[r] Use `libtiledbsoma` for array handles

[johnkerl]

Currently we're retaining a tiledb-r handle at open: https://github.com/single-cell-data/TileDB-SOMA/blob/11d53967ebbdcab3070ef94e0a0a7ac590dfafac/apis/r/R/TileDBArray.R#L31

And we use a temporary open-use-close at every single call to libtiledbsoma -- here is just one of many examples: https://github.com/single-cell-data/TileDB-SOMA/blob/11d53967ebbdcab3070ef94e0a0a7ac590dfafac/apis/r/src/metadata.cpp#L170-L193

On the one hand this might seem lower-pri: the redundant opens are a perf hit but they work.

But as discussed on #3060 we must do this in order to remove the tiledb-r dependency.

See also #3053 which @nguyenv is working on -- this is a case where we do currently require a second open for array reads.

From #3059:

Already done for groups

2406; [sc-55685].

Steps:

• Status quo is
  • There is a private$.tiledb_array with connection to self$object
  • This is set on construction/open, and closed on close
  • The libtiledbsoma handles are currently ephemeral
  • Every read / write / get_metadata etc which is libtiledbsoma-capable currently takes a uri and a ctx and does open/op/close
• Needing to be done:
  • Remove the last refrences to private$.tiledb_array with connection to self$object

  • 3129

  • Glitch: while with 3129 we are removing the last references to self$object within R/*.R, there are hundreds of lines of uses of $object within tests/testthat -- tracked on #3124
  • Then set a libtiledbsoma handle on construction/open, and close it on close
  • Modify all the methods to not take uri+ctx but rather the open handle

To check:

• Currently the array set_metadata does not error out when the array is opened for read -- this is a bug.
  • This should be auto-fixed by this issue, but, it needs to be tested.

[johnkerl]

Blocked by #3051 since here https://github.com/single-cell-data/TileDB-SOMA/blob/1.15.0rc2/apis/r/R/SOMADataFrame.R#L182 the self$object is used. Removing that private$.tiledb_array / self$object from the base class TileDBArray.R causes this query-condition logic to return empty results, in a way that makes unit-test cases fail in non-obvious ways.

We need to move the query-condition logic over from TileDB-R before we can continue here.

[viviannguyen]

I don't believe I'm the correct Vivian to be tagged here. You're probably looking for @nguyenv.

[johnkerl]

Oh no @viviannguyen sorry about that!! Have a nice day though!!