forums.silverfrost.com

[Martin_K]

here is the 15s video - how quickly I moved with slider with no crash!
And I repeated it many times.

Thanks Paul again!

https://www.dropbox.com/s/jl0q284watx3m4f/slider3.mp4?dl=0

[Martin_K]

Ken,

Thanks for your - as always - very useful tips/observations.
I will have a look at this!

A question:

Somehow, I was unable to understand how could I write a short code
for selecting an object in a graphic area by mouse click on it, which could
be used universally (for any graphic) and that it would be optically visible
in the graphic area, which object (point or line or area) was selected
and then - say - deselecting it by additional mouse click (left or right).

Is there a free code which would do it?

[Kenneth_Smith]

Martin,

See the following: https://www.silverfrost.com/ftn95-help/clearwinp/gdialog/graphicsicons.aspx

Using the basic ideas described there I was able to develop a graphics interface which allowed me to draw interactively an electrical network and extract its parameters and connectivity to a file for input to another program.

[Martin_K]

I had a look there some time ago, but was unsure whether it can be applied
also with %PL, since there is explicitly stated the %GR only.

So, can I use these features also with %PL or there is necessary to rewrite the code with %PL to %GR?

Now - an additional remark for Paul regarding my previous
suggestion for enhancing %PL capabilities.:

I still mulled how to simplify my second suggestion for the %PL enhancement (georeferencing of images in the graph) mentioned
earlier in this string of conversations and have the following
- I think, essentially simplified - idea:

Would it be realizable (with no huge effort and amount of time) to add a new option for the %PL like this:
%PL [image=filename.jpg]?

The purpose of this option would be to overlay current vector graphics
in the graph (like red lines and blue points in my case) with a pure jpg
image/picture (like in the case of %IM), with NO georeferencing,
hence no association with real geodetic coordinates which have the points/lines/areas in the graph.
I suppose, it would mean no complicated mathematics behind this.

The only requirement for such option would be that it
would have to stretch the image between X,Y axes in the graph.

Namely, I would be able to prepare a georeferenced orthophotomosaic
images in another software used for mobile mapping, which would at least roughly fit the area where
objects with real geodetic coordinates are located in the graph.
It would mean - at least - a roughly fit/coincide between the
image with no coordinate association a and objects in the graph with real geodetic coordinates.

Such option would just serve as a background source for the vector graphics in the graph. I think this would be very interesting enhancement.

Thanks for your evaluation of this idea and answer.

[Kenneth_Smith]

Martin, take a look at this example.

https://www.dropbox.com/sh/w46k0wdctr3rkc9/AADwV6E2XI7zgizhyH8tVfooa?dl=0

[PaulLaidler]

Very long standing.

[Martin_K]

I had a look at your code and this fulfills the basic idea - thanks, Ken!

But as always - it is rarely the case, when something works
with me on the first time and this is no exception.

So, I have a problem and a question.

PROBLEM:

I have the following plot function(using %PL) calling a callback (PL_CB) called by %PL (an extract of my code):

[Martin_K]

Additional info to my previous post:

I already partially found why the raster JPG image automatically disappears when
I moved with mouse over the graph.

In the callback PL_CB called within %PL I originally used the code:

[Martin_K]

And still: the use of 8913094 instead of 13369376 is no win-win situation:
when I use it (8913094) and I move over the graph with mouse the yellow rectangle, where the relevant data from surveyed points are extracted displayed is always present. It means when I point over a point, then on next and so on,
every yellow box remains permanently visible in the graph.

[Kenneth_Smith]

Martin,

Any call to simpleplot_redraw@ will remove the imported image as the %pl region is reformed. At this point you have to use COPY_GRAPHICS_REGION@ to repaste the image into the %pl graphics, before you call COPY_GRAPHICS_REGION@ to create a "back up" copy of the now modified %pl display which is used to let you display your moving yellow box.

To avoid continually calling IMPORT_IMAGE@, it would make sense to create a new permanent internal graphics region into which you import the image just once (permanent for the duration of the winio@ sequence that is associated with the %pl). Also, remember that IMPORT_IMAGE@ imports to the currently selected graphics region.

