See enhancement issue #128 #136
Conversation
I added more functionality compared to Przemyslaw's initial idea, however, at the expense of adding a further dependency on the rOpenSci package RefManageR. I believe that it is a rather widely used package, and the enhancement pays of for reproducible research.
ghost
assigned 
|
Ok, I pulled to devel. |
|
Ah now I realized that eurostat had this travis failure already before this PR (see issue #129). R CMD build / check --as-cran on command line reveals this: Undocumented arguments in documentation object 'get_bibentry' Can you fix and PR to devel branch? It seems that after this everything is OK. |
|
It seems OK, however, my issue remains on travis and r-hub that the dependency of eurostat, sf does not build and that makes continuous integration impossible. (Same for my iotables package, which is at last really ready for more highlight, but I struggle with this, because it depends on eurostat). Or do you have any workaround to deal with this? I wonder if the sf dependency is necessary at all? It is related to creating choropleths, I guess, which is not a core functionality and could be a vignette topic. I generally find map creation a developing area where the best solutions change from year to year. |
|
In fact map visualizations are a topic where we have been planning to split a separate pkg (eurostat_geospatial or something) with @muuankarski et al. I am OK with removing sf from dependencies if it does not affect core functionality. We have listed all imports in R/firstlib.R - it seems that sf::st_read is needed in |
|
All right. It seems that we should indeed move the GIS aspects (and perhaps the more generic bibtex entry stuff) to another package (eurostat_geospatial?) to solve this. This has been planned quite some time already but not completed. I try to see if we can find time to do this by summer 2019, not sure if faster schedule is possible. A separate geo package would help to simplify dependency structures. |
I added more functionality compared to Przemyslaw's initial idea, however, at the expense of adding a further dependency on the rOpenSci package RefManageR.
I believe that it is a rather widely used package, and the enhancement pays of for reproducible research.