Adding variants iterator and other functions

RcppTskit/NEWS.md

@@ -40,6 +40,28 @@ and releases adhere to [Semantic Versioning](https://semver.org/spec/v2.0.0.html
 - Added `rtsk_mutation_table_add_row()` and
   `TableCollection$mutation_table_add_row()` to append mutation rows from
   \code{R}, mirroring `tsk_mutation_table_add_row()`.
+- Added `rtsk_population_table_add_row()` and
+  `TableCollection$population_table_add_row()` to append population rows from
+  \code{R}, mirroring `tsk_population_table_add_row()`.
+- Added `rtsk_migration_table_add_row()` and
+  `TableCollection$migration_table_add_row()` to append migration rows from
+  \code{R}, mirroring `tsk_migration_table_add_row()`.
+- Added `rtsk_provenance_table_add_row()` and
+  `TableCollection$provenance_table_add_row()` to append provenance rows from
+  \code{R}, mirroring `tsk_provenance_table_add_row()`.
+- Added `rtsk_node_table_get_row()` and `TableCollection$node_table_get_row()`
+  to retrieve node-table rows by 0-based row index.
+- Added `rtsk_table_collection_sort()` and `TableCollection$sort()` to sort
+  table collections with 0-based `edge_start`, `site_start`, and
+  `mutation_start` semantics.
+- Added low-level variant iterators
+  (`rtsk_variant_iterator_init()`/`rtsk_variant_iterator_next()`) and a
+  user-facing `TreeSequence$variants()` method to iterate over decoded
+  site-by-site variants from \code{R}, aligned with `tskit` Python API
+  semantics for `samples`, `isolated_as_missing`, `alleles`, and
+  `left`/`right` intervals.
+- Added `rtsk_treeseq_get_samples()` and `TreeSequence$samples()` to retrieve
+  sample node IDs from a tree sequence.
 - TODO
 
 ### Changed
@@ -53,6 +75,21 @@ and releases adhere to [Semantic Versioning](https://semver.org/spec/v2.0.0.html
   `pointer` to `xptr`.
 - Ensured `TableCollection$tree_sequence()` matches `tskit Python` API:
   it now builds indexes on the `TableCollection`, if indexes are not present.
+- Expanded `TableCollection$*_table_get_row()` documentation with runnable
+  examples, Python `__getitem__` references, and explicit note that numeric
+  row indices are validated and converted to 32-bit integers.
+- Aligned low-level row getters with `tskit C` caveats by returning
+  `mutations = NULL` for `rtsk_site_table_get_row()`, exposing `edge` and
+  `inherited_state` (as `NULL` for table access) in
+  `rtsk_mutation_table_get_row()`, and exposing `nodes` in
+  `rtsk_individual_table_get_row()`.
+- Kept high-level `TableCollection$individual_table_get_row()`,
+  `TableCollection$site_table_get_row()`, and
+  `TableCollection$mutation_table_get_row()` aligned with `tskit Python`
+  row-shape semantics.
+- Refined integer validators to use a single 32-bit integer-like validation
+  path with explicit `strict` handling for integer-only versus numeric-allowed
+  inputs across scalar and optional vector arguments.
 - We now use `bit64::integer64` (signed 64 bit integer) instead of `int` aiming
   to approach `tsk_size_t` in `tskit C` (unsigned 64 bit integer); in low-level
   `rtsk_treeseq_get_num_*()` wrappers and count/metadata-length fields.