def retrievedpost(datadir,
outname,
pnames,
true,
shift,
fpost='output.npy',
outdir=None):
"""
datadir : string. Path/to/directory containing BART-formatted output.
outname : string. File name of resulting plot. Do not include path, use
`outdir` for that purpose.
pnames : list, strings. Parameter names.
true : array. True values.
shift : array. Shift values of the posterior
(e.g., to correct for the assumed starting abundance)
fpost : string. File name of .NPY file containing the posterior.
outdir : string. Path/to/dir to save `outname`. If None, defaults
to the results directory of BARTTest.
"""
# Make sure datadir has a trailing /
if datadir[-1] != '/':
datadir = datadir + '/'
# Set outdir if not specified. If it is, ensure trailing /
if outdir is None:
try:
outdir = 'results'.join(datadir.rsplit('code-output', \
1)).rsplit('/', 2)[0] + '/'
except:
print("Data directory not located within BARTTest.")
print("Please specify an output directory `outdir` and try again.")
sys.exit(1)
elif outdir[-1] != '/':
outdir = outdir + '/'
# Read MCMC output file
MCfile = datadir + 'MCMC.log'
bestP, uncer = bf.read_MCMC_out(MCfile)
allParams = bestP
# Get number of burned iterations
foo = open(MCfile, 'r')
lines = foo.readlines()
foo.close()
line = [
foop for foop in lines if " Burned in iterations per chain:" in foop
]
burnin = int(line[0].split()[-1])
# Load the data
data = np.load(datadir + fpost)
data_stack = data[0, :, burnin:]
for c in np.arange(1, data.shape[0]):
data_stack = np.hstack((data_stack, data[c, :, burnin:]))
# Shift
data_stack += shift[:, None]
if data_stack.shape[0] != len(pnames):
raise ValueError(
"Posterior dimensionality does not match number of parameter names."
)
# Plot it
#mcp.histogram(data_stack, parname=pnames, truepars=true, savefile=outdir+outname)
mcp.pairwise(data_stack,
parname=pnames,
truepars=true,
savefile=outdir + outname)
def retrievedPT(datadir,
atmfile,
tepfile,
nmol,
solution,
outname,
outdir=None,
T_int=100.):
"""
Inputs
------
datadir: string. Path/to/directory containing BART-formatted output.
atmfile: string. Path/to/atmospheric model file.
tepfile: string. Path/to/Transiting ExoPlanet file.
nmol: int. Number of molecules being fit by MCMC.
solution: string. Geometry of the system. 'eclipse' or 'transit'.
outname: string. File name of resulting plot.
outdir: string. Path/to/dir to save `outname`. If None, defaults
to the results directory of BARTTest if `datadir` is
within BARTTest.
T_int: float. Internal planetary temperature. Default is 100 K.
"""
# Set outdir if not specified
if outdir is None:
try:
if datadir[-1] != '/':
datadir = datadir + '/'
outdir = 'results'.join(datadir.rsplit('code-output', \
1)).rsplit('/', 2)[0] + '/'
if not os.path.isdir(outdir):
os.makedirs(outdir)
except:
print("Data directory not located within BARTTest.")
print("Please specify an output directory `outdir` and try again.")
sys.exit(1)
# Read g_surf and R_planet from TEP file
grav, Rp = ma.get_g(tepfile)
# Read star data from TEP file, and semi-major axis
R_star, T_star, sma, gstar = bf.get_starData(tepfile)
# Read atmfile
mols, atminfo = readatm(atmfile)
pressure = atminfo[:, 1]
# Read MCMC output file
MCfile = datadir + 'MCMC.log'
bestP, uncer = bf.read_MCMC_out(MCfile)
allParams = bestP
# Get number of burned iterations
foo = open(MCfile, 'r')
lines = foo.readlines()
foo.close()
line = [
foop for foop in lines if " Burned in iterations per chain:" in foop
]
burnin = int(line[0].split()[-1])
# Figure out number of parameters
nparams = len(allParams)
nradfit = int(solution == 'transit')
nPTparams = nparams - nmol - nradfit
PTparams = allParams[:nPTparams]
# Plot the best PT profile
kappa, gamma1, gamma2, alpha, beta = PTparams
best_T = pt.PT_line(pressure, kappa, gamma1, gamma2, alpha, beta, R_star,
T_star, T_int, sma, grav * 1e2, 'const')
# Load MCMC data
MCMCdata = datadir + 'output.npy'
data = np.load(MCMCdata)
nchains, npars, niter = np.shape(data)
# Make datacube from MCMC data
data_stack = data[0, :, burnin:]
for c in np.arange(1, nchains):
data_stack = np.hstack((data_stack, data[c, :, burnin:]))
# Datacube of PT profiles
PTprofiles = np.zeros((np.shape(data_stack)[1], len(pressure)))
curr_PTparams = PTparams
for k in np.arange(0, np.shape(data_stack)[1]):
j = 0
for i in np.arange(len(PTparams)):
curr_PTparams[i] = data_stack[j, k]
j += 1
kappa, gamma1, gamma2, alpha, beta = curr_PTparams
PTprofiles[k] = pt.PT_line(pressure, kappa, gamma1, gamma2, alpha,
beta, R_star, T_star, T_int, sma,
grav * 1e2, 'const')
# Get percentiles (for 1, 2-sigma boundaries):
low1 = np.percentile(PTprofiles, 16.0, axis=0)
hi1 = np.percentile(PTprofiles, 84.0, axis=0)
low2 = np.percentile(PTprofiles, 2.5, axis=0)
hi2 = np.percentile(PTprofiles, 97.5, axis=0)
median = np.median(PTprofiles, axis=0)
# Plot and save figure
plt.figure(2)
plt.clf()
ax = plt.subplot(111)
ax.fill_betweenx(pressure, low2, hi2, facecolor="#62B1FF", edgecolor="0.5")
ax.fill_betweenx(pressure,
low1,
hi1,
facecolor="#1873CC",
edgecolor="#1873CC")
plt.semilogy(median, pressure, "-", lw=2, label='Median', color="k")
plt.semilogy(best_T, pressure, "-", lw=2, label="Best fit", color="r")
plt.semilogy(atminfo[:, 2], pressure, "--", lw=2, label='Input', color='r')
plt.ylim(pressure[0], pressure[-1])
plt.legend(loc="best")
plt.xlabel("Temperature (K)", size=15)
plt.ylabel("Pressure (bar)", size=15)
plt.savefig(outdir + outname)
plt.close()
])
logfile = './7species_7opac_uniform/MCMC.log'
stepsize = np.array(
[0.01, 0.01, 0.0, 0.0, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01, 0.01])
MCMCdata = './7species_7opac_uniform/output.npy'
# read atmfile
molecules, pressure, temp, abundances = mat.readatm(atmfile)
# get surface gravity
grav, Rp = mat.get_g(tep_name)
# convert gravity to cm/s^2
grav = grav * 1e2
# get star data
R_star, T_star, sma, gstar = bf.get_starData(tep_name)
# get best parameters
bestP, uncer = bf.read_MCMC_out(logfile)
# get all params
allParams = bf.get_params(bestP, stepsize, params)
# get PTparams and abundances factors
nparams = len(allParams)
nmol = len(molfit)
nPTparams = nparams - nmol
PTparams = allParams[:nPTparams]
kappa, gamma1, gamma2, alpha, beta = PTparams
# HARDCODED !
T_int = 100 # K
T_int_type = 'const'
# call PT line profile to calculate temperature
best_T = pt.PT_line(pressure, kappa, gamma1, gamma2, alpha, beta, R_star,
T_star, T_int, sma, grav, T_int_type)