Replace slow dist_math::incomplete_beta with aesara op

.github/workflows/pytest.yml

@@ -12,52 +12,78 @@ jobs:
         os: [ubuntu-18.04]
         floatx: [float32, float64]
         test-subset:
+        # Tests are split into multiple jobs to accelerate the CI.
+        # The first job (starting in the next block) shouldn't run any tests, but
+        # just ignores tests because that don't work at all, or run in other jobs.'
+        # Any test that was not ignored runs in the first job.
+        # A pre-commit hook (scripts/check_all_tests_are_covered.py) enforces that
+        # test run just once.
+
+        # Because YAML doesn't allow comments in the blocks below, here they are..
+        # 1st block: These tests are temporarily disabled, because they are _very_ broken
+        # 2nd block: The JAX tests run through their own workflow: jaxtests.yml
+        # 3nd & 4rd: These tests are covered by other matrix jobs
+        # 5th block: These tests PASS without a single XFAIL
+        # 6th block: These have some XFAILs
           - |
-            --ignore=pymc3/tests/test_dist_math.py
-            --ignore=pymc3/tests/test_distribution_defaults.py
-            --ignore=pymc3/tests/test_distributions.py
-            --ignore=pymc3/tests/test_distributions_random.py
             --ignore=pymc3/tests/test_distributions_timeseries.py
-            --ignore=pymc3/tests/test_examples.py
-            --ignore=pymc3/tests/test_gp.py
+            --ignore=pymc3/tests/test_missing.py
             --ignore=pymc3/tests/test_mixture.py
-            --ignore=pymc3/tests/test_ode.py
+            --ignore=pymc3/tests/test_model_graph.py
+            --ignore=pymc3/tests/test_modelcontext.py
             --ignore=pymc3/tests/test_parallel_sampling.py
-            --ignore=pymc3/tests/test_posteriors.py
-            --ignore=pymc3/tests/test_quadpotential.py
+            --ignore=pymc3/tests/test_profile.py
             --ignore=pymc3/tests/test_random.py
-            --ignore=pymc3/tests/test_sampling.py
-            --ignore=pymc3/tests/test_sampling_jax.py
-            --ignore=pymc3/tests/test_shape_handling.py
             --ignore=pymc3/tests/test_shared.py
             --ignore=pymc3/tests/test_smc.py
+            --ignore=pymc3/tests/test_starting.py
             --ignore=pymc3/tests/test_step.py
-            --ignore=pymc3/tests/test_updates.py
+            --ignore=pymc3/tests/test_tracetab.py
+            --ignore=pymc3/tests/test_tuning.py
+            --ignore=pymc3/tests/test_types.py
             --ignore=pymc3/tests/test_variational_inference.py
+            --ignore=pymc3/tests/test_sampling_jax.py
+            --ignore=pymc3/tests/test_dist_math.py
+            --ignore=pymc3/tests/test_minibatches.py
+            --ignore=pymc3/tests/test_pickling.py
+            --ignore=pymc3/tests/test_plots.py
+            --ignore=pymc3/tests/test_special_functions.py
+            --ignore=pymc3/tests/test_updates.py
+            --ignore=pymc3/tests/test_examples.py
+            --ignore=pymc3/tests/test_gp.py
+            --ignore=pymc3/tests/test_model.py
+            --ignore=pymc3/tests/test_model_func.py
+            --ignore=pymc3/tests/test_ode.py
+            --ignore=pymc3/tests/test_posdef_sym.py
+            --ignore=pymc3/tests/test_quadpotential.py
+            --ignore=pymc3/tests/test_shape_handling.py
+            --ignore=pymc3/tests/test_distributions.py
+            --ignore=pymc3/tests/test_distributions_random.py
+            --ignore=pymc3/tests/test_idata_conversion.py
+
           - |
+            pymc3/tests/test_modelcontext.py
             pymc3/tests/test_dist_math.py
-            pymc3/tests/test_distribution_defaults.py
-            pymc3/tests/test_distributions_random.py
-            pymc3/tests/test_parallel_sampling.py
-            pymc3/tests/test_random.py
-            pymc3/tests/test_shared.py
-            pymc3/tests/test_smc.py
+            pymc3/tests/test_minibatches.py
+            pymc3/tests/test_pickling.py
+            pymc3/tests/test_plots.py
+            pymc3/tests/test_special_functions.py
+            pymc3/tests/test_updates.py
+
           - |
+            pymc3/tests/test_idata_conversion.py
+            pymc3/tests/test_distributions.py
+            pymc3/tests/test_distributions_random.py
             pymc3/tests/test_examples.py
-            pymc3/tests/test_mixture.py
+            pymc3/tests/test_gp.py
+            pymc3/tests/test_model.py
+            pymc3/tests/test_model_func.py
             pymc3/tests/test_ode.py
-            pymc3/tests/test_posteriors.py
+            pymc3/tests/test_posdef_sym.py
             pymc3/tests/test_quadpotential.py
-          - |
-            pymc3/tests/test_distributions_timeseries.py
             pymc3/tests/test_shape_handling.py
             pymc3/tests/test_step.py
-            pymc3/tests/test_updates.py
-            pymc3/tests/test_variational_inference.py
-          - |
-            pymc3/tests/test_distributions.py
-            pymc3/tests/test_gp.py
-            pymc3/tests/test_sampling.py
+
       fail-fast: false
     runs-on: ${{ matrix.os }}
     env:

