The overall order of operations is to build the lsst stack image,
and then use that as a base image to make a MAF iamge.
These instructions will assume you have clone this repository and it is your current working directory.
The stack images can be built from the build_stack.sh script
found in the build_scripts directory in this repo. Without any
command line options this script will build an image with the name:tag,
maf:stack-YYYYMMDD. The date information will automatically be filled in by script.
So if today is October, 4th 2018, running this script will produce a Docker image
called maf:stack-20181004
The stack Docker file includes doing a newinstall and eups distrib install
lsst_sims, so it will take quite a while to finish.
To build the stack imagerun the following command:
build_scripts/buil_stack.sh
When it is finished note the name and tag of the image because it is needed
for the MAFimage.
After the stack image has finished building, open the MAF Docker file found
in docker/maf. You will need to edit the first line of this Docker file to
match the stack image you created in the previous step. Again, assuming today
is October, 4th 2018, the first line of this Docker file will need to be
FROM maf:stack-20181004
Once you have edited and saved the MAF docker file you are read to run
the build_maf.sh script.
build_scripts/buil_maf.sh
Without any command line options this script will build an image with the name:tag,maf:YYYYMMDD. So in the case of our date example this would be
maf:20181004.
This script will run much faster that the build_stack.sh script. The MAF
Docker file is just cloning sims_maf and installing sims_movingObject.
The docker image that results from build_scripts/buil_maf.sh is what you
will want to use when running MAF.
docker run -it maf:20181004
In the bin directory there is a script, that when mounted in a MAF docker image
will run run_all.py from MAF, and save the output to your local machine. For this
example we will assume we have a local directory called maf_local and in the directory
is a run directory and its run database.
mkdir maf_local
cd maf_local
mkdir -p pontus_2556/data
mv pontus_2556.db pontus_2556/data/pontus_2556 (or moved it from where ever you might have it)
Now you need to make sure that the bin/setup_runall.sh script is in the maf_local
directory, so if you did ls you would see the folowing:
pontus_2556/ setup_runall.sh
If this is all setup up correctly you can use a docker run command similar to this
to run the metrics in a non-interactive docker container. In this command replace
/home/oboberg with the full path to your maf_local directory and change oboberg/maf:latest
to the MAF docker image you would like to use.
docker run --rm -v /home/oboberg/maf_local/:/home/docmaf/maf_local \
oboberg/maf:latest \
/home/docmaf/maf_local/setup_runall.sh -r pontus_2556 &