Add windows and macos platforms to CI (#383)

* Add windows and macos to strategy matrix

* Make pypairix dep platform-dependent

* Make pysam dep platform-dependent

* Fix drive letter colon bug on Windows

* test: Skip tabix or pairix loading tests if libs not available

* Do int cast to int64 instead of platform-dependent int

* Force unix-style line endings on tsv outputs

* Clean up temp files manually on Windows

See issues with NamedTemporaryFile on Windows: https://docs.python.org/3.12/library/tempfile.html#tempfile.NamedTemporaryFile

* tests: Skip digest test on windows (irreproducible issue only observed on 32-bit action runner)

.github/workflows/ci.yml

@@ -25,17 +25,25 @@ jobs:
         hatch run lint
 
   Test:
-    runs-on: ubuntu-latest
     strategy:
       matrix:
+        os: [ubuntu-latest]
         python-version: [ "3.8", "3.9", "3.10", "3.11" ]
+        include:
+          - os: windows-latest
+            python-version: "3.11"
+          - os: macos-latest
+            python-version: "3.11"
+
+    runs-on: ${{ matrix.os }}
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python-version }}
-      - run: |
+      - name: Run tests on ${{ matrix.os }}
+        run: |
           python -m pip install --upgrade pip
           pip install -e .[dev]
           # stop the build if there are Python syntax errors or undefined names
@@ -44,6 +52,6 @@ jobs:
 
   Release:
     # Only run on v* tag events
-    if: startsWith(github.ref, 'refs/tags')
+    if: startsWith(github.ref, 'refs/tags/v')
     needs: [Lint, Test]
     uses: open2c/cooler/.github/workflows/publish.yml@master

pyproject.toml

@@ -76,9 +76,9 @@ complete = [
     "ipytree>=0.2.2",
     "ipywidgets>=8.0.0",
     "matplotlib",
-    "pypairix",
+    "pypairix; platform_system != 'Windows'",
     "psutil",
-    "pysam",
+    "pysam; platform_system != 'Windows'",
 ]
 dev = [
     "cooler[complete,test]",

src/cooler/api.py

@@ -596,7 +596,7 @@ def _loc_slice(df, beg, end):
 
     # Select bin annotations that correspond to the bin1 IDs in the pixels df
     if "bin1_id" in columns:
-        bin1 = pixels["bin1_id"].to_numpy().astype(int, copy=False, casting="safe")
+        bin1 = pixels["bin1_id"].to_numpy().astype(np.int64, copy=False, casting="safe")
         if len(bin1) == 0:
             bmin = bmax = 0
         elif len(bins) > len(pixels):
@@ -612,7 +612,7 @@ def _loc_slice(df, beg, end):
 
     # Select bin annotations that correspond to the bin2 IDs in the pixels df
     if "bin2_id" in columns:
-        bin2 = pixels["bin2_id"].to_numpy().astype(int, copy=False, casting="safe")
+        bin2 = pixels["bin2_id"].to_numpy().astype(np.int64, copy=False, casting="safe")
         if len(bin2) == 0:
             bmin = bmax = 0
         elif len(bins) > len(pixels):

src/cooler/cli/_util.py

@@ -98,9 +98,10 @@ def parse_field_param(arg, includes_colnum=True, includes_agg=True):
 
 
 def parse_bins(arg):
-    # Provided chromsizes and binsize
-    if ":" in arg:
-        chromsizes_file, binsize = arg.split(":")
+    # User provided chromsizes path and binsize
+    arg_nodrive = op.splitdrive(arg)[1]  # Remove drive + colon if on Windows
+    if ":" in arg_nodrive:
+        chromsizes_file, binsize = arg.rsplit(":", maxsplit=1)
         if not op.exists(chromsizes_file):
             raise ValueError(f'File "{chromsizes_file}" not found')
         try:

src/cooler/cli/digest.py

@@ -71,6 +71,8 @@ def digest(chromsizes, fasta, enzyme, out, header, rel_ids):
         f = open(out, "w")
 
     if header:
-        frags[0:0].to_csv(f, sep="\t", index=False, header=True)
-    frags.to_csv(f, sep="\t", index=False, header=False)
+        frags[0:0].to_csv(
+            f, sep="\t", lineterminator='\n', index=False, header=True
+        )
+    frags.to_csv(f, sep="\t", lineterminator='\n', index=False, header=False)
     f.flush()

src/cooler/cli/dump.py

@@ -280,13 +280,20 @@ def dump(
         if is_first_chunk:
             if header:
                 chunk[0:0].to_csv(
-                    f, sep="\t", index=False, header=True, float_format=float_format
+                    f,
+                    sep="\t",
+                    lineterminator='\n',
+                    index=False,
+                    header=True,
+                    float_format=float_format,
+                    na_rep=na_rep,
                 )
             is_first_chunk = False
 
         chunk.to_csv(
             f,
             sep="\t",
+            lineterminator='\n',
             index=False,
             header=False,
             float_format=float_format,

src/cooler/cli/makebins.py

@@ -54,6 +54,8 @@ def makebins(chromsizes, binsize, out, header, rel_ids):
         f = open(out, "w")
 
     if header:
-        bins[0:0].to_csv(f, sep="\t", index=False, header=True)
-    bins.to_csv(f, sep="\t", index=False, header=False)
+        bins[0:0].to_csv(
+            f, sep="\t", lineterminator='\n', index=False, header=True
+        )
+    bins.to_csv(f, sep="\t", lineterminator='\n', index=False, header=False)
     f.flush()

src/cooler/create/_create.py

@@ -1,3 +1,4 @@
+import os
 import os.path as op
 import posixpath
 import tempfile
@@ -464,6 +465,7 @@ def create(
 
