CUQIpy 1.1.1.post0.dev78__tar.gz → 1.1.1.post0.dev88__tar.gz

{cuqipy-1.1.1.post0.dev78 → cuqipy-1.1.1.post0.dev88}/CUQIpy.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: CUQIpy
-Version: 1.1.1.post0.dev78
+Version: 1.1.1.post0.dev88
 Summary: Computational Uncertainty Quantification for Inverse problems in Python
 Maintainer-email: "Nicolai A. B. Riis" <nabr@dtu.dk>, "Jakob S. Jørgensen" <jakj@dtu.dk>, "Amal M. Alghamdi" <amaal@dtu.dk>, Chao Zhang <chaz@dtu.dk>
 License:                                  Apache License

{cuqipy-1.1.1.post0.dev78 → cuqipy-1.1.1.post0.dev88}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: CUQIpy
-Version: 1.1.1.post0.dev78
+Version: 1.1.1.post0.dev88
 Summary: Computational Uncertainty Quantification for Inverse problems in Python
 Maintainer-email: "Nicolai A. B. Riis" <nabr@dtu.dk>, "Jakob S. Jørgensen" <jakj@dtu.dk>, "Amal M. Alghamdi" <amaal@dtu.dk>, Chao Zhang <chaz@dtu.dk>
 License:                                  Apache License

{cuqipy-1.1.1.post0.dev78 → cuqipy-1.1.1.post0.dev88}/cuqi/_version.py

@@ -8,11 +8,11 @@ import json
 
 version_json = '''
 {
- "date": "2024-09-11T08:37:02+0200",
+ "date": "2024-09-12T22:06:24+0200",
  "dirty": false,
  "error": null,
- "full-revisionid": "e332fa9b589e70ff3171d724fcc381c556631eb4",
- "version": "1.1.1.post0.dev78"
+ "full-revisionid": "f2fb34068f8c84bf37b3e4049a8a77391d3c432e",
+ "version": "1.1.1.post0.dev88"
 }
 '''  # END VERSION_JSON
 

