python3Packages.cnvkit: fix build error and failing tests (#431215)

pkgs/development/python-modules/cnvkit/default.nix

@@ -3,9 +3,9 @@
   buildPythonPackage,
   fetchFromGitHub,
   fetchpatch,
-
+  python,
+  makeWrapper,
   # dependencies
-  R,
   biopython,
   matplotlib,
   numpy,
@@ -13,15 +13,15 @@
   pomegranate,
   pyfaidx,
   pysam,
-  rPackages,
   reportlab,
+  rPackages,
   scikit-learn,
   scipy,
-
+  R,
   # tests
   pytestCheckHook,
-}:
 
+}:
 buildPythonPackage rec {
   pname = "cnvkit";
   version = "0.9.12";
@@ -47,11 +47,38 @@ buildPythonPackage rec {
     "pomegranate"
   ];
 
-  # Numpy 2 compatibility
-  postPatch = ''
-    substituteInPlace skgenome/intersect.py \
-      --replace-fail "np.string_" "np.bytes_"
-  '';
+  nativeBuildInputs = [
+    makeWrapper
+  ];
+
+  buildInputs = [
+    R
+  ];
+
+  postPatch =
+    let
+      rscript = lib.getExe' R "Rscript";
+    in
+    # Numpy 2 compatibility
+    ''
+      substituteInPlace skgenome/intersect.py \
+        --replace-fail "np.string_" "np.bytes_"
+    ''
+    # Patch shebang lines in R scripts
+    + ''
+      substituteInPlace cnvlib/segmentation/flasso.py \
+        --replace-fail "#!/usr/bin/env Rscript" "#!${rscript}"
+
+      substituteInPlace cnvlib/segmentation/cbs.py \
+        --replace-fail "#!/usr/bin/env Rscript" "#!${rscript}"
+
+      substituteInPlace cnvlib/segmentation/__init__.py \
+        --replace-fail 'rscript_path="Rscript"' 'rscript_path="${rscript}"'
+
+      substituteInPlace cnvlib/commands.py \
+        --replace-fail 'default="Rscript"' 'default="${rscript}"'
+
+    '';
 
   dependencies = [
     biopython
@@ -61,12 +88,42 @@ buildPythonPackage rec {
     pomegranate
     pyfaidx
     pysam
-    rPackages.DNAcopy
     reportlab
+    rPackages.DNAcopy
     scikit-learn
     scipy
   ];
 
+  # Make sure R can find the DNAcopy package
+  postInstall = ''
+    wrapProgram $out/bin/cnvkit.py \
+      --set R_LIBS_SITE "${rPackages.DNAcopy}/library" \
+       --set MPLCONFIGDIR "/tmp/matplotlib-config"
+  '';
+
+  installCheckPhase = ''
+    runHook preInstallCheck
+
+    ${python.executable} -m pytest --deselect=test/test_commands.py::CommandTests::test_batch \
+      --deselect=test/test_commands.py::CommandTests::test_segment_hmm
+
+      cd test
+      # Set matplotlib config directory for the tests
+      export MPLCONFIGDIR="/tmp/matplotlib-config"
+      export HOME="/tmp"
+      mkdir -p "$MPLCONFIGDIR"
+
+      # Use the installed binary - it's already wrapped with R_LIBS_SITE
+      make cnvkit="$out/bin/cnvkit.py" || {
+        echo "Make tests failed"
+        exit 1
+      }
+
+    runHook postInstallCheck
+  '';
+
+  doInstallCheck = true;
+
   pythonImportsCheck = [ "cnvlib" ];
 
   nativeCheckInputs = [
@@ -74,13 +131,6 @@ buildPythonPackage rec {
     R
   ];
 
-  disabledTests = [
-    # AttributeError: module 'pomegranate' has no attribute 'NormalDistribution'
-    # https://github.com/etal/cnvkit/issues/815
-    "test_batch"
-    "test_segment_hmm"
-  ];
-
   meta = {
     homepage = "https://cnvkit.readthedocs.io";
     description = "Python library and command-line software toolkit to infer and visualize copy number from high-throughput DNA sequencing data";

pkgs/development/python-modules/pomegranate/default.nix

@@ -3,20 +3,15 @@
   stdenv,
   buildPythonPackage,
   fetchFromGitHub,
-
-  # build-system
+  fetchpatch,
+  pytestCheckHook,
   setuptools,
-
-  # dependencies
   apricot-select,
   networkx,
   numpy,
   scikit-learn,
   scipy,
   torch,
-
-  # tests
-  pytestCheckHook,
 }:
 
 buildPythonPackage rec {
@@ -27,34 +22,10 @@ buildPythonPackage rec {
   src = fetchFromGitHub {
     repo = "pomegranate";
     owner = "jmschrei";
-    # tag = "v${version}";
-    # No tag for 1.1.2
-    rev = "e9162731f4f109b7b17ecffde768734cacdb839b";
-    hash = "sha256-vVoAoZ+mph11ZfINT+yxRyk9rXv6FBDgxBz56P2K95Y=";
+    tag = "v${version}";
+    hash = "sha256-p2Gn0FXnsAHvRUeAqx4M1KH0+XvDl3fmUZZ7MiMvPSs=";
   };
 
-  # _pickle.UnpicklingError: Weights only load failed.
-  # https://pytorch.org/docs/stable/generated/torch.load.html
-  postPatch = ''
-    substituteInPlace \
-      tests/distributions/test_bernoulli.py \
-      tests/distributions/test_categorical.py \
-      tests/distributions/test_exponential.py \
-      tests/distributions/test_gamma.py \
-      tests/distributions/test_independent_component.py \
-      tests/distributions/test_normal_diagonal.py \
-      tests/distributions/test_normal_full.py \
-      tests/distributions/test_poisson.py \
-      tests/distributions/test_student_t.py \
-      tests/distributions/test_uniform.py \
-      tests/test_bayes_classifier.py \
-      tests/test_gmm.py \
-      tests/test_kmeans.py \
-      --replace-fail \
-        'torch.load(".pytest.torch")' \
-        'torch.load(".pytest.torch", weights_only=False)'
-  '';
-
   build-system = [ setuptools ];
 
   dependencies = [
@@ -72,11 +43,20 @@ buildPythonPackage rec {
     pytestCheckHook
   ];
 
-  disabledTestPaths = lib.optionals (stdenv.hostPlatform.isDarwin && stdenv.hostPlatform.isx86_64) [
+  patches = [
+    # Fix tests for pytorch 2.6
+    (fetchpatch {
+      name = "python-2.6.patch";
+      url = "https://github.com/jmschrei/pomegranate/pull/1142/commits/9ff5d5e2c959b44e569937e777b26184d1752a7b.patch";
+      hash = "sha256-BXsVhkuL27QqK/n6Fa9oJCzrzNcL3EF6FblBeKXXSts=";
+    })
+  ];
+
+  pytestFlagsArray = lib.optionals (stdenv.hostPlatform.isDarwin && stdenv.hostPlatform.isx86_64) [
     # AssertionError: Arrays are not almost equal to 6 decimals
-    "=tests/distributions/test_normal_full.py::test_fit"
-    "=tests/distributions/test_normal_full.py::test_from_summaries"
-    "=tests/distributions/test_normal_full.py::test_serialization"
+    "--deselect=tests/distributions/test_normal_full.py::test_fit"
+    "--deselect=tests/distributions/test_normal_full.py::test_from_summaries"
+    "--deselect=tests/distributions/test_normal_full.py::test_serialization"
   ];
 
   disabledTests = [