     """
     file_path, group_path = parse_cooler_uri(cool_uri)
+    file_path = op.realpath(file_path)
 
     if mode is None:
         mode = "a" if append else "w"
@@ -708,11 +710,16 @@ def create_from_unordered(
 
     dtypes = _get_dtypes_arg(dtypes, kwargs)
 
+    is_windows = os.name == "nt"
+    if is_windows and delete_temp:
+        delete = False
+    else:
+        delete = delete_temp
     temp_files = []
 
     # Sort pass
     tf = tempfile.NamedTemporaryFile(
-        suffix=".multi.cool", delete=delete_temp, dir=temp_dir
+        suffix=".multi.cool", delete=delete, dir=temp_dir
     )
     temp_files.append(tf)
     uris = []
@@ -730,7 +737,7 @@ def create_from_unordered(
         edges = np.linspace(0, n, int(np.sqrt(n)), dtype=int)
 
         tf2 = tempfile.NamedTemporaryFile(
-            suffix=".multi.cool", delete=delete_temp, dir=temp_dir
+            suffix=".multi.cool", delete=delete, dir=temp_dir
         )
         temp_files.append(tf2)
         uris2 = []
@@ -763,6 +770,11 @@ def create_from_unordered(
     logger.info(f"Merging into {cool_uri}")
     create(cool_uri, bins, chunks, columns=columns, dtypes=dtypes, mode=mode, **kwargs)
 
+    if is_windows and delete_temp:
+        for tf in temp_files:
+            if not tf.closed:
+                tf.close()
+            os.remove(tf.name)
     del temp_files
 
 

src/cooler/create/_ingest.py

@@ -581,8 +581,8 @@ def aggregate(self, chrom):  # pragma: no cover
                     np.searchsorted(start_abspos, abspos2, side="right") - 1
                 )
             else:
-                rel_bin1 = np.floor(table["cut1"] / binsize).astype(int)
-                rel_bin2 = np.floor(table["cut2"] / binsize).astype(int)
+                rel_bin1 = np.floor(table["cut1"] / binsize).astype(np.int64)
+                rel_bin2 = np.floor(table["cut2"] / binsize).astype(np.int64)
                 table["bin1_id"] = chrom_binoffset[table["chrom_id1"].values] + rel_bin1
                 table["bin2_id"] = chrom_binoffset[table["chrom_id2"].values] + rel_bin2
 

src/cooler/util.py

@@ -368,8 +368,8 @@ def digest(fasta_records, enzyme):
         raise ValueError(f"Unknown enzyme name: {enzyme}") from e
 
     def _each(chrom):
-        seq = bioseq.Seq(str(fasta_records[chrom]))
-        cuts = np.r_[0, np.array(cut_finder(seq)) + 1, len(seq)].astype(int)
+        seq = bioseq.Seq(str(fasta_records[chrom][:]))
+        cuts = np.r_[0, np.array(cut_finder(seq)) + 1, len(seq)].astype(np.int64)
         n_frags = len(cuts) - 1
 
         frags = pd.DataFrame(

tests/test_cli.py

@@ -6,6 +6,7 @@
 import traceback; traceback.print_tb(result.exception.__traceback__)
 
 """
+import os
 import os.path as op
 
 import click
@@ -76,6 +77,7 @@ def test_makebins():
     )
 
 
+@pytest.mark.skipif(os.name == "nt", reason="Non-reproducible off-by-one bug on Win32 GitHub runner")
 def test_digest():
     runner = CliRunner()
     result = runner.invoke(

tests/test_create_ingest.py

@@ -1,4 +1,5 @@
 # import filecmp
+import importlib.util
 import os
 import os.path as op
 import tempfile
@@ -15,6 +16,8 @@
 from cooler.cli.cload import tabix as cload_tabix
 from cooler.cli.load import load
 
+pysam_missing = importlib.util.find_spec("pysam") is None
+pairix_missing = importlib.util.find_spec("pairix") is None
 _pandas_major_version = int(pd.__version__.split(".")[0])
 
 tmp = tempfile.gettempdir()
@@ -175,6 +178,7 @@ def _mock_hdf5_pairs():
     should_work_with_int32_cols(chromsizes, bintable, mock_pairs)
 
 
+@pytest.mark.skipif(pysam_missing, reason="pysam not installed")
 @pytest.mark.filterwarnings("ignore")
 @pytest.mark.parametrize(
     "bins_path,pairs_path,ref_path,nproc",
@@ -214,6 +218,7 @@ def test_cload_tabix(bins_path, pairs_path, ref_path, nproc):
         pass
 
 
+@pytest.mark.skipif(pairix_missing, reason="pairix not installed")
 @pytest.mark.parametrize(
     "bins_path,pairs_path,ref_path,nproc",
     [