cuqipy-1.1.1.post0.dev88/cuqi/experimental/mcmc/__init__.py

@@ -0,0 +1,123 @@
+"""
+Re-implementation of sampler module in a more object-oriented way.
+
+Main changes for users
+----------------------
+
+1. Sampling API
+   ^^^^^^^^^^^^
+
+   Previously one would call the `.sample` or `sample_adapt` methods of a sampler instance at :py:mod:`cuqi.sampler` to sample from a target distribution and store the samples as the output as follows:
+
+   .. code-block:: python
+
+      from cuqi.sampler import MH
+      from cuqi.distribution import DistributionGallery
+
+      # Target distribution
+      target = DistributionGallery("donut")
+
+      # Set up sampler
+      sampler = MH(target)
+
+      # Sample from the target distribution (Alternatively calling sample with explicit scale parameter set in sampler)
+      samples = sampler.sample_adapt(Ns=100, Nb=100) # Burn-in (Nb) removed by default
+
+   This has now changed to to a more object-oriented API which provides more flexibility and control over the sampling process.
+
+   For example one can now more explicitly control when the sampler is tuned (warmup) and when it is sampling with fixed parameters.
+
+   .. code-block:: python
+
+      from cuqi.experimental.mcmc import MH
+      from cuqi.distribution import DistributionGallery
+
+      # Target distribution
+      target = DistributionGallery("donut")
+
+      # Set up sampler
+      sampler = MH(target)
+
+      # Sample from the target distribution
+      sampler.warmup(Nb=100)  # Explicit warmup (tuning) of sampler
+      sampler.sample(Ns=100)  # Sampling with fixed parameters
+      samples = sampler.get_samples().burnthin(Nb=100) # Getting samples and removing burn-in from warmup
+
+   Importantly, the removal of burn-in from e.g. warmup is now a separate step that is done after the sampling process is complete.
+
+2. Sampling API for BayesianProblem
+   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+   :py:class:`cuqi.problem.BayesianProblem` continues to have the same API for `sample_posterior` and the `UQ` method.
+
+   There is now a flag `experimental` that can be set to `True` to use the new MCMC samplers.
+   
+   By default, the flag is set to `False` and the old samplers are used.
+
+   For this more high-level interface, burn-in is automatically removed from the samples as was the case before.
+
+
+3. More options for Gibbs sampling
+   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+   There are now more options for Gibbs sampling. Previously it was only possible to sample with Gibbs for samplers  :py:class:`cuqi.sampler.LinearRTO`, :py:class:`cuqi.sampler.RegularizedLinearRTO`, :py:class:`cuqi.sampler.Conjugate`, and :py:class:`cuqi.sampler.ConjugateApprox`.
+
+   Now, it is possible to define a Gibbs sampling scheme using any sampler from the :py:mod:`cuqi.experimental.mcmc` module.
+
+   **Example using a NUTS-within-Gibbs scheme for a 1D deconvolution problem:**
+
+   .. code-block:: python
+
+      import cuqi
+      import numpy as np
+      from cuqi.distribution import Gamma, Gaussian, GMRF, JointDistribution
+      from cuqi.experimental.mcmc import NUTS, HybridGibbs, Conjugate
+      from cuqi.testproblem import Deconvolution1D
+
+      # Forward problem
+      A, y_data, info = Deconvolution1D(dim=128, phantom='sinc', noise_std=0.001).get_components()
+
+      # Bayesian Inverse Problem
+      s = Gamma(1, 1e-4)
+      x = GMRF(np.zeros(A.domain_dim), 50)
+      y = Gaussian(A @ x, lambda s: 1 / s)
+
+      # Posterior
+      target = JointDistribution(y, x, s)(y=y_data)
+
+      # Gibbs sampling strategy. Note we can define initial_points and various parameters for each sampler
+      sampling_strategy = {
+          "x": NUTS(max_depth=10, initial_point=np.zeros(A.domain_dim)),
+          "s": Conjugate()
+      }
+
+      # Here we do 10 internal steps with NUTS for each Gibbs step
+      num_sampling_steps = {
+          "x": 10,
+          "s": 1
+      }
+
+      sampler = HybridGibbs(target, sampling_strategy, num_sampling_steps)
+
+      sampler.warmup(50)
+      sampler.sample(200)
+      samples = sampler.get_samples().burnthin(Nb=50)
+
+      samples["x"].plot_ci(exact=info.exactSolution)
+"""
+
+
+
+from ._sampler import Sampler, ProposalBasedSampler
+from ._langevin_algorithm import ULA, MALA
+from ._mh import MH
+from ._pcn import PCN
+from ._rto import LinearRTO, RegularizedLinearRTO
+from ._cwmh import CWMH
+from ._laplace_approximation import UGLA
+from ._hmc import NUTS
+from ._gibbs import HybridGibbs
+from ._conjugate import Conjugate
+from ._conjugate_approx import ConjugateApprox
+from ._direct import Direct
+from ._utilities import find_valid_samplers

{cuqipy-1.1.1.post0.dev78 → cuqipy-1.1.1.post0.dev88}/cuqi/experimental/mcmc/_gibbs.py

@@ -1,6 +1,6 @@
 from cuqi.distribution import JointDistribution
 from cuqi.experimental.mcmc import Sampler
-from cuqi.samples import Samples
+from cuqi.samples import Samples, JointSamples
 from cuqi.experimental.mcmc import NUTS
 from typing import Dict
 import numpy as np
@@ -161,7 +161,7 @@ class HybridGibbs:
 
         """
 
-        for _ in tqdm(range(Ns)):
+        for _ in tqdm(range(Ns), "Sample: "):
 
             self.step()
 
@@ -184,7 +184,7 @@ class HybridGibbs:
 
         tune_interval = max(int(tune_freq * Nb), 1)
 
