False positive probabilities for all KOIs in the Q1-Q17 (DR24) Kepler table. The one-stop-shop table for the summarized FPP results is fpp_final_table.csv.
If you're curious about individual KOIs and want to browse the diagnostic/result plots, please go here.
If you're really curious, all the data files produced by these
calculations are currently hosted at Princeton. To browse the
results from a given KOI, visit
tigress-web.princeton.edu/~tmorton/koi-fpp/K?????.??, with the
appropriate KOI identifier. These directories contain the entire
isochrones outputs for every KOI (including the stellar
posteriors for the single-, binary- and triple-star fits). If you
have specific questions, please feel free to contact me.
The manuscript PDF can be generated by entering the
contain all the commands to make the figures & tables that live in
document directory, with the exception of the Santerne+ (2015)
comparison table, which is created by
You can run your own FPP calculations using all of the same data and constraints that I used for this work. If you wish to do this, first test out the calculation for the KOI for which I provide example data in this repository (KOI-7016.01/Kepler-452b).
First, clone this repository and hop on in:
git clone https://github.com/timothydmorton/koi-fpp cd koi-fpp
Then define some temporary enivronment variables (assuming you're calling this from inside this repository):
export KOI_FPPDIR="$PWD/example_data" export JROWE_DIR="$PWD/example_data/photometry"
Then, set up the python environment using the
environment.yml file in the top
level of this repository, as follows. [If you do not have
conda available, install miniconda first.]
conda env create -f environment.yml source activate koifpp
This should install all the required packages. Now from within this environment you can run the following:
This will set up the
isochrones config files in a
K07016.01 directory under
KOI_FPPDIR (if you haven't yet used
isochrones, then be prepared for a few minutes' worth of stellar
model downloads). It will also do the
trapezoid shape fitting to the photometry. Take a look at the
star.ini files to see what the inputs look like. You
can then run the calculation as follows:
cd $KOI_FPPDIR calcfpp K07016.01
Note that the stellar parameter inference (and thus the FPP
calculation) will be much more reliable if you have
pymultinest installed. The calculations will run just
using the default
emcee sampler, but I stronly recommend using
multinest becaues of the inherently multimodal nature of the
problem. Follow these
In order to get the data necessary for all the KOI FPP calculations beyond this single example, you will need the following:
- TRILEGAL starfield simulations. Define a
KOI_FPPDIRenvironment variable to be somewhere sensible to you, and unpack this
starfielddirectory within that.
- Kepler photometry. This is in two separate tarballs (Zenodo max file size is 2Gb), so unpack them both and combine their contents (all the koi*.n sub-directories) into a single directory. Define a
JROWE_DIRenvironment variable referring to this location.
Then you can run the
calcfpp commands for your favorite KOI(s).
data directory contains the following files:
fpp_all.txt: concatenation of all the
results.txtfiles from all the successful
vespacalculations. [TDM note: created by
fpp_err.txt: concatenation of all the exceptions raised by all the failed
vespacalculations. [TDM note: created by
ttvdata.txt: table of whether TTV information was used to create the folded Kepler photometry used for the trapezoid fits. [TDM note: created by
starprops_all.txt: summarized information from all the single-star
isochronesfits to all the KOI stars. [TDM note: created by
starfit-summarizeon all KOIs]
positional_probability.csv: table of positional probability data from Steve Bryson. This should correspond to the "positional probability" table at the NASA Exoplanet Archive.
make_finale_fpptable.py script uses the data from these files
as well as the DR24 table to produce the
fpp_final_table.csv file in the top
level of this repository.