[Draft] Support for converting from dynesty and emcee samples

anesthetic/convert.py

@@ -1,4 +1,5 @@
 """Tools for converting to other outputs."""
+from anesthetic.samples import NestedSamples, MCMCSamples
 import numpy as np
 
 
@@ -25,3 +26,97 @@ def to_getdist(samples):
                                        names=samples.columns,
                                        ranges=ranges,
                                        labels=labels)
+
+
+def from_anesthetic(dynesty_sampler, columns=None, tex=None, limits=None):
+    """Convert from dynesty to anesthetic samples.
+
+    Parameters
+    ----------
+    dynesty_sampler: dynesty.dynesty.NestedSampler or dynesty.results.Results
+        dynesty NestedSampler instance to copy results from
+        (must have been run already), or results of a dynesty run.
+
+    columns: list, optional
+        List of (python) parameter names
+        default: None
+
+    tex: list, optional
+        List of LaTeX parameter names
+        default: None
+
+    limits: list, optional
+        List of parameter limits
+        default: None
+
+    Returns
+    -------
+    nested_samples: NestedSamples
+        anesthetic nested samples
+    """
+    import dynesty.results
+    import dynesty.dynesty
+
+    if isinstance(dynesty_sampler, dynesty.dynesty.NestedSampler):
+        dynesty_results = dynesty_sampler.results
+    elif isinstance(dynesty_sampler, dynesty.results.Results):
+        dynesty_results = dynesty_sampler
+    else:
+        raise ValueError("Unknown dynesty input type")
+
+    data = dynesty_results['samples']
+    weights = np.exp(dynesty_results['logwt'])
+    logl = dynesty_results['logl']
+    # dynesty_results['logz']
+    # logL_birth
+    # label
+    # beta
+    # logzero
+    return NestedSamples(data=data,
+                         weights=weights,
+                         logL=logl,
+                         columns=columns,
+                         tex=tex,
+                         limits=limits)
+
+
+def from_emcee(emcee_sampler, columns=None, tex=None, limits=None):
+    """Convert from emcee to anesthetic samples.
+
+    Parameters
+    ----------
+    emcee_sampler: emcee.ensemble.EnsembleSampler
+        emcee sampler to copy results from
+
+    columns: list, optional
+        List of (python) parameter names
+        default: None
+
+    tex: list, optional
+        List of LaTeX parameter names
+        default: None
+
+    limits: list, optional
+        List of parameter limits
+        default: None
+
+    Returns
+    -------
+    mcmc_samples: MCMCSamples
+        anesthetic MCMC samples
+    """
+    import emcee.ensemble
+    if not isinstance(emcee_sampler, emcee.ensemble.EnsembleSampler):
+        raise ValueError("Wrong input type, please pass\
+                          emcee.ensemble.EnsembleSampler")
+
+    data = emcee_sampler.flatchain
+    # weights = 1
+    # logL
+    # label
+    # logzero
+    # burn_in
+    return MCMCSamples(data=data,
+                       columns=columns,
+                       tex=tex,
+                       limits=limits)