-        for idx in tqdm(range(Nb)):
+        for idx in tqdm(range(Nb), "Warmup: "):
 
             self.step()
 
@@ -197,7 +197,7 @@ class HybridGibbs:
         return self
 
     def get_samples(self) -> Dict[str, Samples]:
-        samples_object = {}
+        samples_object = JointSamples()
         for par_name in self.par_names:
             samples_array = np.array(self.samples[par_name]).T
             samples_object[par_name] = Samples(samples_array, self.target.get_density(par_name).geometry)

{cuqipy-1.1.1.post0.dev78 → cuqipy-1.1.1.post0.dev88}/cuqi/experimental/mcmc/_sampler.py

@@ -220,7 +220,7 @@ class Sampler(ABC):
         if hasattr(self, "_pre_sample"): self._pre_sample()
 
         # Draw samples
-        for _ in tqdm( range(Ns) ):
+        for _ in tqdm( range(Ns), "Sample: "):
             
             # Perform one step of the sampler
             acc = self.step()
@@ -260,7 +260,7 @@ class Sampler(ABC):
         if hasattr(self, "_pre_warmup"): self._pre_warmup()
 
         # Draw warmup samples with tuning
-        for idx in tqdm(range(Nb)):
+        for idx in tqdm(range(Nb), "Warmup: "):
 
             # Perform one step of the sampler
             acc = self.step()

cuqipy-1.1.1.post0.dev88/cuqi/samples/__init__.py

@@ -0,0 +1 @@
+from ._samples import Samples, JointSamples

{cuqipy-1.1.1.post0.dev78 → cuqipy-1.1.1.post0.dev88}/cuqi/samples/_samples.py

@@ -867,3 +867,23 @@ class Samples(object):
                "Geometry:\n {}\n\n".format(self.geometry) + \
                "Shape:\n {}\n\n".format(self.shape) + \
                "Samples:\n {}\n\n".format(self.samples)
+
+class JointSamples(dict):
+    """ An object used to store samples from :class:`cuqi.distribution.JointDistribution`. 
+
+    This object is a simple overload of the dictionary class to allow easy access to certain methods 
+    of Samples objects without having to iterate over each key in the dictionary. 
+    
+    """
+
+    def burnthin(self, Nb, Nt=1):
+        """ Remove burn-in and thin samples for all samples in the dictionary. Returns a copy of the samples stored in the dictionary. """
+        return JointSamples({key: samples.burnthin(Nb, Nt) for key, samples in self.items()})
+
+    def __repr__(self) -> str: 
+        return "CUQIpy JointSamples Dict:\n" + \
+               "-------------------------\n\n" + \
+               "Keys:\n {}\n\n".format(list(self.keys())) + \
+               "Ns (number of samples):\n {}\n\n".format({key: samples.Ns for key, samples in self.items()}) + \
+               "Geometry:\n {}\n\n".format({key: samples.geometry for key, samples in self.items()}) + \
+               "Shape:\n {}\n\n".format({key: samples.shape for key, samples in self.items()})

cuqipy-1.1.1.post0.dev78/cuqi/experimental/mcmc/__init__.py

@@ -1,15 +0,0 @@
-""" Re-implementation of sampler module in a more object oriented way. """
-
-from ._sampler import Sampler, ProposalBasedSampler
-from ._langevin_algorithm import ULA, MALA
-from ._mh import MH
-from ._pcn import PCN
-from ._rto import LinearRTO, RegularizedLinearRTO
-from ._cwmh import CWMH
-from ._laplace_approximation import UGLA
-from ._hmc import NUTS
-from ._gibbs import HybridGibbs
-from ._conjugate import Conjugate
-from ._conjugate_approx import ConjugateApprox
-from ._direct import Direct
-from ._utilities import find_valid_samplers

cuqipy-1.1.1.post0.dev78/cuqi/samples/__init__.py

@@ -1 +0,0 @@
-from ._samples import Samples