def plot_mcmc_bestfit():
"""
Load a mcmc bestfit and plot it
"""
from pyRSD.rsdfit import FittingDriver
import plotify as pfy
parser = argparse.ArgumentParser()
parser.formatter_class = argparse.RawTextHelpFormatter
# arguments
h = 'the name of directory holding the results'
parser.add_argument('results_dir', type=str, help=h)
h = 'the name of the results file to load'
parser.add_argument('results_file', type=str, help=h)
h = 'the name of the model to load'
parser.add_argument('model', type=str, help=h)
h = 'the output file name'
parser.add_argument('-o', '--output', type=str, help=h)
h = 'the number of burnin steps to set'
parser.add_argument('-b', '--burnin', type=int, help=h)
args = parser.parse_args()
# load the driver
kwargs = {'results_file':args.results_file, 'model_file':args.model}
driver = FittingDriver.from_directory(args.results_dir, **kwargs)
if args.burnin is not None:
driver.results.burnin = args.burnin
driver.set_fit_results()
driver.plot()
if args.output is not None:
pfy.savefig(args.output)
pfy.show()
def compare_bestfits(mode, **kwargs):
"""
Compare the best-fit parameters to the data.
Parameters
----------
mode : str, {`function`, `params`, `gp`, `spline`}
either compare to bestfit function or bestfit params
kwargs : key/value pairs
data : subclass of lsskit.speckmod.plugins.ModelInput
plugin instance specifying the data
model : subclass of lsskit.speckmod.plugins.ModelInput
plugin instance specifying the model, only needed
for `mode == params`
bestfit_file : str
the name of the file holding the pickled dataframe
select : list of str
a list holding strings with the format should
`index_col`:value'. This specifies which bin to be
compared
"""
import plotify as pfy
import pandas as pd
if mode == 'gp':
from pyRSD.rsd.mu0_modeling import GPModelParams
elif mode == 'spline':
from pyRSD.rsd.mu0_modeling import SplineTableModelParams
if mode not in ['function', 'params', 'gp', 'spline']:
raise ValueError("``mode`` in compare_bestfits must be `function`, `params`, `gp`, or `spline`")
# make the index cols
try:
index_cols = [x.split(':')[0] for x in kwargs['select']]
select = [int(x.split(':')[1]) for x in kwargs['select']]
except:
raise ValueError("``select`` should have format: `index_col`:value")
# read the bestfits file and select
df = pd.read_pickle(kwargs['bestfit_file'])
valid = index_cols
if mode == 'function': valid += ['k']
if not all(x in df.columns for x in valid):
raise ValueError("please specify a bestfit file with columns: %s" %(", ".join(valid)))
df = df.set_index(valid)
# get the key dictionary and print out what we are selecting
key = dict((df.index.names[i], df.index.levels[i][v]) for i, v in enumerate(select))
msg = ", ".join("%s = %s" %(k,v) for k,v in key.items())
print("selecting " + msg)
# select the bestfit
select = tuple(df.index.levels[i][v] for i, v in enumerate(select))
# load the GP
if mode == 'gp':
gp = GPModelParams(kwargs['gp_file'])
args = tuple(df.loc[select, col] for col in kwargs['interp_cols'])
bestfits = gp.to_dict(*args)
print("gp bestfit values:\n-------------")
print("\n".join("%s = %s" %(k,str(v)) for k,v in bestfits.items()))
actual = {k:df.loc[select, k] for k in kwargs['model'].param_names}
print("actual bestfit values:\n-------------")
print("\n".join("%s = %s" %(k,str(v)) for k,v in actual.items()))
# load the spline table
elif mode == 'spline':
table = SplineTableModelParams(kwargs['spline_file'])
s8_z = df.loc[select, 's8_z']
b1 = df.loc[select, 'b1']
bestfits = table(s8_z, b1)
print("spline table bestfit values:\n-------------")
print("\n".join("%s = %s" %(k,str(v)) for k,v in bestfits.items()))
actual = {k:df.loc[select, k] for k in kwargs['model'].param_names}
print("actual bestfit values:\n-------------")
print("\n".join("%s = %s" %(k,str(v)) for k,v in actual.items()))
elif mode == 'params':
bestfits = {k:df.loc[select, k] for k in kwargs['model'].param_names}
if mode == 'params' or mode == 'gp' or mode == 'spline':
kwargs['data'].select = key
# this should hopefully only loop over one thing
for key, extra, data_df in kwargs['data']:
# plot the data
pfy.errorbar(data_df.index.values, data_df['y'], data_df['error'])
# plot the bestfit parameters
x = data_df.index.values
y = kwargs['model'](x, **dict(bestfits, **extra))
lines = pfy.plot(x, y)
else: # mode is `function`
df = df.xs(select)
# select the data
data_df = kwargs['data'].to_dataframe(key)
# plot the data
pfy.errorbar(data_df.index.values, data_df['y'], data_df['error'])
# plot the bestfit function mean
x = df.index.values
y = df['mean']
errs = df['error']
lines = pfy.plot(x, y)
pfy.plt.fill(np.concatenate([x, x[::-1]]),
np.concatenate([y - errs,
(y + errs)[::-1]]),
alpha=.5, fc=lines[0].get_color(), ec='None')
ax = pfy.gca()
ax.title.update('Bestfit (%s) comparison for %s' %(mode,msg), fontsize=16)
ax.xlabel.update(r"$k$ ($h$/Mpc)", fontsize=16)
ax.ylabel.update(kwargs['data'].variable_str, fontsize=16)
pfy.show()