DM-48642: Use private RNG rather than affecting global rng state

.github/workflows/test_treegp.yaml

@@ -14,7 +14,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [ 3.8, 3.9, "3.10", "3.11" ]
+        python-version: [ 3.8, 3.9, "3.10", "3.11", "3.12" ]
 
     steps:
       - uses: actions/checkout@v2
@@ -26,7 +26,7 @@ jobs:
 
       - name: Setup environment
         run: |
-          sudo apt-get install -y libffi-dev
+          sudo apt-get install -y libffi-dev libbz2-dev
 
       - name: Install dependencies
         run: |
@@ -50,9 +50,8 @@ jobs:
           pytest --cov=treegp --cov-report=xml --cov-fail-under=80
 
       - name: Upload coverage to Codecov
-        uses: codecov/codecov-action@v2
+        uses: codecov/codecov-action@v5
         with:
-          file: ./tests/coverage.xml
+          files: ./tests/coverage.xml
           fail_ci_if_error: true  # Optional: specify if CI should fail when Codecov upload fails
-        env:
-          CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
+          token: ${{ secrets.CODECOV_TOKEN }}

requirements.txt

@@ -2,8 +2,8 @@ setuptools>=65.5.1
 numpy>=1.16.5
 scipy
 matplotlib
-iminuit>1.3,<1.4; python_version < '3.0'  # iminuit 1.4 fails on python 2.7.
-iminuit>2;        python_version >= '3.0' # iminuit changed API in v2.0.  We use the 2.x API.
+
+iminuit>2  # iminuit changed API in v2.0.  We use the 2.x API.
 treecorr>=4.2
 fitsio>=0.9.12
 scikit-learn>=0.18

treegp/two_pcf.py

@@ -220,6 +220,7 @@ class two_pcf(object):
                         fiting part of the GP. (Boolean)
     :param robust_fit:  Used minuit to fit hyperparameter. Works only
                         anisotropic is True. (Boolean)
+    :param seed:        Seed to use for random number generator.
     """
 
     def __init__(
@@ -233,6 +234,7 @@ def __init__(
         anisotropic=False,
         robust_fit=False,
         p0=[3000.0, 0.0, 0.0],
+        seed=610639139,
     ):
         self.ndim = np.shape(X)[1]
         if self.ndim not in [1, 2]:
@@ -252,14 +254,22 @@ def __init__(
         self.anisotropic = anisotropic
         self.robust_fit = robust_fit
         self.p0_robust_fit = p0
+        self.seed = seed
+        self._rng = None
+
+    @property
+    def rng(self):
+        if self._rng is None:
+            self._rng = np.random.default_rng(self.seed)
+        return self._rng
 
     def resample_bootstrap(self):
         """
         Make a single bootstrap resampling on data.
         """
         npsfs = len(self.y)
 
-        ind_object = np.random.randint(0, npsfs - 1, size=npsfs)
+        ind_object = self.rng.integers(0, npsfs - 1, size=npsfs)
         u_ressample = self.X[:, 0][ind_object]
         v_ressample = self.X[:, 1][ind_object]
         y_ressample = self.y[ind_object]
@@ -332,7 +342,8 @@ def comp_xi_covariance(self, n_bootstrap=1000, mask=None, seed=610639139):
 
         :param seed: seed of the random generator.
         """
-        np.random.seed(seed)
+        self.seed = seed
+        self._rng = None
         xi_bootstrap = []
         for i in range(n_bootstrap):
             u, v, y, y_err = self.resample_bootstrap()