Witchblade101
commented
2 weeks ago I can't attach .ecf or .epf files, so they're pasted below:
# Eureka! Control File for Stage 5: Lightcurve Fitting
# Stage 5 Documentation: https://eurekadocs.readthedocs.io/en/latest/ecf.html#stage-5
ncpu 12 # The number of CPU threads to use when running emcee or dynesty in parallel
multwhite False # Are you simultaneously fitting multiple white lightcurves? (if so, use inputdirlist and see docs)
allapers False # Run S5 on all of the apertures considered in S4? Otherwise will use newest output in the inputdir
# Manual clipping in time
manual_clip None # A list of lists specifying the start and end integration numbers for manual removal.
fit_par ./S5_t1be_3.epf # What fitting epf do you want to use?
fit_method [dynesty] #options are: lsq, emcee, dynesty (can list multiple types separated by commas)
run_myfuncs [batman_tr,polynomial] # options are: batman_tr, batman_ecl, sinusoid_pc, quasilambert_pc, poet_tr, poet_ecl, poet_pc, fleck_tr, expramp, polynomial, step, xpos, ypos, xwidth, ywidth, lorentzian, damped_osc, and GP (can list multiple types separated by commas)
compute_ltt False # options are: True (correct model for the light travel time effect), or False (ignore the light travel time effect)
num_planets 2
force_positivity False # Optional boolean for sinusoid_pc and poet_pc models. Set True to force positive phase variations.
# Limb darkening controls
# IMPORTANT: limb-darkening coefficients are not automatically fixed then, change to 'fixed' in .epf file whether they should be fixed or fitted!
use_generate_ld None # use the generated limb-darkening coefficients from Stage 4? Options: exotic-ld, spam, None. For exotic-ld and spam, the limb-darkening laws available are linear, quadratic, 3-parameter and 4-parameter non-linear.
ld_file None # Fully qualified path to the location of a limb darkening file that you want to use
ld_file_white None # Fully qualified path to the location of a limb darkening file that you want to use for the white-light light curve (required if ld_file is not None and any EPF parameters are set to white_free or white_fixed).
# General fitter
old_fitparams None # filename relative to topdir that points to a fitparams csv to resume where you left off (set to None to start from scratch)
# lsq fitter
lsq_method Powell # The scipy.optimize.minimize optimization method to use
lsq_tol 1e-7 # The tolerance for the scipy.optimize.minimize optimization method
lsq_maxiter None # Maximum number of iterations to perform. Depending on the method each iteration may use several function evaluations. Set to None to use the default value
# emcee fitter
old_chain None # Output folder relative to topdir that contains an old emcee chain to resume where you left off (set to None to start from scratch)
lsq_first True # Initialize with an initial lsq call (can help shorten burn-in, but turn off if lsq fails). Only used if old_chain is None
run_nsteps 1000
run_nwalkers 200
run_nburn 500 # How many of run_nsteps should be discarded as burn-in steps
# dynesty fitter
run_nlive 'min' # Must be > ndim * (ndim + 1) // 2
run_bound 'multi'
run_sample 'auto'
run_tol 0.1
# PyMC3 NUTS fitter
exoplanet_first False # Initialize with an initial exoplanet optimizer call (generally not recommended, but can sometimes be helpful if your initial manual guess is quite poor)
tune 3000
draws 1500
chains 3
target_accept 0.85
#GP inputs
kernel_inputs ['time'] #options: time
kernel_class ['Matern32'] #options: ExpSquared, Matern32, Exp, RationalQuadratic for george, Matern32 for celerite (sums of kernels possible for george separated by commas)
GP_package 'celerite' #options: george, celerite
# Plotting controls
interp False # Should astrophysical model be interpolated (useful for uneven sampling like that from HST)
# Diagnostics
isplots_S5 5 # Generate few (1), some (3), or many (5) figures (Options: 1 - 5)
nbin_plot None # The number of bins that should be used for figures 5104 and 5304
hide_plots True # If True, plots will automatically be closed rather than popping up
verbose True # If True, more details will be printed about steps
# Project directory
topdir /Users/dlong/DataANalysis/JWST/eureka/t1be/obs3/
# Directories relative to topdir
inputdir Stage4 # The folder containing the outputs from Eureka!'s S4 pipeline (will be overwritten if calling S4 and S5 sequentially)
inputdirlist None # list of extra folders to read in from after inputdir for multwhite fits. Must be strings.
outputdir Stage5
# inputdirlist [] # A list of additional Stage 4 input folders to use in a joint white lightcurve fit (e.g., NIRSpec NRS1+NRS2, or NIRCam+MIRI). Only relevant if multwhite is set to True# Stage 5 Fit Parameters Documentation: https://eurekadocs.readthedocs.io/en/latest/ecf.html#stage-5-fit-parameters
