Dockerfile to build containers with Bob Carswell's VPFIT software installed

The voigt profile fitting program VPFIT written by Bob Carswell is a fortran program to fit voigt profiles to absorption features in astronomical spectra.

This dockerfile is to download and install version VPFIT.12.4 along with all its dependencies (PGPLOT, CFITSIO). The dockerfile also modifies the makefile to use the newer fortran compiler gfortran over g77.

Step 0. Clone the repo

git clone https://github.com/tansb/VPFIT_docker.git
cd VPFIT_docker

1. Set up X11 forwarding to XQuartz

VPFIT uses the PGPLOT GUI, so you need some way to forward the display from the container to your actual computer's display. On a mac to allow X11 access:

(a) Open XQuartz and go to Preferences > Security. Make sure "Allow connections from network clients" is enabled.

(b) Also, on your computer run the command xhost + to allow the Docker container to connect to any X11 server.

Step 2. Build the docker container

Docker containers are designed to work and operate in isolation, but when using VPFIT you need an easy way to pass data in and out. Therefore, specify a directory to mount into the container so you can access that data from within the container (and save the reduced files out).

export MY_MOUNTED_DATA_DIR=<your_data_dir>

Create the docker image and build a container

docker compose up -d

Enter the container

docker exec -it vpfit_container bash

First check that the X11 forwarding is working correctly by typing xeyes into the container terminal. A little display with cartoon eyes should pop up in an XQuartz window and follow your cursor around. If not, not sure try step 1 again?

Step 3. Run VPFIT

The container will open directly into your mounted data directory you defined in MY_MOUNTED_DATA_DIR. To run vpfit, just type vpfit. For instructions on running VPFIT, see the user manual written by Bob Carswell

If you want to run it in non-interactive mode, I've written a python script which wraps around VPFIT and runs it on an input file. From within the container do:

python3 run_vpfit_noninteractive.py <path_to_your_input_file.txt>