COPY_GRAPHICS_REGION@ will automatically rescale if the destination and source x/y ranges are different and since you can specify an x/y offset as well as width/height for both destination and source; this gives you a means of pasting a smaller region of the image when you have for example zoomed into an area of the graph - you know the %pl xmin/xmax/ymin/ymax values and you know the pixel width/height dimensions of the image as well as the physical origin and x/y extent coordinates the image refers to.

If your yellow box is permanently displayed you are creating the back up copy at the wrong point in the code. You just need to ensure that the copy operations are in the correct sequence.

[Martin_K]

Thanks Ken for your important remarks!

However, at this moment I have something wrong somewhere in the code
of one (or more?) of the three callbacks dealing with the graphics.

Their full code can be seen/downloaded here:
https://www.dropbox.com/s/ben76ov4n02ivs0/PL%2BCB.f95?dl=0

What I am getting now is this:



A short 10s video can be seen here:
https://www.dropbox.com/s/k12axnjebx21bzf/jpg_problem.mp4?dl=0

So, the facts are as follows:

1. I need the JPG to be displayed (stretched) ONLY between X,Y axes, not over the whole graph.
I have differences between source and destination widths and heights.

I tried it to adjust tentatively in a few iteration by specifying the offset
of the image - but I achieve practically nothing, moreover - I do not want to do it manually (to find the values of offset by manually and iteratively trying by typing in their estimated values, since this will always be different from one jpg to the another one).
So, in this particular case I would need to shorten (to squeeze) the image
both in north-south and east-west direction to get it between X,Y axes and I was unable to find how
(I would also need to specify an offset from right bottom corner).

2. As can be seen in the short video, when I load in the raster onto the graph and when I move with mouse (placed outside graph) back over the
graph, the image automatically disappears although I used the COPY_GRAPHICS_REGION@ from the buffer on the graph (the ID of the graph is variable handle_pl_OK and equals 0) also in the condition:

IF (cb_reason .eq. 'MOUSE_MOVE' ) THEN

which is in the PL_CB function called within %PL.

[Kenneth_Smith]

Looking only at your graphic above, I suspect that in your call to COPY_GRAPHICS_REGION@

Syntax

[Kenneth_Smith]

John,

Only in the sense that it's sufficiently long to avoid character buffer errors for something like:

'%pl[x_min=-1000.0,x_max=+1000,dx=100,x_sigfigs=2,y_min=-1000.0,y_max=+1000,dy=100,y_sigfigs=2]'

[Martin_K]

Thanks Ken!

I read the section regarding the COPY_GRAPHICS_REGION@ many times in the on-line manual (which also you pasted in your previous post), but due to its vague formulation I did not understand it as a whole, since I did not know what is the real meaning of DX,DY and SX,SY values. Now, thanks to your explanation I know that he values DX,DY,SX,SY represent margins.

Nevertheless, I still have couple of questions. Please � have a look at the picture below (outlined a pixel coordinate system):



If my understanding (imagination) of a pixel coordinate system pictured above is correct (if not, correct me) the questions are:

A)
I defined the whole window area (DW,DH dimensions: 1400x800) to accommodate all texts, buttons, titles, caption, graphs but with NO margin option for the %PL (I did not use the option margin at all). The graph itself, where points and lines are drawn, was automatically created by %PL and bound by X,Y axes. This graph is smaller than 1400x800 window.
How can I automatically (by program or an option in %PL) detect the pixel coordinates of at least one corner of the graph (say � top left) bound by X,Y axes, where the input data (points) are drawn?
The procedure with GUESSING the offsets is not usable.

B)
When I loaded in the raster jpg image, it was placed within whole white area of the graph (overlaying graph title, X,Y captions and legend). This was � of coarse � not wanted, so I started to try to lower the image dimensions. Unfortunately, I did it manually, tentatively and in few iterations (I started to use the DX, DY values (margins) with 120, 80 � it did not fit, then I used 150, 90, it did not fit and finally � I guessed!!! the correct DX, DY values (200,200), where the top left corner of the vector graphs coincided with top-left corner of the loaded raster image. But this was all, no other dimension coincided, so I posted my questions in my previous post. Based on your answer, I used the MARGIN option for %PL and the result can be seen below:



I used the code (for adjusting raster image to vector graph):

[Martin_K]

still last sentence of my previous post:


I do not mean current scale options for %PL like SCALE=LINEAR,� I mean to tell %PL to draw all input data in linear scale � let say � 1:1000, it means SCALE=1000 or SCALE = 500, �).