[GRASS-user] d.out.file / d.save - Shell-script use issues

[César Augusto Ramírez Franco]

Hi Peter,

As far as I understand, d.out.file only works with X monitors, the workflow
for a PNG monitor would be as follows:

export GRASS_PNGFILE=test.png
export GRASS_WIDTH=800
export GRASS_HEIGHT=600
d.mon start=PNG
d.rast somerastermap
d.vect somevectormap
d.mon stop=PNG

You could do this entirely on the shell or on the script, but you must
specify the PNG Driver variables before starting the PNG monitor.

This is what I tested with the PNG monitor:

d.mon start=PNG
PNG: GRASS_TRUECOLOR status: TRUE
PNG: collecting to file: map.png,
GRASS_WIDTH=640, GRASS_HEIGHT=480
Graphics driver [PNG] started

d.rast somerastermap
d.vect somevectormap

./test.sh
Currently selected monitor: PNG
ERROR: Only X monitors are supported.

d.mon stop=PNG
Monitor 'PNG' terminated

Please let me know if what you're trying to achieve with the PNG monitor
differs from what I tried

El Tue Nov 11 2014 at 4:02:10 p. m., "Peter Löwe" <peter.loewe at gmx.de>
escribió:

> Hi César,
>
> thanks for your follow-up work in this issue: That's exactly the workflow
> I was looking for. So the issue got already fixed between GRASS6.4.2 and
> 6.4.4. Could you please verify the behaviour of d.out.file in GRASS6.4.4
> when a PNG monitor is being used ?
>
> Best,
> Peter
>
> <peter.loewe at gmx.de>
>   Hello Peter,
>
> I'm sorry for misunderstanding your question, I thought that starting the
> monitor INSIDE the script would be adequate for you, I have tested this and
> couldn't reproduce your problem:
>
> I created a test.sh file on my current directory and made it executable
> with the following contents:
>
>  #!/bin/bash
> d.mon -p
> d.out.file output=test20141110 format=geotiff resolution=1 size=1440,720
>
> Then I started a monitor with some maps: d.mon x0, d.rast somerastermap,
> d.vect somevectormap
>
> Then I executed the script with ./test.sh
>
> I got this output:
>
>  Currently selected monitor: x0
> Saving display from Monitor: [x0] to <test20141110.tif>.
> Image size [1440 x 720]
>  100%
> Screen export complete. (writing the file may take a small amount of time)
> Image crop [720 x 720]
> Waiting for file to write ...
> Translating to GeoTIFF format
> Done.
>
> I'm using GRASS 6.4.4 on Ubuntu 14.04 from ubuntugis PPA, let me now if
> I'm not testing it correctly
>
> El Tue Nov 11 2014 at 11:03:24 a. m., "Peter Löwe" <peter.loewe at gmx.de>
> escribió:
>>
>>  Dear César, dear all,
>>
>> thanks for your input, but the core problem remains:
>>
>> When a Shell-script is invoked from the GRASS-shell (GRASS6.4.2 on Suse
>> Linux on a HPC cluster) with a monitor (x0, PNG, etc.) being already
>> activated (d.mon start / d.mon select), a g.gisenv _within_ the script will
>> properly reference the active monitor ("MONITOR=x0"), BUT a "d.save -c"
>> will throw a "WARNING: No socket to connect to for monitor" while
>> "d.out.file" gives an "ERROR: no monitor selected".
>>
>> A feasible workaround would be to store the monitors content with d.save
>> before running the script and recreating the content afterwards, but this
>> is less than elegant.
>>
>> Suggestions, anyone ?
>>
>> Best,
>> Peter
>>
>>
>>
>> <peter.loewe at gmx.de>
>>
>>
>> What I use on scripts is the PNG monitor setting the GRASS_PNGFILE,
>> GRASS_HEIGHT and GRASS_WIDTH before starting the monitor.
>>
>> Take a look at the PNG Driver manual pages for more variables:
>> http://grass.osgeo.org/grass64/manuals/pngdriver.html
>>
>> El lun, 10 de noviembre de 2014 17:42, "Peter Löwe" <peter.loewe at gmx.de>
>> escribió:
>>>
>>> Hi all,
>>>
>>> I'm trying call up d.out.file from within an add-on Shell-script for
>>> GRASS 6.>=4, but keep getting the following error:
>>> ERROR: No such monitor as <GRASS_MONITOR>
>>> d.out.file output=test20141110 format=geotiff resolution=1 size=1440,720
>>> ERROR: You must select a display monitor.
>>>
>>> Let's assume that a GRASS monitor has been _started_ and _selected_ (x0,
>>> PNG, or similar) in the current GRASS session and some rasters and vectors
>>> have been drawn on it. Then, the script is invoked which needs to write out
>>> the content of the currently active monitor.
>>>
>>> Up till now neither a simple "export `g.gisenv`" or an explicit
>>> re-selecting the monitor via "d.mon select=$THECURRENTMONITOR" within the
>>> script improved the situation.
>>>
>>> Hints and suggestions would be greatly appreciated.
>>>
>>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20141111/16b90b2e/attachment.html>