got the logcdf working without refactoring so the example code works

Update CODE_OF_CONDUCT.md Changed contact email ChiSquared now returns a Gamma random variable (pymc-devs#7007) Update devcontainer (pymc-devs#7017) * Update dev container on release and schedule * Use latest version of upstream dev container image * Remove pre-commit cache hack * Add Jupyter extension to container Rename RVTransform to Transform Remove duplicate ChainTransform Move ZeroSumTransform methods inside respective class Remove deprecated function rvs_to_value_vars Merge functionality of pytensorf and logprob/utils Also fixes circular imports Deprecate unused function walk_model update theme (pymc-devs#7018) * update theme version * update conf.py Fix docs formatting in `shape_utils` (pymc-devs#7025) * Small docs update shape_utils.py Fix docs markup * Update to_tuple docs shape_utils.py * Types --------- Co-authored-by: Denis Kataev <[email protected]> Use PyTensor StudentT RV Update GOVERNANCE.md (pymc-devs#7031) * Update GOVERNANCE.md * Update GOVERNANCE.md Co-authored-by: Purna Chandra Mansingh <[email protected]> --------- Co-authored-by: Purna Chandra Mansingh <[email protected]> Remove deprecated model methods Deprecate Model.model property Deprecate pytensor_config Bump Pytensor dependency to >=2.18.1,<2.19 Better float32 sampling support for TruncatedNormal (pymc-devs#7026) * manually inv transform; force rng same type * always upcast f64 and downcast to dtype of param * add comment * use class attr dtype * need else stmt for dtype * actually no need to downcast in this method * rm unused import Another small fixes in docs and imports (pymc-devs#7030) * Remove old import from autosummary doc pages * Fix small error * Correct place for model_to_networkx & model_to_graphviz * Correct currentmodule source for pytensorf.rst * Update docs/source/api/model/core.rst Co-authored-by: Thomas Wiecki <[email protected]> * Fix sampling & jax docs imports --------- Co-authored-by: Denis Kataev <[email protected]> Co-authored-by: Thomas Wiecki <[email protected]> Update installation instructions to remove incompatibility of JAX on Windows
LukeLB · Nov 29, 2023 · bf619e0 · bf619e0
1 parent 44e1c5e
commit bf619e0
Show file tree

Hide file tree

Showing 49 changed files with 619 additions and 1,073 deletions.
diff --git a/.github/workflows/devcontainer-docker-image.yml b/.github/workflows/devcontainer-docker-image.yml
@@ -2,6 +2,10 @@ name: devcontainer-docker-image
 
 on:
   workflow_dispatch:
+  schedule:
+    - cron: "48 19 * * 5"  # Fridays at 19:48 UTC
+  release:
+    types: [published]
 
 env:
   REGISTRY: ghcr.io

diff --git a/.gitpod.yml b/.gitpod.yml
@@ -36,6 +36,7 @@ vscode:
     - eamodio.gitlens
     - ms-python.python
     - ms-pyright.pyright
+    - ms-toolsai.jupyter
     - donjayamanne.githistory
 
 github:

diff --git a/CODE_OF_CONDUCT.md b/CODE_OF_CONDUCT.md
@@ -56,7 +56,7 @@ further defined and clarified by project maintainers.
 
 Instances of abusive, harassing, or otherwise unacceptable behavior may be
 reported by contacting PyMC developer Christopher Fonnesbeck via email
-(chris.fonnesbeck@vanderbilt.edu) or phone (615-955-0380). Alternatively, you
+(fonnesbeck@gmail.com) or phone (615-955-0380). Alternatively, you
 may also contact NumFOCUS Executive Director Leah Silen (512-222-5449), as PyMC
 is a member of NumFOCUS and subscribes to their code of conduct as a
 precondition for continued membership. All complaints will be reviewed and

diff --git a/GOVERNANCE.md b/GOVERNANCE.md
@@ -191,6 +191,7 @@ Contributors don't need to be part of any dedicated team.
 * Michael Osthege (dev)
 * Oriol Abril-Pla (docs, community)
 * Osvaldo Martin (dev, docs)
+* Purna Chandra Mansingh (community)
 * Ravin Kumar (dev, community, docs)
 * Reshama Shaikh (community - PyMC Labs)
 * Ricardo Vieira (dev, community)

diff --git a/conda-envs/environment-dev.yml b/conda-envs/environment-dev.yml
@@ -14,7 +14,7 @@ dependencies:
 - numpy>=1.15.0
 - pandas>=0.24.0
 - pip
-- pytensor>=2.17.0,<2.18
+- pytensor>=2.18.1,<2.19
 - python-graphviz
 - networkx
 - scipy>=1.4.1

diff --git a/conda-envs/environment-docs.yml b/conda-envs/environment-docs.yml
@@ -12,7 +12,7 @@ dependencies:
 - numpy>=1.15.0
 - pandas>=0.24.0
 - pip
-- pytensor>=2.17.0,<2.18
+- pytensor>=2.18.1,<2.19
 - python-graphviz
 - scipy>=1.4.1
 - typing-extensions>=3.7.4
@@ -24,7 +24,7 @@ dependencies:
 - numpydoc
 - polyagamma
 - pre-commit>=2.8.0
-- pymc-sphinx-theme==0.13
+- pymc-sphinx-theme==0.14
 - sphinx-copybutton
 - sphinx-design
 - sphinx-notfound-page

diff --git a/conda-envs/environment-jax.yml b/conda-envs/environment-jax.yml
@@ -21,7 +21,7 @@ dependencies:
 - numpyro>=0.8.0
 - pandas>=0.24.0
 - pip
-- pytensor>=2.17.0,<2.18
+- pytensor>=2.18.1,<2.19
 - python-graphviz
 - networkx
 - scipy>=1.4.1

diff --git a/conda-envs/environment-test.yml b/conda-envs/environment-test.yml
@@ -17,7 +17,7 @@ dependencies:
 - numpy>=1.15.0
 - pandas>=0.24.0
 - pip
-- pytensor>=2.17.0,<2.18
+- pytensor>=2.18.1,<2.19
 - python-graphviz
 - networkx
 - scipy>=1.4.1

diff --git a/conda-envs/windows-environment-dev.yml b/conda-envs/windows-environment-dev.yml
@@ -14,7 +14,7 @@ dependencies:
 - numpy>=1.15.0
 - pandas>=0.24.0
 - pip
-- pytensor>=2.17.0,<2.18
+- pytensor>=2.18.1,<2.19
 - python-graphviz
 - networkx
 - scipy>=1.4.1

diff --git a/conda-envs/windows-environment-test.yml b/conda-envs/windows-environment-test.yml
@@ -17,7 +17,7 @@ dependencies:
 - numpy>=1.15.0
 - pandas>=0.24.0
 - pip
-- pytensor>=2.17.0,<2.18
+- pytensor>=2.18.1,<2.19
 - python-graphviz
 - networkx
 - scipy>=1.4.1

diff --git a/docs/source/api/math.rst b/docs/source/api/math.rst
@@ -67,11 +67,7 @@ Functions exposed in pymc.math
    sgn
    ceil
    floor
-   det
    matrix_inverse
-   extract_diag
-   matrix_dot
-   trace
    sigmoid
    logsumexp
    invlogit

diff --git a/docs/source/api/model/core.rst b/docs/source/api/model/core.rst
@@ -6,8 +6,6 @@ Model creation and inspection
    :toctree: generated/
 
    Model
-   model_to_graphviz
-   model_to_networkx
    modelcontext
 
 Others
@@ -22,3 +20,14 @@ Others
    set_data
    Point
    compile_fn
+
+
+Graph visualization
+-------------------
+
+.. currentmodule:: pymc.model_graph
+.. autosummary::
+   :toctree: generated/
+
+   model_to_networkx
+   model_to_graphviz
diff --git a/docs/source/api/pytensorf.rst b/docs/source/api/pytensorf.rst
@@ -1,7 +1,7 @@
 PyTensor utils
 **************
 
-.. currentmodule:: pymc
+.. currentmodule:: pymc.pytensorf
 
 .. autosummary::
    :toctree: generated/

diff --git a/docs/source/api/samplers.rst b/docs/source/api/samplers.rst
@@ -4,27 +4,39 @@ Samplers
 This submodule contains functions for MCMC and forward sampling.
 
 
-.. currentmodule:: pymc
+.. currentmodule:: pymc.sampling.forward
 
 .. autosummary::
    :toctree: generated/
 
-   sample
    sample_prior_predictive
    sample_posterior_predictive
-   sample_posterior_predictive_w
-   sampling.jax.sample_blackjax_nuts
-   sampling.jax.sample_numpyro_nuts
-   init_nuts
    draw
 
+
+.. currentmodule:: pymc.sampling.mcmc
+
+.. autosummary::
+   :toctree: generated/
+
+   sample
+   init_nuts
+
+.. currentmodule:: pymc.sampling.jax
+
+.. autosummary::
+   :toctree: generated/
+
+   sample_blackjax_nuts
+   sample_numpyro_nuts
+
+
 Step methods
 ************
 
-.. currentmodule:: pymc
-
 HMC family
 ----------
+.. currentmodule:: pymc.step_methods.hmc
 
 .. autosummary::
    :toctree: generated/
@@ -34,6 +46,7 @@ HMC family
 
 Metropolis family
 -----------------
+.. currentmodule:: pymc.step_methods
 
 .. autosummary::
     :toctree: generated/
@@ -53,6 +66,7 @@ Metropolis family
 
 Other step methods
 ------------------
+.. currentmodule:: pymc.step_methods
 
 .. autosummary::
    :toctree: generated/

diff --git a/docs/source/conf.py b/docs/source/conf.py
@@ -156,7 +156,7 @@
 myst_substitutions = {
     "version_slug": rtd_version,
 }
-myst_heading_anchors = None
+myst_heading_anchors = 0
 
 v3_example_tutorials = [
     "case_studies/BEST",

diff --git a/docs/source/installation.md b/docs/source/installation.md
@@ -31,8 +31,6 @@ Similarly, to use BlackJAX sampler instead:
 conda install blackjax
 ```
 
-Note that JAX is not directly supported on Windows systems at the moment.
-
 ## Nutpie sampling
 
 You can also enable sampling with [nutpie](https://github.com/pymc-devs/nutpie).
@@ -41,5 +39,3 @@ Nutpie uses numba as the compiler and a sampler written in Rust for faster perfo
 ```console
 conda install -c conda-forge nutpie
 ```
-
-Unlike JAX, nutpie is directly supported on Windows.
diff --git a/pymc/distributions/continuous.py b/pymc/distributions/continuous.py
@@ -38,7 +38,6 @@
     BetaRV,
     _gamma,
     cauchy,
-    chisquare,
     exponential,
     gumbel,
     halfcauchy,
@@ -49,6 +48,7 @@
     lognormal,
     normal,
     pareto,
+    t,
     triangular,
     uniform,
     vonmises,
@@ -571,11 +571,13 @@ def rng_fn(
         upper: Union[np.ndarray, float],
         size: Optional[Union[List[int], int]],
     ) -> np.ndarray:
+        # Upcast to float64. (Caller will downcast to desired dtype if needed)
+        #   (Work-around for https://github.com/scipy/scipy/issues/15928)
         return stats.truncnorm.rvs(
-            a=(lower - mu) / sigma,
-            b=(upper - mu) / sigma,
-            loc=mu,
-            scale=sigma,
+            a=((lower - mu) / sigma).astype("float64"),
+            b=((upper - mu) / sigma).astype("float64"),
+            loc=(mu).astype("float64"),
+            scale=(sigma).astype("float64"),
             size=size,
             random_state=rng,
         )
@@ -744,6 +746,42 @@ def logp(value, mu, sigma, lower, upper):
 
         return logp
 
+    def logcdf(value, mu, sigma, lower, upper):
+        logcdf = normal_lccdf(mu, sigma, value)
+        lower_logcdf = normal_lccdf(mu, sigma, lower)
+        upper_logcdf = normal_lccdf(mu, sigma, upper)
+
+        is_lower_bounded = not (
+            isinstance(lower, TensorConstant) and np.all(np.isneginf(lower.value))
+        )
+        is_upper_bounded = not (isinstance(upper, TensorConstant) and np.all(np.isinf(upper.value)))
+
+        lognorm = 0
+        if is_lower_bounded and is_upper_bounded:
+            lognorm = logdiffexp(upper_logcdf, lower_logcdf)
+        elif is_lower_bounded:
+            lognorm = pt.log1mexp(lower_logcdf)
+        elif is_upper_bounded:
+            lognorm = upper_logcdf
+
+        logcdf_numerator = logdiffexp(logcdf, lower_logcdf) if is_lower_bounded else logcdf
+        logcdf_trunc = logcdf_numerator - lognorm
+
+        if is_lower_bounded:
+            logcdf_trunc = pt.switch(value < lower, -np.inf, logcdf_trunc)
+
+        if is_upper_bounded:
+            logcdf_trunc = pt.switch(value <= upper, logcdf_trunc, 0.0)
+
+        if is_lower_bounded and is_upper_bounded:
+            logcdf_trunc = check_parameters(
+                logcdf_trunc,
+                pt.le(lower, upper),
+                msg="lower_bound <= upper_bound",
+            )
+
+        return logcdf_trunc
+
 
 @_default_transform.register(TruncatedNormal)
 def truncated_normal_default_transform(op, rv):
@@ -1744,21 +1782,6 @@ def icdf(value, mu, sigma):
 Lognormal = LogNormal
 
 
-class StudentTRV(RandomVariable):
-    name = "studentt"
-    ndim_supp = 0
-    ndims_params = [0, 0, 0]
-    dtype = "floatX"
-    _print_name = ("StudentT", "\\operatorname{StudentT}")
-
-    @classmethod
-    def rng_fn(cls, rng, nu, mu, sigma, size=None) -> np.ndarray:
-        return np.asarray(stats.t.rvs(nu, mu, sigma, size=size, random_state=rng))
-
-
-studentt = StudentTRV()
-
-
 class StudentT(Continuous):
     r"""
     Student's T log-likelihood.
@@ -1823,7 +1846,7 @@ class StudentT(Continuous):
         with pm.Model():
             x = pm.StudentT('x', nu=15, mu=0, lam=1/23)
     """
-    rv_op = studentt
+    rv_op = t
 
     @classmethod
     def dist(cls, nu, mu=0, *, sigma=None, lam=None, **kwargs):
@@ -2384,16 +2407,21 @@ def logcdf(value, alpha, beta):
         )
 
 
-class ChiSquared(PositiveContinuous):
+class ChiSquared:
     r"""
     :math:`\chi^2` log-likelihood.
 
+    This is the distribution from the sum of the squares of :math:`\nu` independent standard normal random variables or a special
+    case of the gamma distribution with :math:`\alpha = \nu/2` and :math:`\beta = 1/2`.
+
     The pdf of this distribution is
 
     .. math::
 
        f(x \mid \nu) = \frac{x^{(\nu-2)/2}e^{-x/2}}{2^{\nu/2}\Gamma(\nu/2)}
 
+    Read more about the :math:`\chi^2` distribution at https://en.wikipedia.org/wiki/Chi-squared_distribution
+
     .. plot::
         :context: close-figs
 
@@ -2423,24 +2451,13 @@ class ChiSquared(PositiveContinuous):
     nu : tensor_like of float
         Degrees of freedom (nu > 0).
     """
-    rv_op = chisquare
-
-    @classmethod
-    def dist(cls, nu, *args, **kwargs):
-        nu = pt.as_tensor_variable(floatX(nu))
-        return super().dist([nu], *args, **kwargs)
 
-    def moment(rv, size, nu):
-        moment = nu
-        if not rv_size_is_none(size):
-            moment = pt.full(size, moment)
-        return moment
+    def __new__(cls, name, nu, **kwargs):
+        return Gamma(name, alpha=nu / 2, beta=1 / 2, **kwargs)
 
-    def logp(value, nu):
-        return _logprob_helper(Gamma.dist(alpha=nu / 2, beta=0.5), value)
-
-    def logcdf(value, nu):
-        return _logcdf_helper(Gamma.dist(alpha=nu / 2, beta=0.5), value)
+    @classmethod
+    def dist(cls, nu, **kwargs):
+        return Gamma.dist(alpha=nu / 2, beta=1 / 2, **kwargs)
 
 
 # TODO: Remove this once logp for multiplication is working!