R help - Oct 2023 - rnrfa package problems

R help email

Hello  I?m trying to work my way through the rnrfa package

RJ-2016-036.pdf (r-project.org)
<https://journal.r-project.org/archive/2016/RJ-2016-036/RJ-2016-036.pdf>

I?ve got it loaded OK and can access the data by the catalogue() command -
gives me a tibble with names, grid refs etc

allStations <- catalogue()

allStations

This works OK



But when I try to run the example codes there are various problems

The code is lifted directly from the document btw it?s not mine

This bit to convert gives an error message:

# Option a: from OS grid reference to WGS84 >

osg_parse(gridRef = "NC581062", CoordSystem = "WGS84")

 # Option b: from OS grid reference to BNG

osg_parse(gridRef = "NC581062", CoordSystem = "BNG")

##

osg_parse seems to be built into the rnrfa package so I don?t know why it
doesn?t work?




Then this bounding box code works but doesn?t seem to capture any stations
although rubric says that it should capture stations in mid Wales - the
tibble is empty

# Define a bounding box.

bbox <- list(lonMin = -3.76, latMin = 52.43, lonMax = -3.67, latMax = 52.48)




The ggmap code following gives an error ? this may be because the bbox data
is empty but I?m not sure  The needed packages are loaded:

library(ggmap)

library(ggrepel)

m <- get_map(location = as.numeric(bbox), maptype = 'terrain')

ggmap(m) + geom_point(data = someStations, aes(x = lon, y = lat),

                        + col = "red", size = 3, alpha = 0.5) +

geom_text_repel(data = someStations, aes(x = lon, y = lat, label = name),

                    + size = 3, col = "red")




I thought that I?d try the next bit to see anyway:

# Select stations with more than 100 years of recordings.

s100Y <- catalogue(minRec = 100, all = FALSE)

# Print s100Y to the screen.

s100Y

But again I get an error message

I haven?t gone any further through the article because there doesn?t seem
much point unless I can sort things.  I am puzzled as to why the examples
don?t work ? I?ve got all the needed packages loaded etc

If anyone can help/explain I?d be v grateful

Thanks Nick Wray

	[[alternative HTML version deleted]]

On Fri, 13 Oct 2023 09:11:25 +0100
Nick Wray <nickmwray at gmail.com> wrote:
> But when I try to run the example codes there are various problems
> 
> The code is lifted directly from the document btw it?s not mine
> 
> This bit to convert gives an error message:
> 
> # Option a: from OS grid reference to WGS84 >
> 
> osg_parse(gridRef = "NC581062", CoordSystem = "WGS84")
When reporting problems, it's best to list (1) your actions, (2) their
expected result and (3) their actual result:
https://www.chiark.greenend.org.uk/~sgtatham/bugs.html

In particular, it's best to report error messages together with the
code that causes them. This is what I get:

osg_parse(gridRef = "NC581062", CoordSystem = "WGS84")
# ?????? ? osg_parse(gridRef = "NC581062", CoordSystem =
"WGS84") :
#   ???????????????? ????????? (gridRef = "NC581062", CoordSystem #  
"WGS84")

Presumably, your error message looks a bit differently, but the
underlying cause is the same: the package has evolved since the R
Journal article has been published. In particular, the argument names
have been changed in 2018 with a giant commit commented "lint free":
https://github.com/ilapros/rnrfa/commit/258f11d89726c1031cb5bfa73e2f819d984914f3#diff-acbd0646da1333775e213e83567eaf53cb9e791ff7ba1200d462005e65fd7694L24-R30

I think you can use the argument name grid_refs instead of gridRef
and coord_system instead of CoordSystem:

osg_parse(grid = "NC581062", coo = "WGS84")
# $lon
# [1] -4.404049
# 
# $lat
# [1] 58.02222

Same goes for your catalogue() problems, which can be easily resolved
by calling help() on the respective functions to look up their new
argument names.
> The ggmap code following gives an error ? this may be because the bbox
> data is empty but I?m not sure  The needed packages are loaded:
> m <- get_map(location = as.numeric(bbox), maptype = 'terrain')
Is this the error you get?
>> Error in `get_googlemap()`:
>> ! Google now requires an API key; see `ggmap::register_google()`.
Does it help to follow the instructions in help(register_google,
package = ggmap)? Google inflected this change on the world in 2018,
approximately 1.5 years after the article had been published.

OpenStreetMap support is broken in ggmap
<https://github.com/dkahle/ggmap/issues/117>, but passing the source
'stamen' argument to ggmap::get_map seems to work. Well, partially. I
also get half an error message rendered right into the map telling me
something about "these basemap[s]... no longer... as of October..."
and
asking me to upgrade by visiting stamen.com.

This is more or less inevitable when a package depends on an external
service to keep providing map data (you can't bundle the whole map of
the world into a CRAN package, after all), so unless someone steps up
and implements OpenStreetMap's stable API support in ggmap, your best
bet is to ask Google for an API key and pray they won't alter the deal
any further.

-- 
Best regards,
Ivan