Merge pull request yt-project#4733 from neutrinoceros/ruff-npy

RFC: refactor usage of numpy.random legacy API

doc/source/examining/Loading_Generic_Particle_Data.ipynb

@@ -21,9 +21,10 @@
    "source": [
     "import numpy as np\n",
     "\n",
-    "n_particles = 5000000\n",
+    "rng = np.random.default_rng()\n",
+    "n_particles = 5_000_000\n",
     "\n",
-    "ppx, ppy, ppz = 1e6 * np.random.normal(size=[3, n_particles])\n",
+    "ppx, ppy, ppz = 1e6 * rng.normal(size=(3, n_particles))\n",
     "\n",
     "ppm = np.ones(n_particles)"
    ]
@@ -146,19 +147,19 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "n_gas_particles = 1000000\n",
-    "n_star_particles = 1000000\n",
-    "n_dm_particles = 2000000\n",
+    "n_gas_particles = 1_000_000\n",
+    "n_star_particles = 1_000_000\n",
+    "n_dm_particles = 2_000_000\n",
     "\n",
-    "ppxg, ppyg, ppzg = 1e6 * np.random.normal(size=[3, n_gas_particles])\n",
+    "ppxg, ppyg, ppzg = 1e6 * rng.normal(size=(3, n_gas_particles))\n",
     "ppmg = np.ones(n_gas_particles)\n",
     "hsml = 10000 * np.ones(n_gas_particles)\n",
     "dens = 2.0e-4 * np.ones(n_gas_particles)\n",
     "\n",
-    "ppxd, ppyd, ppzd = 1e6 * np.random.normal(size=[3, n_dm_particles])\n",
+    "ppxd, ppyd, ppzd = 1e6 * rng.normal(size=(3, n_dm_particles))\n",
     "ppmd = np.ones(n_dm_particles)\n",
     "\n",
-    "ppxs, ppys, ppzs = 5e5 * np.random.normal(size=[3, n_star_particles])\n",
+    "ppxs, ppys, ppzs = 5e5 * rng.normal(size=(3, n_star_particles))\n",
     "ppms = 0.1 * np.ones(n_star_particles)\n",
     "\n",
     "bbox = 1.1 * np.array(\n",

pyproject.toml

@@ -289,13 +289,16 @@ select = [
     "YTT",  # flake8-2020
     "UP",   # pyupgrade
     "I",    # isort
+    "NPY",  # numpy specific rules
 ]
 ignore = [
     "E501",  # line too long
     "E741",  # Do not use variables named 'I', 'O', or 'l'
     "B018",  # Found useless expression. # disabled because ds.index is idiomatic
 ]
 
+[tool.ruff.per-file-ignores]
+"test_*" = ["NPY002"]
 
 [tool.ruff.isort]
 combine-as-imports = true

yt/fields/field_detector.py

@@ -199,7 +199,8 @@ def deposit(self, *args, **kwargs):
     def mesh_sampling_particle_field(self, *args, **kwargs):
         pos = args[0]
         npart = len(pos)
-        return np.random.rand(npart)
+        rng = np.random.default_rng()
+        return rng.random(npart)
 
     def smooth(self, *args, **kwargs):
         tr = np.random.random((self.nd, self.nd, self.nd))

yt/frontends/enzo/data_structures.py

@@ -441,7 +441,8 @@ def _detect_output_fields(self):
 
     def _generate_random_grids(self):
         if self.num_grids > 40:
-            starter = np.random.randint(0, 20)
+            rng = np.random.default_rng()
+            starter = rng.integers(0, 20)
             random_sample = np.mgrid[starter : len(self.grids) - 1 : 20j].astype(
                 "int32"
             )
@@ -608,7 +609,8 @@ def _generate_random_grids(self):
         my_rank = self.comm.rank
         my_grids = self.grids[self.grid_procs.ravel() == my_rank]
         if len(my_grids) > 40:
-            starter = np.random.randint(0, 20)
+            rng = np.random.default_rng()
+            starter = rng.integers(0, 20)
             random_sample = np.mgrid[starter : len(my_grids) - 1 : 20j].astype("int32")
             mylog.debug("Checking grids: %s", random_sample.tolist())
         else:

yt/frontends/gadget/testing.py

@@ -95,8 +95,9 @@ def fake_gadget_binary(
             dtype = endian + dtype
             # Generate and write field block
             data = []
+            rng = np.random.default_rng()
             for pt in ptype:
-                data += [np.random.rand(npart[pt], dim)]
+                data += [rng.random((npart[pt], dim))]
             data = np.concatenate(data).astype(dtype)
             if field in block_ids:
                 block_id = block_ids[field]

yt/testing.py

@@ -501,7 +501,7 @@ def small_fake_hexahedral_ds():
 def fake_stretched_ds(N=16):
     from yt.loaders import load_uniform_grid
 
-    np.random.seed(0x4D3D3D3)
+    np.random.RandomState().seed(0x4D3D3D3)
     data = {"density": np.random.random((N, N, N))}
 
     cell_widths = []

yt/utilities/particle_generator.py

@@ -405,10 +405,11 @@ def __init__(
 
         pbar = get_pbar("Generating Particles", num_particles)
         tot_num_accepted = 0
+        rng = np.random.default_rng()
 
         while num_particles_left > 0:
-            m = np.random.uniform(high=1.01 * max_mass, size=num_particles_left)
-            idxs = np.random.randint(low=0, high=num_cells, size=num_particles_left)
+            m = rng.uniform(high=1.01 * max_mass, size=num_particles_left)
+            idxs = rng.integers(low=0, high=num_cells, size=num_particles_left)
             m_true = (
                 data_source[density_field] * data_source[("gas", "cell_volume")]
             ).flat[idxs]
@@ -418,17 +419,17 @@ def __init__(
 
             xpos = (
                 data_source[("index", "x")].flat[accepted_idxs]
-                + np.random.uniform(low=-0.5, high=0.5, size=num_accepted)
+                + rng.uniform(low=-0.5, high=0.5, size=num_accepted)
                 * data_source[("index", "dx")].flat[accepted_idxs]
             )
             ypos = (
                 data_source[("index", "y")].flat[accepted_idxs]
-                + np.random.uniform(low=-0.5, high=0.5, size=num_accepted)
+                + rng.uniform(low=-0.5, high=0.5, size=num_accepted)
                 * data_source[("index", "dy")].flat[accepted_idxs]
             )
             zpos = (
                 data_source[("index", "z")].flat[accepted_idxs]
-                + np.random.uniform(low=-0.5, high=0.5, size=num_accepted)
+                + rng.uniform(low=-0.5, high=0.5, size=num_accepted)
                 * data_source[("index", "dz")].flat[accepted_idxs]
             )
 

yt/visualization/fixed_resolution_filters.py

@@ -123,4 +123,5 @@ def apply(self, buff):
         else:
             amp = self.bg_lvl
         npm, nqm = np.shape(buff)
-        return buff + np.random.randn(npm, nqm) * amp
+        rng = np.random.default_rng()
+        return buff + rng.standard_normal((npm, nqm)) * amp

yt/visualization/volume_rendering/camera_path.py

@@ -118,7 +118,8 @@ def setup_tsp(self, niter=50000, init_temp=10.0, alpha=0.999, fixed_start=False)
         """
         # randomize tour
         self.tour = list(range(self.nframes))
-        np.random.shuffle(self.tour)
+        rng = np.random.default_rng()
+        rng.shuffle(self.tour)
         if fixed_start:
             first = self.tour.index(0)
             self.tour[0], self.tour[first] = self.tour[first], self.tour[0]