Support for manually created tree sequences with spatial information

[bodkan]

While doing some follow up inference work, it turned out that slendr's ts_load() function makes an implicit assumption that the only spatial tree sequences (those which store spatial coordinates in the individual tree sequence table) are those coming from SLiM. Naturally, loading a manually created tree sequence (a tree sequence composed from individual tables) ignored spatial coordinate columns.

So, with a tree sequence created like this:

nodes <- tribble(
  ~id, ~is_sample, ~time, ~individual,
  0,   1,          0,     0,
  1,   1,          0,     1,
  2,   0,          3,     2,
  3,   1,          3,     3,
  4,   0,          6,     4
)

edges <- tribble(
  ~left, ~right, ~parent, ~child,
  0,     10,     2,       0,
  0,     10,     2,       1,
  0,     10,     4,       3,
  0,     10,     4,       2

)

individuals <- tribble(
  ~id, ~flags, ~x,  ~y,
  0,   0,      2,   4,
  1,   0,      8,   7,
  2,   0,      5,   5.5,
  3,   0,      4,   2,
  4,   0,      4.5, 3.75
)

ts <- compose_ts(nodes, edges, individuals)

calling ts_nodes(ts) on a serialized and ts_load()-ed tree sequence returns this (note the absence of coordinates of nodes):

This PR is a mildly terrifying attempt at a few surgical cuts deep into the slendr internals to make ts_load() load spatial columns even for such tree sequences without breaking anything else.

With changes introduced by this PR, ts_load() and subsequent ts_nodes(ts) do what they're supposed to (note the presence of the location column as well as the fact that the data is now in the spatial sf data format):

The question now is, did something else break in the process? Guess we will find out!

[bodkan]

Oh, while I'm at this -- I had to turn off the testthat pipeline because of sudden timeout issues on GitHub Actions' part. I should try to switch this back up, to see if things have been fixed in the meantime.

[codecov[bot]]

Codecov Report

Merging #144 (d3be498) into main (001ee58) will increase coverage by 7.25%. The diff coverage is 81.81%.

@@            Coverage Diff             @@
##             main     #144      +/-   ##
==========================================
+ Coverage   78.34%   85.60%   +7.25%     
==========================================
  Files           7        7              
  Lines        3117     3125       +8     
==========================================
+ Hits         2442     2675     +233     
+ Misses        675      450     -225     

Files	Coverage Δ
R/tree-sequences.R	88.18% <81.81%> (+3.14%)	:arrow_up:

... and 6 files with indirect coverage changes

:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more