sgongar / serval

calculate radial velocities from stellar spectra

SERVAL (SpEctrum Radial Velocity AnaLyser)

calculate radial velocities from stellar spectra

The concept of SERVAL is described in http://adsabs.harvard.edu/abs/2017A&A...609A..12Z [pdf].

Currently, SERVAL can process data from CARM_VIS, CARM_NIR, ELODIE, HARPS, HARPN, and SOPHIE.

Recent changes

• request to simbad to get RADE, PM, and PLX via -targ
• barycentric correction with Wright & Eastman (2014) is now default (requires https://github.com/shbhuk/barycorrpy)

Install instruction

Requirements:

• python 2.7 + numpy, scipy, pyfits, astropy
• gnuplot

Setup the path:

export SERVALHOME=~/mzechmeister
export SERVAL=$SERVALHOME/serval/
export PYTHONPATH=$PYTHONPATH:$SERVALHOME/python/

You can include these lines into your ~/.bashrc. Above is for bash; for tcsh:

setenv SERVALHOME ~
...

Download SERVAL and required tools:

mkdir $SERVALHOME
cd $SERVALHOME
git clone https://github.com/mzechmeister/serval.git
git clone https://github.com/mzechmeister/python.git

Make main files executable:

chmod u+x $SERVAL/src/serval.py
chmod u+x $SERVAL/src/read_spec.py

Install barycorrpy:

pip install barycorrpy

or if you don't have root rights:

pip install --user barycorrpy

See also https://github.com/shbhuk/barycorrpy/wiki/1.-Installation for other possibilities. Note there is a numpy 1.14.0 einsum issue which I reported in astropy/astropy#7051 (comment). I fixed this by removing the unicode_literals in https://github.com/astropy/astropy/blob/v2.0.x/astropy/coordinates/builtin_frames/utils.py. But there might be other ways.

A few c programs come precompiled. But probably it is necessary to compile (e.g. Mac OS) ...

cd $SERVAL/src/
gcc -c  -Wall -O2 -ansi -pedantic -fPIC polyregression.c; gcc -o polyregression.so -shared polyregression.o
gcc -c  -Wall -O2 -ansi -pedantic -fPIC cbspline.c; gcc -o cbspline.so -shared cbspline.o
python -m numpy.f2py -c -m spl_int spl_int.f
cd $SERVAL/src/BarCor
gfortran bary.f -o bary.e

A first try to check whether there are any conflicts. It should list all available options:

$SERVAL/src/serval.py --help

If you have a ~/bin folder, a useful shortcut is:

ln -s $SERVAL/src/serval.py ~/bin/serval
ln -s $SERVAL/src/srv.py ~/bin/srv

Otherwise, an alias can be create and included in ~/.bashrc.

alias serval=$SERVAL/src/serval.py

and you can run it as

serval --help

SERVAL in action

A basic example is:

mkdir data
(cd data; git clone https://github.com/mzechmeister/HARPS.git)
serval gj699 data/HARPS/gj699/ -inst HARPS -targ gj699

-targ requests the coordinates from simbad (otherwise RA and DEC from fits header is used)

After serval has finished, you can inspect the results with srv.py, for instance

srv gj699 -rv -x

Tip

You may want to include the following lines in your ~/.gnuplot:

set colors classic
load "~/mzechmeister/python/zoom.gnu"

In gnuplot 5 this uses the old color scheme from gnuplot 4.

And zoom.gnu gives you some additional features, like pan and zoom with keyboard and arrows keys. In particular, very useful for the look and lookt options to explore the spectra.

Further tips are giving in the wiki.