R compatibility

This PR makes sure Phlorest phylogenies can be read by several common
R packages.

closes #36

.github/workflows/python-package.yml

@@ -15,7 +15,7 @@ jobs:
         python-version: [3.8, 3.9, "3.10", "3.11"]
 
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
     - name: Set up Python ${{ matrix.python-version }}
       uses: actions/setup-python@v4
       with:

README.md

@@ -79,9 +79,10 @@ cldfbench makecldf --glottolog-version vX.Y --with-cldfreadme cldfbench_<id>.py
 pytest
 cldfbench zenodo --communities phlorest cldfbench_<id>.py
 cldfbench readme cldfbench_<id>.py
+phlorest check cldfbench_<id>.py --with-R
 git commit -a -m"release vX.Y"
 git push origin
-cldfbench release cldfbench_<id>.py vX.Y
+phlorest release cldfbench_<id>.py vX.Y
 ```
 
 

RELEASING.md

@@ -3,12 +3,12 @@ Releasing phlorest
 ==================
 
 - Do platform test via tox:
-```
+```shell
 tox -r
 ```
 
 - Make sure flake8 passes:
-```
+```shell
 flake8 src
 ```
 
@@ -34,7 +34,7 @@ twine upload dist/*
 ```
 
 - Push to github:
-```
+```shell
 git push origin
 git push --tags
 ```

setup.cfg

@@ -23,6 +23,7 @@ classifiers =
     Programming Language :: Python :: 3.9
     Programming Language :: Python :: 3.10
     Programming Language :: Python :: 3.11
+    Programming Language :: Python :: 3.12
     License :: OSI Approved :: Apache Software License
 
 [options]
@@ -33,7 +34,7 @@ package_dir =
 python_requires = >=3.8
 install_requires =
     newick>=1.9
-    commonnexus>=1.5
+    commonnexus>=1.8
     cldfviz>=0.10
     pycldf>=1.33.0
     clldutils
@@ -105,7 +106,7 @@ show_missing = true
 skip_covered = true
 
 [tox:tox]
-envlist = py38, py39, py310, py311
+envlist = py38, py39, py310, py311, py312
 isolated_build = true
 skip_missing_interpreter = true
 

src/phlorest/check.R

@@ -0,0 +1,58 @@
+# install.packages('ape')
+# if (!require("BiocManager", quietly = TRUE))
+#     install.packages("BiocManager")
+# BiocManager::install("treeio")
+# install.packages('rncl')
+# remotes::install_github("ropensci/tracerer")
+# install.packages('testthat')
+
+library(ape) # APE: the standard package, no annotations
+suppressMessages(library(treeio)) # treeio: considered the 'newer' and 'better' package for trees
+library(rncl) # rncl: based on nexus class library, so should be fast but no annotations.
+library(tracerer)  # tracerer: loads BEAST annotations
+
+library(testthat)
+
+args <- commandArgs(trailingOnly = TRUE)
+treefile <- args[1]
+ntrees <- strtoi(args[2])
+res <- 0
+
+readers <- list(
+  "ape" = function(x) ape::read.nexus(x, force.multi = TRUE),
+  "treeio" = function(x) treeio::read.beast(x),
+  "rncl" = function(x) {
+    trees <- rncl::read_nexus_phylo(x)
+    # We force multi "by hand":
+    if (class(trees) != 'multiPhylo'){
+      trees <- list(c(trees))
+      class(trees) <- "multiPhylo"
+    }
+    trees
+  },
+  "tracerer" = function(x) tracerer::parse_beast_trees(x)
+)
+
+for (rdr in names(readers)) {
+  # Only use tracerer with files with multiple trees, because it requires tree names to start
+  # with STATE_ ...
+  if (!(ntrees == 1 && rdr == 'tracerer')) {
+    cat(sprintf("READER: %s -- %s\n", rdr, treefile))
+    tryCatch(
+      expr = {
+        t <- readers[[rdr]](treefile)
+        # Make sure we read the correct number of trees:
+        testthat::expect_equal(length(t), ntrees)
+        cat('OK\n')
+      },
+      error = function(e) {
+        res <<- 1
+        print(e)
+      },
+      warning = function(w) {
+        print(w)
+      }
+    )
+  }
+}
+quit(status=res)

src/phlorest/cldfwriter.py

@@ -164,7 +164,8 @@ def add_posterior(self,
             self.add_tree(
                 tree,
                 self.posterior,
-                'posterior-{}'.format(i),
+                # We use a name format that works with the `tracerer` package for R:
+                'STATE_{}'.format(i),
                 metadata,
                 log,
