r_to_py() does not successfully convert sf objects to geopandas.GeoDataFrame

[philiporlando]

When passing an sf data.frame to r_to_py(), it is implicitly converted to a non-spatial pandas.DataFrame, and the geometry field is converted to a normal list. This transformation results in the loss of the coordinate reference system (CRS) information, which is crucial for spatial data analysis.

Are there any opportunities to convert sf objects to a geopandas.GeoDataFrame when using r_to_py()?

library(sf)
library(reticulate)

nc <- sf::st_read(system.file("shape/nc.shp", package="sf"))
class(nc)
# [1] "sf"         "data.frame"
all(sf::st_is_valid(nc$geometry))
# [1] TRUE


nc_py <- reticulate::r_to_py(nc)
class(nc_py)
# [1] "pandas.core.frame.DataFrame"        "pandas.core.generic.NDFrame"       
# [3] "pandas.core.base.PandasObject"      "pandas.core.accessor.DirNamesMixin"
# [5] "pandas.core.indexing.IndexingMixin" "pandas.core.arraylike.OpsMixin"    
# [7] "python.builtin.object" 
all(sf::st_is_valid(nc_py$geometry))
# Error in UseMethod("st_is_valid") : 
#   no applicable method for 'st_is_valid' applied to an object of class "c('pandas.core.series.Series', 'pandas.core.base.IndexOpsMixin', 'pandas.core.arraylike.OpsMixin', 'pandas.core.generic.NDFrame', 'pandas.core.base.PandasObject', 'pandas.core.accessor.DirNamesMixin', 'pandas.core.indexing.IndexingMixin', 'python.builtin.object')"

Here are my session details:

reticulate::py_config()
# python:         /home/porlando/Projects/do4ds/.venv/bin/python
# libpython:      /usr/lib/python3.10/config-3.10-x86_64-linux-gnu/libpython3.10.so
# pythonhome:     /home/porlando/Projects/do4ds/.venv:/home/porlando/Projects/do4ds/.venv
# version:        3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0]
# numpy:          /home/porlando/Projects/do4ds/.venv/lib/python3.10/site-packages/numpy
# numpy_version:  1.24.3
# 
# NOTE: Python version was forced by use_python() function

utils::sessionInfo()
# R version 4.2.1 (2022-06-23)
# Platform: x86_64-pc-linux-gnu (64-bit)
# Running under: Ubuntu 22.04.3 LTS
# 
# Matrix products: default
# BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.10.0
# LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.10.0
# 
# locale:
# [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8       
# [4] LC_COLLATE=en_US.UTF-8     LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
# [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C              
# [10] LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       
# 
# attached base packages:
# [1] stats     graphics  grDevices datasets  utils     methods   base     
# 
# other attached packages:
# [1] sf_1.0-14         dplyr_1.1.2       reticulate_1.34.0
# 
# loaded via a namespace (and not attached):
# [1] Rcpp_1.0.11        dbplyr_2.2.1       compiler_4.2.1     pillar_1.9.0       class_7.3-20      
# [6] RPostgres_1.4.5    tools_4.2.1        digest_0.6.29      bit_4.0.4          jsonlite_1.8.0    
# [11] lattice_0.20-45    lubridate_1.8.0    lifecycle_1.0.3    tibble_3.2.1       png_0.1-8         
# [16] pkgconfig_2.0.3    rlang_1.1.1        Matrix_1.5-1       DBI_1.1.3          cli_3.6.1         
# [21] rstudioapi_0.13    e1071_1.7-11       s2_1.1.0           janitor_2.1.0      stringr_1.4.0     
# [26] generics_0.1.3     vctrs_0.6.3        hms_1.1.1          classInt_0.4-7     bit64_4.0.5       
# [31] grid_4.2.1         tidyselect_1.2.0   snakecase_0.11.0   glue_1.6.2         R6_2.5.1          
# [36] fansi_1.0.3        tidyr_1.2.0        tzdb_0.3.0         purrr_0.3.4        readr_2.1.2       
# [41] blob_1.2.3         logger_0.2.2       magrittr_2.0.3     ellipsis_0.3.2     units_0.8-0       
# [46] assertthat_0.2.1   renv_0.16.0        KernSmooth_2.23-20 utf8_1.2.2         stringi_1.7.8     
# [51] proxy_0.4-27       wk_0.6.0     

[t-kalinowski]

Hi, that seems reasonable to me, though I'm not sure if taking a "Suggests" dependency to {sf} in reticulate is prudent. @kevinushey Do you have thoughts?

@philiporlando Would you like to submit a PR adding r_to_py() and py_to_r() methods?

[philiporlando]

@t-kalinowski, if adding {sf} to "Suggests" isn't an issue, then I can start working on a PR but I may need some additional support throughout the process.

[kevinushey]

I don't feel too strongly either way; I think adding sf as a Suggests is okay.

[gufranpathan]

Thank you for working on this. It would be a very useful feature. Looking forward!