truvari: init at 1.3.2 (#70877)

pythonPackages.pyvcf: init at 0.6.8

pkgs/applications/science/biology/truvari/default.nix

@@ -0,0 +1,43 @@
+{ lib
+, fetchFromGitHub
+, python3Packages
+}:
+
+python3Packages.buildPythonApplication rec {
+  pname = "truvari";
+  version = "1.3.2";
+
+  src = fetchFromGitHub {
+    owner = "spiralgenetics";
+    repo = "truvari";
+    rev = "v${version}";
+    sha256 = "0wmjz8nzibvj0ixky1m0qi7iyd204prk7glbvig1cvaab33k19f1";
+  };
+
+  propagatedBuildInputs = with python3Packages; [
+    pyvcf
+    python-Levenshtein
+    progressbar2
+    pysam
+    pyfaidx
+    intervaltree
+  ];
+
+  prePatch = ''
+    substituteInPlace ./setup.py \
+      --replace '"progressbar2==3.41.0",' "" \
+      --replace '"pysam==0.15.2",' ""
+  '';
+
+  meta = with lib; {
+    description = "Structural variant comparison tool for VCFs";
+    license = licenses.mit;
+    maintainers = with maintainers; [ scalavision ];
+    longDescription = ''
+      Truvari is a benchmarking tool for comparison sets of SVs.
+      It can calculate the recall, precision, and f-measure of a
+      vcf from a given structural variant caller. The tool
+      is created by Spiral Genetics.
+    '';
+  };
+}

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

@@ -0,0 +1,32 @@
+{ lib
+, buildPythonPackage
+, fetchFromGitHub
+, pytest
+}:
+
+buildPythonPackage rec {
+  pname = "PyVCF";
+  version = "0.6.8";
+
+  src = fetchFromGitHub {
+    owner = "jamescasbon";
+    repo = "PyVCF";
+    rev = "476169cd457ba0caa6b998b301a4d91e975251d9";
+    sha256 = "0qf9lwj7r2hjjp4bd4vc7nayrhblfm4qcqs4dbd43a6p4bj2jv5p";
+  };
+
+  checkInputs = [ pytest ];
+
+  meta = with lib; {
+    homepage = "https://pyvcf.readthedocs.io/en/latest/index.html";
+    description = "A VCF (Variant Call Format) Parser for Python, supporting version 4.0 and 4.1";
+    license = licenses.bsd3;
+    maintainers = with maintainers; [ scalavision ];
+    longDescription = ''
+      The intent of this module is to mimic the csv module in the Python stdlib, 
+      as opposed to more flexible serialization formats like JSON or YAML. 
+      vcf will attempt to parse the content of each record based on the data 
+      types specified in the meta-information lines
+    '';
+  };
+}

pkgs/top-level/all-packages.nix

@@ -23282,6 +23282,8 @@ in
 
   trimal = callPackage ../applications/science/biology/trimal { };
 
+  truvari = callPackage ../applications/science/biology/truvari { };
+
   varscan = callPackage ../applications/science/biology/varscan { };
 
   hmmer = callPackage ../applications/science/biology/hmmer { };

pkgs/top-level/python-packages.nix

@@ -1167,6 +1167,8 @@ in {
 
   pyvcd = callPackage ../development/python-modules/pyvcd { };
 
+  pyvcf = callPackage ../development/python-modules/pyvcf { };
+
   pyvoro = callPackage ../development/python-modules/pyvoro { };
 
   relatorio = callPackage ../development/python-modules/relatorio { };