[Bug] i.image.bathymetry needs corrections to the R scripting calls to run all options about grass-addons

[newcomb-d]

Upon further investigation, it was noted that rgeos was not loaded,
Checking rgeos availability: FALSE
Note: when rgeos is not available, polygon geometry computations in maptools depend on gpclib,
which has a restricted licence. It is disabled by default;
to enable gpclib, type gpclibPermit()
and the reason that the raster module was failing was that the sp libary sp() was not being used.
When line 236 if the script was modified from :
libs = ['GWmodel', 'data.table', 'rgrass7', 'rgdal', 'raster']
to:
libs = ['GWmodel', 'data.table', 'sp', 'rgeos', 'rgrass7', 'rgdal', 'raster']
and the following line was added after line 248
r_file.write('use_sp()\n')

The script was then able to get past this error. However, in line 293
predict = g.read_command("g.tempfile", pid=os.getpid()).strip() + '.txt'
a temporary file is written to hold the results of the GWR analysis.
On windows, this leads to a filename like this:
D:\grass\lake\u18wgs84\bathy.tmp/unknown\13752.5.txt , which throws an escape character error in R on windows.
This was worked around by setting a static filename in the directory that the script is run from:
predict="tempgwr.txt"
While this works for a desktop user and makes things os-independent, it requires the user have write access to the directory the script is running in.
It might be better to set the filename with a suffix more dynamically from the os.getpid process for a result like tempgwr13725.txt in more dynamic settings.

[neteler]

Please consider to edit the Python script directly here and submit the changes as a pull request:
https://github.com/OSGeo/grass-addons/blob/master/grass7/imagery/i.image.bathymetry/i.image.bathymetry.py

[newcomb-d]

Sorry for the delay. Working with external user who can submit via web interface , but is still getting up to speed on command line git , which seems to be the project preference. Doug

…

________________________________ From: Markus Neteler <[email protected]> Sent: Sunday, October 11, 2020 8:43 AM To: OSGeo/grass-addons <[email protected]> Cc: Newcomb, Doug <[email protected]>; Author <[email protected]> Subject: [EXTERNAL] Re: [OSGeo/grass-addons] [Bug] i.image.bathymetry needs corrections to the R scripting calls to run all options (#289) This email has been received from outside of DOI - Use caution before clicking on links, opening attachments, or responding. Please consider to edit the Python script directly here and submit the changes as a pull request: https://github.com/OSGeo/grass-addons/blob/master/grass7/imagery/i.image.bathymetry/i.image.bathymetry.py — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#289 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABSYN7XQNMTIUKE7FRRFFVTSKGR6BANCNFSM4R3WYH2Q>.

[tmszi]

The script was then able to get past this error. However, in line 293
predict = g.read_command("g.tempfile", pid=os.getpid()).strip() + '.txt'
a temporary file is written to hold the results of the GWR analysis.
On windows, this leads to a filename like this:
D:\grass\lake\u18wgs84\bathy.tmp/unknown\13752.5.txt , which throws an escape character error in R on windows.

This is bug in the G__temp_element() function (line 121).

[neteler]

On windows, this leads to a filename like this:
D:\grass\lake\u18wgs84\bathy.tmp/unknown\13752.5.txt , which throws an escape character error in R on windows.

This is bug in the G__temp_element() function (line 121).

Probably G_convert_dirseps_to_host() is missing? See
https://grass.osgeo.org/programming7/paths_8c.html#a0bd23f09c109997ed8ddc25c5d650698

[tmszi]

On windows, this leads to a filename like this:
D:\grass\lake\u18wgs84\bathy.tmp/unknown\13752.5.txt , which throws an escape character error in R on windows.

This is bug in the G__temp_element() function (line 121).

Probably G_convert_dirseps_to_host() is missing? See
https://grass.osgeo.org/programming7/paths_8c.html#a0bd23f09c109997ed8ddc25c5d650698

Yes you are right or use HOST_DIRSEP.

Yesterday I prepared patch (need test on MS Win), but I have trouble with Cross-Compilation (environment) GRASS GIS for Windows OS on GNU/Linux. Otherwise I would have opened a new PR.

[neteler]

Right, HOST_DIRSEP is widely used in GRASS-core.

[neteler]

Yes you are right or use HOST_DIRSEP.

Yesterday I prepared patch (need test on MS Win), but I have trouble with Cross-Compilation (environment) GRASS GIS for Windows OS on GNU/Linux. Otherwise I would have opened a new PR.

While checking the entire core source code I found, besides lib/gis/tempfile.c two more files in which HOST_DIRSEP seems to be missing (unrelated but like worth to also fix in the same PR):

• lib/display/r_raster.c (2 times)
• display/d.mon/list.c (4 times)

[wenzeslaus]

Isn't the issue more that the backslash is used rather than missing? If the error is about "an escape character" in R on Windows, that suggests that the backslashes in path are not escaped or that the string in generated R code (is there some in this case?) needs to be raw. It seems to me then that if you want to change it in grass core, you need to remove the backslashes from path rather than adding more of them and use forward slashes instead.

The question also is if g.tempfile is the right tool. Is the file huge or copied to mapset non-temporary directories? If not, standard Python temporary directory object used with a with statement and a hardcoded file name might be a better option.

[newcomb-d]

I just noticed a similar error for r.in.wms for 7.8.7 on Windows.