool
should then allow the user to to slightly alter the positions of
the endpoints of the drawn lines and to tell the user specifically
(to within a a user specified number of decimal places) where in
the data space the endpoints are currently located, in real time.
I would like this tool to be able to work with two separate lines,
but I guess one would be enough if only that was possible.
It would basically allow me to more easliy determine positions
in a data space which I have been having to do interactively
which takes a long time. Any ideas anybody?
cheers, Howie
|
|
|
|
|
|
| Re: Line-Mouse widget tool [message #43252 is a reply to message #43237] |
Tue, 29 March 2005 17:27  |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Robert Barnett writes:
> That was very insightful. It seems that I totally forgot about the
> Convert_Coord function. Given that Convert_Coord relies on "system
> variables to compute coordinate conversion factors" I can see the
> significance of needing to call
> self -> ApplyCoordinates ; Set up data coordinate space.
Indeed, Convert_Coord *does* rely on system variables, but
not as many as you would think. In fact, setting ![XY].Window
and ![XY].S is really all that is needed for many (most?)
coordinate systems. With these two (or four, depending on
how you count) pieces of information, Convert_Coord is
happy to do its thing. :-)
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
| Re: Line-Mouse widget tool [message #43253 is a reply to message #43237] |
Tue, 29 March 2005 16:05 |
Robert Barnett
Messages: 70
Registered: May 2004
|
Member |
|
|
Thanks David,
That was very insightful. It seems that I totally forgot about the
Convert_Coord function. Given that Convert_Coord relies on "system
variables to compute coordinate conversion factors" I can see the
significance of needing to call
self -> ApplyCoordinates ; Set up data coordinate space.
It's so nice to see code that makes sense. It makes IDL such a
delightfully pleasant experience.
--
nrb@
Robbie Barnett
imag
Research Assistant
wsahs
Nuclear Medicine & Ultrasound
nsw
Westmead Hospital
gov
Sydney Australia
au
+61 2 9845 7223
|
|
|
|
| Re: Line-Mouse widget tool [message #43254 is a reply to message #43237] |
Tue, 29 March 2005 15:29 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
David Fanning writes:
> The Pixel_To_Value method looks something like this (stripped
> to the essentials):
>
> PRO myObject::Pixel_To_Value, x, y, INSIDE=inside
Well, of course, that would be a FUNCTION, not a procedure. :-)
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
| Re: Line-Mouse widget tool [message #43255 is a reply to message #43237] |
Tue, 29 March 2005 15:25 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Robert Barnett writes:
> I render a direct graphics plot onto a widget_draw.
> How do you transform a mouse click event which is specified in pixels to
> data-space when using direct graphics? I can compute it from the
> xmargin, ymargin, xtick_get, ytick_get, xsize and ysize. But, surely
> there must be an easier way?
Well, remember that every object I create and every-
thing I draw into a window has a coordinate system
associated with it. This is, effectively, the *data*
coordinate space. So, a medical image, for example,
might have a coordinate system in mm or cm, something
like that. And typically, these objects (images are
good examples, as are plots) know something about
where they are located in the display window.
So, these objects have a Pixel_To_Value method that
simply takes a window pixel X and Y value and can
return the proper "data" coordinate. (And also tell
me if I am inside or outside the object.)
So a call looks like this:
data = theObject -> Pixel_To_Value(event.x, event.y, Inside=in)
IF in THEN BEGIN
Print, 'Data Coordinates are: ', data
ENDIF ELSE Print, 'Outside of object'
The Pixel_To_Value method looks something like this (stripped
to the essentials):
PRO myObject::Pixel_To_Value, x, y, INSIDE=inside
inside = 0
self -> ApplyCoordinates ; Set up data coordinate space.
c = Convert_Coord(x, y, /Device, /To_Data)
retval = retValue = [c[0,0], c[1,0]]
IF (retValue[0] GE !X.CRange[0]) AND $
(retValue[0] LE !X.Crange[1]) AND $
(retValue[1] GE !Y.CRange[0]) AND $
(retValue[1] LE !Y.Crange[1]) THEN inside = 1
RETURN, retval
END
The "inside" test varies a little bit, depending upon the object.
What I've shown you is for some kind of plot object. If this were
an image, I would be checking to see if the point is inside the
image's "location" in the window.
The notion of a coordinate system associated with every object
is amazingly powerful. Today I built a "ruler" object for an
image, which is just a small scale that you want to display
on an image to give you some idea of the size of features
you are looking at. The ruler has it's own coordinate
system, so I can position it in a window, and it has a holder
for the parent coordinate system. It uses the parent coordinate
system to see how big it needs to be, and its own coordinate
system to display itself.
So I just popped it into an image object and it grows or retracts
in size as I zoom in and out of the image, all the while staying
in the same place in the window. Amazing!! Magic!
Let's just say there are days when you've *got* to love IDL!
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
