Running R now. Following output is R output. During startup - Warning messages: 1: Setting LC_CTYPE=en_GB.cp1252 failed 2: Setting LC_COLLATE=en_GB.cp1252 failed 3: Setting LC_TIME=en_GB.cp1252 failed 4: Setting LC_MONETARY=en_GB.cp1252 failed Loading required package: caret Loading required package: lattice Loading required package: ggplot2 Error: ‘\S’ is an unrecognized escape in character string starting "“C:\S” Execution halted ERROR: There was an error in the execution of the R script. Please check the R output. (Wed May 24 11:05:30 2017) Command finished (17 sec)
My GRASS mapset is located in - C:\Sandbox\RGRASS\TEST\MAP1
Given that R takes paths with / rather than ****, I think the issue lies at the start of the R part of the script which reads:
require(caret) features ← read.csv(“C:\Sandbox\RGRASS\TEST\MAP1.tmp/unknown\5636.0”, sep=“|”, header=TRUE, row.names=1) training ← read.csv(“C:\Sandbox\RGRASS\TEST\MAP1.tmp/unknown\5636.1”, sep=“|”, header=TRUE, row.names=1)
I have run v.class.mlR on an osgeolive build before, but not on Windows. Is this a Windows specific issue that needs addressing?
Le Thu, 25 May 2017 00:25:57 +0200,
Markus Neteler <neteler@osgeo.org> a écrit :
On Wed, May 24, 2017 at 12:14 PM, James Duffy
<james.philip.duffy@gmail.com> wrote:
...
> v.class.mlR --verbose segments_map=seg_stats_vec@MAP1
...
> My GRASS mapset is located in - C:\Sandbox\RGRASS\TEST\MAP1
>
> Given that R takes paths with / rather than \, I think the issue
> lies at the start of the R part of the script which reads:
>
> require(caret)
> features <-
> read.csv("C:\Sandbox\RGRASS\TEST\MAP1\.tmp/unknown\5636.0",
> sep="|", header=TRUE, row.names=1) training <-
> read.csv("C:\Sandbox\RGRASS\TEST\MAP1\.tmp/unknown\5636.1",
> sep="|", header=TRUE, row.names=1)
>
> I have run v.class.mlR on an osgeolive build before, but not on
> Windows. Is this a Windows specific issue that needs addressing?
In the script I see things like
reclass_files[classifier] = tmpfilename.replace("\\",
"/")
which is probably unhelpful on Windows. Not sure how to make that
portable, though.
I had forgotten that I had already done this elsewhere in the
module...
It is actually there exactly for Windows: tempfile() creates a path
with \, but in R these have to either be escaped, or modified into / as
R (and I think Windows in general) can handle paths with forward
slashes.
I had answered James, but I now notice that it was from a wrong
address, so the mail didn't get through to the list.
I've now just comitted a fix using the same technique.
James, can you just reinstall v.class.mlR and try again ?
Your suggested edits to v.class.mlR.py worked (except you had one too many brackets on the end of each line, so I removed those). Once I updated and saved that script file, the error disappeared.
I have just reinstalled the module using g.extension and tried running v.class.mlR and I again have the original error regarding escaped slashes.
v.class.mlR --verbose segments_map=seg_stats_vec@MAP1
…
My GRASS mapset is located in - C:\Sandbox\RGRASS\TEST\MAP1
Given that R takes paths with / rather than , I think the issue
lies at the start of the R part of the script which reads:
require(caret)
features ←
read.csv(“C:\Sandbox\RGRASS\TEST\MAP1.tmp/unknown\5636.0”,
sep=“|”, header=TRUE, row.names=1) training ←
read.csv(“C:\Sandbox\RGRASS\TEST\MAP1.tmp/unknown\5636.1”,
sep=“|”, header=TRUE, row.names=1)
I have run v.class.mlR on an osgeolive build before, but not on
Windows. Is this a Windows specific issue that needs addressing?
In the script I see things like
reclass_files[classifier] = tmpfilename.replace(“\”,
“/”)
which is probably unhelpful on Windows. Not sure how to make that
portable, though.
I had forgotten that I had already done this elsewhere in the
module…
It is actually there exactly for Windows: tempfile() creates a path
with , but in R these have to either be escaped, or modified into / as
R (and I think Windows in general) can handle paths with forward
slashes.
I had answered James, but I now notice that it was from a wrong
address, so the mail didn’t get through to the list.
I’ve now just comitted a fix using the same technique.
James, can you just reinstall v.class.mlR and try again ?
Moritz
–
James Duffy
PhD Researcher
Environment and Sustainability Institute
Penryn Campus
University of Exeter
Penryn
Cornwall
TR10 9FE
Le Thu, 25 May 2017 10:54:18 +0100,
James Duffy <james.philip.duffy@gmail.com> a écrit :
Hi Moritz,
Your suggested edits to v.class.mlR.py worked (except you had one too
many brackets on the end of each line, so I removed those). Once I
updated and saved that script file, the error disappeared.
I have just reinstalled the module using g.extension and tried running
v.class.mlR and I again have the original error regarding escaped
slashes.
I had just copied and pasted from my message, so the problem you found
with one too many brackets was also in the file. I just fixed that.
Sorry, don't have the opportunity to test myself right now, so doing
this a bit blindly.
I’ve just tried using a csv as input instead of a vector file and I get the same error, now relating to the path of the csv. Could I suggest also putting in a check / replace line for csv paths if they are entered from a windows machine (especially using the GUI).
Your suggested edits to v.class.mlR.py worked (except you had one too
many brackets on the end of each line, so I removed those). Once I
updated and saved that script file, the error disappeared.
I have just reinstalled the module using g.extension and tried running
v.class.mlR and I again have the original error regarding escaped
slashes.
I had just copied and pasted from my message, so the problem you found
with one too many brackets was also in the file. I just fixed that.
Sorry, don’t have the opportunity to test myself right now, so doing
this a bit blindly.
Moritz
–
James Duffy
PhD Researcher
Environment and Sustainability Institute
Penryn Campus
University of Exeter
Penryn
Cornwall
TR10 9FE
Le Thu, 25 May 2017 12:18:52 +0100,
James Duffy <james.philip.duffy@gmail.com> a écrit :
I've just tried using a csv as input instead of a vector file and I
get the same error, now relating to the path of the csv. Could I
suggest also putting in a check / replace line for csv paths if they
are entered from a windows machine (especially using the GUI).
I really don't understand why this hasn't hit anyone before. Maybe its
because my MS Windows colleagues work in a Jupyter notebook environment
and things are different there.
I've just committed a fix that uses the same replace technique for the
csv file inputs.
I don't know if for Python addon modules you also have to wait 24 to
get access via g.extension...