.github/workflows/windows.yml

@@ -7,6 +7,7 @@ on:
 
 jobs:
   pytest:
+    if: false
     strategy:
       matrix:
         os: [windows-latest]

RELEASE-NOTES.md

@@ -4,15 +4,21 @@
 ### Breaking Changes
 - ⚠ Theano-PyMC has been replaced with Aesara, so all external references to `theano`, `tt`, and `pymc3.theanof` need to be replaced with `aesara`, `at`, and `pymc3.aesaraf` (see [4471](https://github.com/pymc-devs/pymc3/pull/4471)).
 - ArviZ `plots` and `stats` *wrappers* were removed. The functions are now just available by their original names (see [#4549](https://github.com/pymc-devs/pymc3/pull/4471) and `3.11.2` release notes).
+- The GLM submodule has been removed, please use [Bambi](https://bambinos.github.io/bambi/) instead.
+- The `Distribution` keyword argument `testval` has been deprecated in favor of `initval`.
+- The `incomplete_beta` function in `pymc3.distributions.dist_math` was replaced by an equivalent faster vectorized `betainc` Aesara `Op` (see [4519](https://github.com/pymc-devs/pymc3/pull/4519)).
 - ...
 
 ### New Features
 - The `CAR` distribution has been added to allow for use of conditional autoregressions which often are used in spatial and network models.
+- Add `logcdf` method to Kumaraswamy distribution (see [#4706](https://github.com/pymc-devs/pymc3/pull/4706)).
 - ...
 
 ### Maintenance
 - Remove float128 dtype support (see [#4514](https://github.com/pymc-devs/pymc3/pull/4514)).
 - Logp method of `Uniform` and `DiscreteUniform` no longer depends on `pymc3.distributions.dist_math.bound` for proper evaluation (see [#4541](https://github.com/pymc-devs/pymc3/pull/4541)).
+- `Model.RV_dims` and `Model.coords` are now read-only properties. To modify the `coords` dictionary use `Model.add_coord`. Also `dims` or coordinate values that are `None` will be auto-completed (see [#4625](https://github.com/pymc-devs/pymc3/pull/4625)).
+- The length of `dims` in the model is now tracked symbolically through `Model.dim_lengths` (see [#4625](https://github.com/pymc-devs/pymc3/pull/4625)).
 - ...
 
 ## PyMC3 3.11.2 (14 March 2021)

docs/source/Gaussian_Processes.rst

@@ -158,7 +158,7 @@ other type of random variable.  The first argument is the name of the random
 variable representing the function we are placing the prior over.
 The second argument is the inputs to the function that the prior is over,
 :code:`X`.  The inputs are usually known and present in the data, but they can
-also be PyMC3 random variables.  If the inputs are a Aesara tensor or a
+also be PyMC3 random variables.  If the inputs are an Aesara tensor or a
 PyMC3 random variable, the :code:`shape` needs to be given.
 
 Usually at this point, inference is performed on the model.  The

docs/source/Probability_Distributions.rst

@@ -117,20 +117,15 @@ For example, the gamma distribution is positive-valued. If we define one for a m
     with pm.Model() as model:
         g = pm.Gamma('g', 1, 1)
 
-We notice a modified variable inside the model ``vars`` attribute, which holds the free variables in the model.
+We notice a modified variable inside the model ``value_vars`` attribute.  These variables represent the values of each random variable in the model's log-likelihood.
 