                 'sample',

src/phlorest/commands/check.py

@@ -1,14 +1,28 @@
 """
 Checks datasets for compliance
 """
+import shutil
+import pathlib
+import zipfile
+import subprocess
+
+from clldutils.path import TemporaryDirectory
 from termcolor import colored
 from cldfbench.cli_util import add_dataset_spec, get_dataset
+from commonnexus import Nexus
 
 from phlorest.check import run_checks
 
 
 def register(parser):  # pragma: no cover
     add_dataset_spec(parser)
+    parser.add_argument(
+        '--with-R',
+        action='store_true',
+        help="Make sure the NEXUS files of the dataset can be read with commonly used R packages."
+             "\nNOTE: This requires the Rscript command and an R installation with the relevant "
+             "packages.",
+        default=False)
 
 
 def run(args, d=None):
@@ -19,5 +33,29 @@ def run(args, d=None):
             args.log.error("Unable to load %s - %s" % (args.dataset, e))
             raise
 
-    msg, color = ('PASS', 'green') if run_checks(d, args.log) else ('FAIL', 'red')
+    success = True
+    if args.with_R:  # pragma: no cover
+        for fname in {'summary.trees', 'posterior.trees.zip'}:
+            if d.cldf_dir.joinpath(fname).exists():
+                p = d.cldf_dir / fname
+                with TemporaryDirectory() as tmp:
+                    if fname.endswith('.zip'):
+                        with zipfile.ZipFile(p) as zip:
+                            zip.extract(zip.infolist()[0], tmp)
+                            assert tmp.joinpath(p.stem).exists()
+                            p = tmp / p.stem
+                    else:
+                        shutil.copy(p, tmp / p.name)
+                        p = tmp / p.name
+                    ntrees = len(Nexus.from_file(p).TREES.commands['TREE'])
+                    res = subprocess.call([
+                        'Rscript',
+                        str(pathlib.Path(__file__).parent.parent / 'check.R'),
+                        str(p),
+                        str(ntrees),
+                    ])
+                    if res:
+                        success = False
+
+    msg, color = ('PASS', 'green') if run_checks(d, args.log) and success else ('FAIL', 'red')
     print('{} {}'.format(colored(msg, color, attrs=['bold']), d.id))

src/phlorest/nexuslib.py

@@ -33,9 +33,9 @@ def _rescaler(factor, n):
             nwk = tree.newick
             nwk.visit(functools.partial(_rescaler, RESCALE_TO_YEARS[orig_scaling]))
             trees.append((tree.name, nwk, tree.rooted))
-        nex.replace_block(
-            nex.TREES,
-            Trees.from_data(*trees, **nex.TREES.TRANSLATE.mappings if nex.TREES.TRANSLATE else {}))
+        kwarg = nex.TREES.TRANSLATE.mappings if nex.TREES.TRANSLATE else {}
+        kwarg.update(lowercase_command=True)
+        nex.replace_block(nex.TREES, Trees.from_data(*trees, **kwarg))
         return nex
     raise ValueError('Cannot rescale {} to years.'.format(orig_scaling))
 
@@ -104,7 +104,8 @@ def __enter__(self):
 
     def __exit__(self, exc_type, exc_val, exc_tb):
         if self._trees:
-            nex = Nexus.from_blocks(Trees.from_data(*self._trees))
+            nex = Nexus.from_blocks(
+                Trees.from_data(*self._trees, **dict(lowercase_command=True)))
             nex.to_file(self.path)
             if self.zipped:
                 with zipfile.ZipFile(

tests/test_cli.py

@@ -8,7 +8,7 @@
 
 
 def test_check(dataset, caplog):
-    check.run(argparse.Namespace(log=logging.getLogger(__name__)), dataset)
+    check.run(argparse.Namespace(log=logging.getLogger(__name__), with_R=False), dataset)
     assert len(caplog.records) >= 4
 
 
@@ -26,7 +26,7 @@ class DS(Dataset):
         dir = tmp_repos
         id = 'phy'
 
-    check.run(argparse.Namespace(log=logging.getLogger(__name__)), DS())
+    check.run(argparse.Namespace(log=logging.getLogger(__name__), with_R=False), DS())
 
 
 def test_main(dataset):