Merge pull request #88 from flatironinstitute/jax_key_annotate

updates PRNGKey to remove deprecation warning

Author: BalzaniEdoardo

docs/examples/plot_04_glm_demo.py

@@ -215,7 +215,7 @@
 # with the same number of neurons and features (mandatory)
 Xnew = np.random.normal(size=(20, ) + X.shape[1:])
 # generate a random key given a seed
-random_key = jax.random.PRNGKey(123)
+random_key = jax.random.key(123)
 spikes, rates = model.simulate(random_key, Xnew)
 
 plt.figure()
@@ -343,7 +343,7 @@
 # call simulate, with both the recurrent coupling
 # and the input
 spikes, rates = model.simulate_recurrent(
-    jax.random.PRNGKey(123),
+    jax.random.key(123),
     feedforward_input=feedforward_input,
     coupling_basis_matrix=coupling_basis,
     init_y=init_spikes

src/nemos/base_class.py

@@ -205,7 +205,7 @@ def score(
     @abc.abstractmethod
     def simulate(
         self,
-        random_key: jax.random.PRNGKeyArray,
+        random_key: jax.Array,
         feed_forward_input: DESIGN_INPUT_TYPE,
     ):
         """Simulate neural activity in response to a feed-forward input and recurrent activity."""

src/nemos/glm.py

@@ -393,15 +393,15 @@ def fit(
 
     def simulate(
         self,
-        random_key: jax.random.PRNGKeyArray,
+        random_key: jax.Array,
         feedforward_input: DESIGN_INPUT_TYPE,
     ) -> Tuple[jnp.ndarray, jnp.ndarray]:
         """Simulate neural activity in response to a feed-forward input.
 
         Parameters
         ----------
         random_key :
-            PRNGKey for seeding the simulation.
+            jax.random.key for seeding the simulation.
         feedforward_input :
             External input matrix to the model, representing factors like convolved currents,
             light intensities, etc. When not provided, the simulation is done with coupling-only.
@@ -485,7 +485,7 @@ def __init__(
 
     def simulate_recurrent(
         self,
-        random_key: jax.random.PRNGKeyArray,
+        random_key: jax.Array,
         feedforward_input: Union[NDArray, jnp.ndarray],
         coupling_basis_matrix: Union[NDArray, jnp.ndarray],
         init_y: Union[NDArray, jnp.ndarray],
@@ -501,7 +501,7 @@ def simulate_recurrent(
         Parameters
         ----------
         random_key :
-            PRNGKey for seeding the simulation.
+            jax.random.key for seeding the simulation.
         feedforward_input :
             External input matrix to the model, representing factors like convolved currents,
             light intensities, etc. When not provided, the simulation is done with coupling-only.
@@ -586,7 +586,7 @@ def simulate_recurrent(
         )
 
         def scan_fn(
-            data: Tuple[jnp.ndarray, int], key: jax.random.PRNGKeyArray
+            data: Tuple[jnp.ndarray, int], key: jax.Array
         ) -> Tuple[Tuple[jnp.ndarray, int], Tuple[jnp.ndarray, jnp.ndarray]]:
             """Scan over time steps and simulate activity and rates.
 

src/nemos/observation_models.py

@@ -9,8 +9,6 @@
 from . import utils
 from .base_class import Base
 
-KeyArray = Union[jnp.ndarray, jax.random.PRNGKeyArray]
-
 __all__ = ["PoissonObservations"]
 
 
@@ -138,7 +136,7 @@ def negative_log_likelihood(self, predicted_rate, y):
 
     @abc.abstractmethod
     def sample_generator(
-        self, key: KeyArray, predicted_rate: jnp.ndarray
+        self, key: jax.Array, predicted_rate: jnp.ndarray
     ) -> jnp.ndarray:
         """
         Sample from the estimated distribution.
@@ -399,7 +397,7 @@ def negative_log_likelihood(
         return jnp.mean(predicted_rate - x)
 
     def sample_generator(
-        self, key: KeyArray, predicted_rate: jnp.ndarray
+        self, key: jax.Array, predicted_rate: jnp.ndarray
     ) -> jnp.ndarray:
         """
         Sample from the Poisson distribution.
@@ -538,7 +536,7 @@ def check_observation_model(observation_model):
             "test_scalar_func": True,
         },
         "sample_generator": {
-            "input": [jax.random.PRNGKey(123), 0.5 * jnp.array([1.0, 1.0, 1.0])],
+            "input": [jax.random.key(123), 0.5 * jnp.array([1.0, 1.0, 1.0])],
             "test_preserve_shape": True,
         },
     }

tests/conftest.py

@@ -82,7 +82,7 @@ def poissonGLM_coupled_model_config_simulate():
             - coupling_basis (jax.numpy.ndarray): Coupling basis values from the config.
             - feedforward_input (jax.numpy.ndarray): Feedforward input values from the config.
             - init_spikes (jax.numpy.ndarray): Initial spike values from the config.
-            - jax.random.PRNGKey(123) (jax.random.PRNGKey): A pseudo-random number generator key.
+            - jax.random.key(123) (jax.Array): A pseudo-random number generator key.
     """
     observations = nmo.observation_models.PoissonObservations(jnp.exp)
     regularizer = nmo.regularizer.Ridge("BFGS", regularizer_strength=0.1)
@@ -119,7 +119,7 @@ def poissonGLM_coupled_model_config_simulate():
         coupling_basis,
         feedforward_input,
         init_spikes,
-        jax.random.PRNGKey(123),
+        jax.random.key(123),
     )
 
 

tests/test_base_class.py

@@ -42,7 +42,7 @@ def score(
 
     def simulate(
         self,
-        random_key: jax.random.PRNGKeyArray,
+        random_key: jax.Array,
         feed_forward_input: Union[NDArray, jnp.ndarray],
         **kwargs,
     ):

tests/test_glm.py

@@ -1038,14 +1038,14 @@ def test_simulate_feedforward_GLM_not_fit(self, poissonGLM_model_instantiation):
         with pytest.raises(
             nmo.exceptions.NotFittedError, match="This GLM instance is not fitted yet"
         ):
-            model.simulate(jax.random.PRNGKey(123), X)
+            model.simulate(jax.random.key(123), X)
 
     def test_simulate_feedforward_GLM(self, poissonGLM_model_instantiation):
         """Test that simulate goes through"""
         X, y, model, params, rate = poissonGLM_model_instantiation
         model.coef_ = params[0]
         model.intercept_ = params[1]
-        ysim, ratesim = model.simulate(jax.random.PRNGKey(123), X)
+        ysim, ratesim = model.simulate(jax.random.key(123), X)
         # check that the expected dimensionality is returned
         assert ysim.ndim == 2
         assert ratesim.ndim == 2

tests/test_observation_models.py

@@ -160,7 +160,7 @@ def test_emission_probability(selfself, poissonGLM_model_instantiation):
         Check that the emission probability is set to jax.random.poisson.
         """
         _, _, model, _, _ = poissonGLM_model_instantiation
-        key_array = jax.random.PRNGKey(123)
+        key_array = jax.random.key(123)
         counts = model.observation_model.sample_generator(key_array, np.arange(1, 11))
         if not jnp.all(counts == jax.random.poisson(key_array, np.arange(1, 11))):
             raise ValueError(