# Name Value Free? PriorPar1 PriorPar2 PriorType
# "Free?" can be free, fixed, white_free, white_fixed, shared, or independent
# PriorType can be U (Uniform), LU (Log Uniform), or N (Normal).
# If U/LU, PriorPar1 and PriorPar2 represent upper and lower limits of the parameter/log(the parameter).
# If N, PriorPar1 is the mean and PriorPar2 is the standard deviation of a Gaussian prior.
#-------------------------------------------------------------------------------------------------------
# ------------------
# ** Transit/eclipse parameters **
# ------------------
rprs 0.07 'free' 0.000000 0.200000 U # Natalie
rprs_pl1 0.08 'free' 0.000000 0.200000 U # Natalie
# ------------------
# ** Orbital parameters **
# ------------------
per 6.101012 'free' 6.101012 0.000017 N # Natalie
per_pl1 1.510826 'free' 1.510826 0.000006 N # Natalie
t0 .24 'free' 0.23 0.25 U #
t0_pl11 .42 'free' 0.41 0.43 U #
time_offset 60502.0 'independent'
inc 89.813 'fixed' 89.813 60 N #
inc_pl1 89.809 'fixed' 89.809 60 N #
ars 52.72 'free' 52.72 0.13 N # Natalie
ars_pl1 20.809884 'free' 20.84 0.16 N # Natalie
ecc 0.0 'fixed' 0 1 U
ecc_pl1 0.0 'fixed' 0 1 U
w 90. 'fixed' 0 180 U
w_pl1 90. 'fixed' 0 180 U
# ----------------------
# ** Sinusoid_pc Phase Curve Parameters **
# ----------------------
# AmpCos1 0.4 'free' 0 1 U
# AmpSin1 0.01 'free' -1 1 U
# AmpCos2 0.01 'free' -1 1 U
# AmpSin2 0.01 'free' -1 1 U
# ----------------------
# ** POET_pc Phase Curve Parameters **
# ----------------------
# cos1_amp 0.4 'free' 0 1 U
# cos1_off 0.00 'free' -180 180 U
# cos2_amp 0.1 'free' 0 1 U
# cos2_off 0.00 'free' 0 180 U
# ----------------------
# ** quasilambert_pc Phase Curve Parameters **
# ----------------------
# quasi_gamma 4 'free' 2 40 U
# quasi_offset 0 'free' -90 90 U
# -------------------------
# ** Limb darkening parameters **
# For BATMAN, choose limb_dark from ['uniform', 'linear', 'quadratic', 'kipping2013', 'squareroot', 'logarithmic', 'exponential','3-parameter', '4-parameter']
# For POET, choose limb_dark from ['uniform', 'linear', 'quadratic', '4-parameter']
# When using generated limb-darkening coefficients from exotic-ld or spam, choose from ['linear', 'quadratic', '3-parameter', '4-parameter'] and set their state to 'fixed' or use `recenter_ld_prior`.
# -------------------------
limb_dark 'kipping2013' 'independent'
u1 0.2 'free' 0 1 U
u2 0.2 'free' 0 1 U
# -------------------------
# ** Stellar variability parameters **
# A modified Lorentzian model can be used to fit stellar flares
# A damped oscillator model can be used to fit quasi-periodic pulsations
# -------------------------
# lor_amp_lhs 0.03 'free' 0 0.1 U
# lor_amp_rhs 0.03 'free' 0 0.1 U
# lor_hwhm_lhs 0 'fixed' 0 0.1 U
# lor_hwhm_rhs 1e-5 'free' 0 0.1 U
# lor_t0 0.3163 'free' 0.3160 0.3166 U
# lor_power 0.7 'free' 0.1 2 U
# osc_amp 0.0003 'free' 0 0.01 U
# osc_amp_decay 1 'free' 0 20 U
# osc_per 0.03 'free' 0 0.1 U
# osc_per_decay 0. 'fixed' 0 10 U
# osc_t0 0.3163 'fixed' 0 1 U
# osc_t1 0.3163 'free' 0.3 0.4 U
# ------------------
# ** Star Spot parameters **
# Each spot needs rad, lat, lon. See docs for details!
# For fleck: can only have one spot contrast (spotcon), spotnpts is the number of temporal points to evaluate at (~200-500)
# For starry: assign one spotcon per spot (spotcon, spotcon1, etc), spotnpts is the ydeg of the star map (~30)
# ------------------
# spotstari 90 'fixed'
# spotrot 100 'fixed'
# spotnpts 300 'independent'
# spotcon 0.9 'free' 0 1 U # Should be between 0 and 1
# spotrad 0.2 'free' 0 1 U # Should be between 0 and 1
# spotlat 0 'free' -90 90 U # Should be between -90 and 90
# spotlon 0 'free' -90 90 U # Should be between -90 and 90 to be on observer-facing hemisphere
# spotcon1 0.95 'free' 0 1 U # Only use for starry! fleck cannot have different contrasts for each spot
# spotrad1 0.1 'free' 0 1 U # Should be between 0 and 1
# spotlat1 10 'free' -90 90 U # Should be between -90 and 90
# spotlon1 45 'free' -90 90 U # Should be between -90 and 90 to be on observer-facing hemisphere
# --------------------
# ** Systematic variables **
# Polynomial model variables (c0--c9 for 0th--3rd order polynomials in time); Fitting at least c0 is very strongly recommended!
# Exponential ramp model variables (r0--r1 for one exponential ramp, r2--r3 for a second exponential ramp)
# GP model parameters (A, m for the first kernel; A1, m1 for the second kernel; etc.) in log scale
# Step-function model variables (step# and steptime# for step-function model #; e.g. step0 and steptime0)
# Drift model variables (xpos, ypos, xwidth, ywidth)
# --------------------
c0 1.015 'free' 1 0.05 N
c1 0 'free' 0 0.01 N
# A -15 'free' -25 0 U
# m -4 'free' -7 0 U
# -----------
# ** White noise **
# Use scatter_mult to fit a multiplier to the expected noise level from Stage 3 (recommended)
# Use scatter_ppm to fit the noise level in ppm
# -----------
scatter_mult 1.1 'free' 0.8 10 U
kevin218
taylorbell57
FAQ check
Instrument
Light curve fitting (Stages 4-6)
What happened?
I installed the latest version:
I reran stages 2 - 4 using new .ecf files to avoid any potential problems. Everything looked good.
On Stage 5, I added
num_planetsandpl1where needed as described in the updated documentation. Stage 5 crashed.Error traceback output
What operating system are you using?
Mac OS Sonoma 14.6.1
What version of Python are you running?
Python 3.10.14
What Python packages do you have installed?
Code of Conduct