add missing file

cosmosis/postprocessing/main.py

@@ -0,0 +1,87 @@
+from .postprocess import postprocessor_for_sampler
+from .inputs import read_input
+from .plots import Tweaks
+from ..utils import mkdir
+
+
+
+def run_cosmosis_postprocess(inputs, **kwargs):
+    """
+    Run the postprocessing steps on the given input file(s).
+
+    Parameters
+    ----------
+    inputs : list of str or output objects
+        The input file(s) to run postprocessing on.
+
+
+    """
+    #Make the directory for the outputs to go in.
+    outdir = kwargs.get("outdir", ".")
+    legend = kwargs.get("legend")
+    run_max_post = kwargs.get("run_max_post", False)
+    text = kwargs.get("text", False)
+    weights = kwargs.get("weights", False)
+    extra = kwargs.get("extra")
+    tweaks = kwargs.get("tweaks")
+
+    mkdir(outdir)
+    outputs = {}
+
+    #Deal with legends, if any
+    if legend:
+        labels = legend.split("|")
+        if len(labels)!=len(inputs):
+            raise ValueError("You specified {} legend names but {} files to plot".format(len(labels), len(inputs)))
+    elif isinstance(inputs[0], str):
+        labels = inputs
+    else:
+        labels = [str(i) for i in range(len(inputs))]
+
+    if len(inputs)>1 and run_max_post:
+        raise ValueError("Can only use the --run-max-post argument with a single parameter file for now")
+
+    for i, ini_filename in enumerate(inputs):
+        sampler, ini = read_input(ini_filename, text, weights)
+        processor_class = postprocessor_for_sampler(sampler.split()[-1])
+
+        #We do not know how to postprocess everything.
+        if processor_class is None:
+            print("I do not know how to postprocess output from the %s sampler"%sampler)
+            sampler = None
+            continue
+
+        #Create and run the postprocessor
+
+        processor = processor_class(ini, labels[i], i, **kwargs)
+
+        #Inherit any plots from the previous postprocessor
+        #so we can make plots with multiple datasets on
+        processor.outputs.update(outputs)
+
+        #We can load extra plots to make from a python
+        #script here
+        if extra:
+            processor.load_extra_steps(extra)
+
+        #Optionally add a step in which we re-run the best-fit
+        if run_max_post:
+            processor.add_rerun_bestfit_step(run_max_post)
+
+        #Run the postprocessor and make the outputs for this chain
+        processor.run()
+
+        #Save the outputs ready for the next post-processor in case
+        #they want to add to it (e.g. two constriants on the same axes)
+        outputs = processor.outputs
+
+    if sampler is None:
+        return
+
+    #Run any tweaks that the user specified
+    if tweaks:
+        tweaks = Tweaks.instances_from_file(tweaks)
+        for tweak in tweaks:
+            processor.apply_tweaks(tweak)
+
+    return processor