sf and sftime changes instead of sdst functions

[kyle-messier]

No description provided.

[eva0marques]

https://stackoverflow.com/questions/21435339/data-table-vs-dplyr-can-one-do-something-well-the-other-cant-or-does-poorly#:~:text=or%20less%20equivalent.-,data.,whereas%20dplyr%20generates%20it%20automatically

[eva0marques]

To do list

• sf and sftime added to NAMESPACE

In R/manipulate_spacetime_data.R (and tests/testthat/test-manipulate_spacetime_data.R):

• adapt is_stdt() to a new fonction check_mysftime() to see if mysftime has the right format (also a version for check_mysf()?)
• test for check_mysftime() and check_mysf()
• adapt convert_stobj_to_stdt() to as_mysftime() macro (function and UT), calling to following functions (already available and tested: dt_as_sf() and dt_as_sftime()):
  • dt_as_mysftime() (function and UT)
  • sf_as_mysftime() (function and UT)
  • sftime_as_mysftime() (function and UT)
  • spatraster_as_sftime() (function and UT)
  • spatvector_as_sftime() (function and UT)
  • spatrds_as_sftime() (function and UT)
• same but for as_mysf() from data.frame, data.table, sf, SpatRaster, SpatVector
• turn convert_stdt() into:
  • sftime_as_sf() (function and UT)
  • sftime_as_spatvector() (function and UT)
  • sftime_as_spatraster() (function and UT)
  • sftime_as_spatrds() (function and UT)
• remove redundancy in test for as_mysftime() (reuse all the other functions)

Future work: see how stars format conversion to sf/sftime works

[eva0marques]

In case anything goes wrong with this issue, I duplicated main branch before changes in a new branch "save_branch_stdt.

[eva0marques]

Notes:

• terra::vect() function works to convert sf class to SpatVector (but not for sftime class: see new function).

[eva0marques]

@mitchellmanware @sigmafelix
I have added to my todo to remove all calls to stdt functions in calc_covariates.R but I cannot find any of them anymore (which is very cool!!) But I just wanna check with you I am not hallucinating (it can happen sometimes 🫨🫣): did I miss something, you've already removed them?

[sigmafelix]

@eva0marques At least in amadeus functions I've been working on, I replaced lines with stdt calls with generic functions (e.g., sf or terra functions). stdt compatibility is necessary at this moment as cross-validation indexing functions in beethoven rely on stdt.

[eva0marques]

Ok, thanks! So to let you know, I implemented all the conversion functions you can see in my comment above.

[eva0marques]

I close this issue on amadeus and open one in beethoven.
As @sigmafelix said, we need to modify cross-validation functions in beethoven to remove the calls to stdt format. It implies to load Amadeus in Beethoven or to copy manipulate_space_time_data.R.