This package started with some tools and scripts I use for Lyman-alpha forest analyses, but has since evolved into the mock production and analysis pipeline for the DESI Lyman-alpha BAO measurements. It contains functions and scripts for running the different parts of the mock production process which involves the QuickQuasars script in desisim, and the mock analysis using picca, QSOnic, and/or lya2pt.
- Free software: GPL-3.0 License
Lyatools has similar dependencies with picca, and even directly calls some picca code. Therefore I recommend installing it in the same conda environment with picca. To do this, first follow the installation instructions in picca, and then do:
git clone https://github.com/andreicuceu/lyatools.git
cd lyatools
pip install -e .
Lyatools has a number of scripts (checkout scripts to see all of them), however the most common one is lyatools-run, which requires a configuration file to run. You can find example config files here. To run, simply do:
lyatools-run -i path/to/config.ini
Note that Lyatools has a no_submit option in the config file, which controls the submission of jobs to NERSC. I strongly recommend first setting this to True and running once to check whether all the scripts that Lyatools produces are correct for your desired run.
To run Lyatools, you will also need two environment commands, one for picca, and one for the DESI environment. These could either be a bash function or the name of an alias. For picca, I recommend to add something like this to you bashrc file:
piccaenv () {
module load python
conda activate picca
}
You would then need to set the name of this function in the Lyatools config file: env_command = piccaenv. For the DESI environment, you can do the same if you want to have your own custom environment (see example below), but you can also comment out the desi_env_command option, which will just use the default DESI master branch at NERSC.
Here are instructions to setup the same environment as the one used to run the set of Y1 mocks for Key Project 6 in DESI Y1.
Warning: Make sure your bashrc script does not activate anything automatically. Not even module load python. Otherwise, the scripts may not work in some cases. It is best to keep it clean of direct calls and keep any environment activations in bash functions or in aliases.
First you need to setup your own custom DESI environment. You will need three packages: desisim, specsim, and desimodel. For each of these make note of the path where you cloned these packages, as it will be needed later.
Starting with specsim, do:
git clone https://github.com/desihub/specsim.git
cd specsim
git fetch
git checkout binning
Moving on to desimodel, do:
git clone https://github.com/desihub/desimodel.git
cd desimodel
svn export https://desi.lbl.gov/svn/code/desimodel/tags/0.19.0/data
cd data
cp -r /global/cfs/cdirs/desi/users/dkirkby/desimodel-data/* .
For more details on why these two are needed, see this pull request.
Then moving to desisim, do:
git clone https://github.com/HiramHerrera/desisim.git
cd desisim
git fetch
git checkout y1mocks_v3
Once you have done these three steps, you can setup your DESI environment command. Open your bashrc file and add this function:
mydesienv () {
source /global/common/software/desi/desi_environment.sh master
module unload desimodel
export PYTHONPATH=/path/to/desimodel/py:$PYTHONPATH
export PATH=/path/to/desimodel/bin:$PATH
export DESIMODEL=/path/to/desimodel/
module unload specsim
export PYTHONPATH=/path/to/specsim:$PYTHONPATH
export PATH=/path/to/specsim/bin:$PATH
export SPECSIM=/path/to/specsim
module unload desisim
export PYTHONPATH=/path/to/desisim/py:$PYTHONPATH
export PATH=/path/to/desisim/bin:$PATH
export DESISIM=/path/to/desisim
export DESI_BASIS_TEMPLATES=/global/cfs/cdirs/desi/spectro/templates/basis_templates/v3.2
}
Make sure to replace all the paths in the function above with the path where you installed the three packages. Once this is done, you can now set desi_env_command = mydesienv in the Lyatools config file.
Finally, you will need a picca environment with Lyatools installed. For the first part, just follow the instructions in picca, or if you have a picca environment already, just go to picca, check you are on the master branch, and do git pull. To install Lyatools follow the instructions above, and then put this function in your bashrc file:
piccaenv () {
module load python
conda activate picca
}
and add env_command = piccaenv to your config file.
You are now ready to run DESI Lyman-alpha Y1 mocks. There are two example Y1 config files you can checkout for running the two types of mocks: LyaCoLoRe and Saclay.