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

   1 {
   2   lib,
   3   buildPythonPackage,
   4   fetchFromGitHub,
   5
   6   # dependencies
   7   networkx,
   8   numpy,
   9   scipy,
  10   scikit-learn,
  11   pandas,
  12   pyparsing,
  13   torch,
  14   statsmodels,
  15   tqdm,
  16   joblib,
  17   opt-einsum,
  18   xgboost,
  19   google-generativeai,
  20
  21   # tests
  22   pytestCheckHook,
  23   pytest-cov,
  24   coverage,
  25   mock,
  26   black,
  27 }:
  28 buildPythonPackage rec {
  29   pname = "pgmpy";
  30   version = "0.1.26";
  31   pyproject = true;
  32
  33   src = fetchFromGitHub {
  34     owner = "pgmpy";
  35     repo = "pgmpy";
  36     rev = "refs/tags/v${version}";
  37     hash = "sha256-RusVREhEXYaJuQXTaCQ7EJgbo4+wLB3wXXCAc3sBGtU=";
  38   };
  39
  40   dependencies = [
  41     networkx
  42     numpy
  43     scipy
  44     scikit-learn
  45     pandas
  46     pyparsing
  47     torch
  48     statsmodels
  49     tqdm
  50     joblib
  51     opt-einsum
  52     xgboost
  53     google-generativeai
  54   ];
  55
  56   disabledTests = [
  57     # flaky:
  58     # AssertionError: -45.78899127622197 != -45.788991276221964
  59     "test_score"
  60
  61     # self.assertTrue(np.isclose(coef, dep_coefs[i], atol=1e-4))
  62     # AssertionError: False is not true
  63     "test_pillai"
  64
  65     # requires optional dependency daft
  66     "test_to_daft"
  67   ];
  68
  69   nativeCheckInputs = [
  70     pytestCheckHook
  71     # xdoctest
  72     pytest-cov
  73     coverage
  74     mock
  75     black
  76   ];
  77
  78   meta = {
  79     description = "Python Library for learning (Structure and Parameter), inference (Probabilistic and Causal), and simulations in Bayesian Networks";
  80     homepage = "https://github.com/pgmpy/pgmpy";
  81     changelog = "https://github.com/pgmpy/pgmpy/releases/tag/v${version}";
  82     license = lib.licenses.mit;
  83     maintainers = with lib.maintainers; [ happysalada ];
  84   };
  85 }