 ::
 
-    >>> model.vars
+    >>> model.value_vars
     [g_log__]
 
-As the name suggests, the variable ``g`` has been log-transformed, and this is the space over which sampling takes place.
+As the name suggests, the variable ``g`` has been log-transformed, and this is the space over which posterior sampling takes place.
 
-The original variable is simply treated as a deterministic variable, since the value of the transformed variable is simply back-transformed when a sample is drawn in order to recover the original variable. Hence, ``g`` resides in the ``model.deterministics`` list.
-
-::
-
-    >>> model.deterministics
-    [g]
+The value of the transformed variable is simply back-transformed when a sample is drawn in order to recover the original variable.
 
 By default, auto-transformed variables are ignored when summarizing and plotting model output.

docs/source/PyMC3_and_Aesara.rst

@@ -12,7 +12,7 @@ What is Aesara
 
 Aesara is a package that allows us to define functions involving array
 operations and linear algebra. When we define a PyMC3 model, we implicitly
-build up a Aesara function from the space of our parameters to
+build up an Aesara function from the space of our parameters to
 their posterior probability density up to a constant factor. We then use
 symbolic manipulations of this function to also get access to its gradient.
 
@@ -159,7 +159,7 @@ where with the normal likelihood :math:`N(x|μ,σ^2)`
 
 To build that function we need to keep track of two things: The parameter
 space (the *free variables*) and the logp function. For each free variable
-we generate a Aesara variable. And for each variable (observed or otherwise)
+we generate an Aesara variable. And for each variable (observed or otherwise)
 we add a term to the global logp. In the background something similar to
 this is happening::
 
@@ -177,7 +177,7 @@ So calling `pm.Normal()` modifies the model: It changes the logp function
 of the model. If the `observed` keyword isn't set it also creates a new
 free variable. In contrast, `pm.Normal.dist()` doesn't care about the model,
 it just creates an object that represents the normal distribution. Calling
-`logp` on this object creates a Aesara variable for the logp probability
+`logp` on this object creates an Aesara variable for the logp probability
 or log probability density of the distribution, but again without changing
 the model in any way.
 

docs/source/api/distributions/discrete.rst

@@ -15,10 +15,12 @@ Discrete
    ZeroInflatedNegativeBinomial
    DiscreteUniform
    Geometric
+   HyperGeometric
    Categorical
    DiscreteWeibull
    Constant
    OrderedLogistic
+   OrderedProbit
 
 .. automodule:: pymc3.distributions.discrete
    :members:

docs/source/api/distributions/utilities.rst

@@ -12,9 +12,6 @@ Distribution utility classes and functions
   DensityDist
   TensorType
 
-  draw_values
-  generate_samples
-
 
 .. autoclass:: Distribution
 .. autoclass:: Discrete
@@ -23,6 +20,3 @@ Distribution utility classes and functions
 .. autoclass:: DensityDist
     :members:
 .. autofunction:: TensorType
-
-.. autofunction:: draw_values
-.. autofunction:: generate_samples

docs/source/api/glm.rst

@@ -4,5 +4,7 @@ Generalized Linear Models
 
 .. currentmodule:: pymc3.glm.linear
 
-.. automodule:: pymc3.glm.linear
-   :members:
+Generalized Linear Models are delegated to the
+`Bambi <https://bambinos.github.io/bambi>`_.
+library, a high-level Bayesian model-building
+interface built on top of the PyMC3.

docs/source/api/math.rst

@@ -4,7 +4,7 @@ Math
 
 This submodule contains various mathematical functions. Most of them are imported directly
 from aesara.tensor (see there for more details). Doing any kind of math with PyMC3 random
-variables, or defining custom likelihoods or priors requires you to use these aesara
+variables, or defining custom likelihoods or priors requires you to use these Aesara
 expressions rather than NumPy or Python code.
 
 .. currentmodule:: pymc3.math

docs/source/api/variables.rst

@@ -6,22 +6,5 @@ Random Variables
 The normal PyMC3 programmer will typically not need to interact with these classes, except possibly when debugging.  Otherwise they are primarily of interest to developers.
 
 
-.. autoclass:: PyMC3Variable
-    :members:
-
-
 .. autoclass:: ValueGradFunction
     :members:
-
-
-.. autoclass:: FreeRV
-    :members:
-
-.. autoclass:: ObservedRV
-    :members:
-
-.. autoclass:: MultiObservedRV
-    :members:
-
-.. autoclass:: TransformedRV
-    :members: