From a7654deb7b973b69b68245524faa051fc256924b Mon Sep 17 00:00:00 2001
From: Matthew Pitkin <matthew.pitkin@ligo.org>
Date: Mon, 13 Aug 2018 11:24:39 +0100
Subject: [PATCH] Allow PyMC3 prior to be set through the Prior ln_prob method
- added this to PyMC3 example
---
...near_regression_pymc3_custom_likelihood.py | 28 +++++++++-
tupak/core/prior.py | 24 ---------
tupak/core/sampler.py | 53 +++++++++++++++++--
3 files changed, 76 insertions(+), 29 deletions(-)
diff --git a/examples/other_examples/linear_regression_pymc3_custom_likelihood.py b/examples/other_examples/linear_regression_pymc3_custom_likelihood.py
index 9535d58d9..3b3ab8af1 100644
--- a/examples/other_examples/linear_regression_pymc3_custom_likelihood.py
+++ b/examples/other_examples/linear_regression_pymc3_custom_likelihood.py
@@ -109,11 +109,37 @@ class GaussianLikelihoodPyMC3(tupak.Likelihood):
# the time, data and signal model
likelihood = GaussianLikelihoodPyMC3(time, data, sigma, model)
+
+# Define a custom prior for one of the parameter for use with PyMC3
+class PriorPyMC3(tupak.core.prior.Prior):
+ def __init__(self, minimum, maximum, name=None, latex_label=None):
+ """
+ Uniform prior with bounds (should be equivalent to tupak.prior.Uniform)
+ """
+
+ tupak.core.prior.Prior.__init__(self, name, latex_label,
+ minimum=minimum,
+ maximum=maximum)
+
+ def ln_prob(self, sampler=None):
+ """
+ Change ln_prob method to take in a Sampler and return a PyMC3
+ distribution.
+ """
+
+ from tupak.core.sampler import Pymc3
+
+ if not isinstance(sampler, Pymc3):
+ raise ValueError("Sampler is not a tupak Pymc3 sampler object")
+
+ return pm.Uniform(self.name, lower=self.minimum,
+ upper=self.maximum)
+
# From hereon, the syntax is exactly equivalent to other tupak examples
# We make a prior
priors = {}
priors['m'] = tupak.core.prior.Uniform(0, 5, 'm')
-priors['c'] = tupak.core.prior.Uniform(-2, 2, 'c')
+priors['c'] = PriorPyMC3(-2, 2, 'c')
# And run sampler
result = tupak.run_sampler(
diff --git a/tupak/core/prior.py b/tupak/core/prior.py
index 31f04c93f..0a4151601 100644
--- a/tupak/core/prior.py
+++ b/tupak/core/prior.py
@@ -467,30 +467,6 @@ class Prior(object):
label = self.name
return label
- def pymc3_prior(self, sampler):
- """
- 'pymc3_prior' A user defined PyMC3 prior.
-
- This should be overwritten by each subclass if needed.
-
- Parameters
- ----------
- sampler: `tupak.core.sampler.Sampler`
- A Sampler class
-
- Returns
- -------
- None
-
- """
-
- from tupak.core.sampler import Sampler
-
- if not isinstance(sampler, Sampler):
- raise ValueError("'sampler' is not a Sampler class")
-
- return None
-
class DeltaFunction(Prior):
diff --git a/tupak/core/sampler.py b/tupak/core/sampler.py
index 045d1f667..eca15e6c8 100644
--- a/tupak/core/sampler.py
+++ b/tupak/core/sampler.py
@@ -957,6 +957,46 @@ class Pymc3(Sampler):
"""
pass
+ def _initialise_parameters(self):
+ """
+ Change `_initialise_parameters()`, so that it does call the `sample`
+ method in the Prior class.
+
+ """
+
+ self.__search_parameter_keys = []
+ self.__fixed_parameter_keys = []
+
+ for key in self.priors:
+ if isinstance(self.priors[key], Prior) \
+ and self.priors[key].is_fixed is False:
+ self.__search_parameter_keys.append(key)
+ elif isinstance(self.priors[key], Prior) \
+ and self.priors[key].is_fixed is True:
+ self.__fixed_parameter_keys.append(key)
+
+ logger.info("Search parameters:")
+ for key in self.__search_parameter_keys:
+ logger.info(' {} = {}'.format(key, self.priors[key]))
+ for key in self.__fixed_parameter_keys:
+ logger.info(' {} = {}'.format(key, self.priors[key].peak))
+
+ def _initialise_result(self):
+ """
+ Initialise results within Pymc3 subclass.
+ """
+ result = Result()
+ result.sampler = self.__class__.__name__.lower()
+ result.search_parameter_keys = self.__search_parameter_keys
+ result.fixed_parameter_keys = self.__fixed_parameter_keys
+ result.parameter_labels = [
+ self.priors[k].latex_label for k in
+ self.__search_parameter_keys]
+ result.label = self.label
+ result.outdir = self.outdir
+ result.kwargs = self.kwargs
+ return result
+
@property
def kwargs(self):
""" Ensures that proper keyword arguments are used for the Pymc3 sampler.
@@ -1253,10 +1293,15 @@ class Pymc3(Sampler):
# set the parameter prior distributions (in the model context manager)
with self.pymc3_model:
for key in self.priors:
- # if the prior contains a pymc3_prior method use that otherwise try
- # and find the PyMC3 distribution
- if self.priors[key].pymc3_prior(self) is not None:
- self.pymc3_priors[key] = self.priors[key].pymc3_prior(self)
+ # if the prior contains ln_prob method that takes a 'sampler' argument
+ # then try using that
+ lnprobargs = inspect.getargspec(self.priors[key].ln_prob).args
+ if 'sampler' in lnprobargs:
+ try:
+ self.pymc3_priors[key] = self.priors[key].ln_prob(sampler=self)
+ except RuntimeError:
+ raise RuntimeError(("Problem setting PyMC3 prior for ",
+ "'{}'".format(key)))
else:
# use Prior distribution name
distname = self.priors[key].__class__.__name__
--
GitLab