ForetFusion

Code to compute coadds for Lya-forest data in BOSS and eBOSS, new gen.

This is a refactoring of the code of the same name, initially put together by Jose Vazquez and Anze Slosar (here github.com/slosar/ForetFusion and github.com/ja-vazquez/ForetFusion). Due to weight of the previously created repo, this is a freshly created repo.

Installation

The package comprises of the ffusion module and a driver in exec/ff_driver.py

First install the module by running

./setup.py install

or to leaving things within your own area on e.g. NERSC:

./setup.py install --user

Then the driver should be run like this, specifying config file

./exec/ff_driver.py config/nersc.ini

Multiprocessing is supporter either by MPI or by multiprocessing module. See config file. Config file format should be fairly self-explanatory.

Output format

Output format is composed of a master fits file and per healpix pixel spectra files.

master file

is a FITS formatted file in root/master.fits with a single table, containing THING_ID, PIX, PLATE, MJD, FIBER. PIX is the healpix pixel (NSIDE in header's NSIDE entry). Data are ordered by PIX, then THING_ID.

pixel files

are FITS fromatted files in root/pixs/pix/pix_###.fits. It contains the following extensions:

• Extension 0 THING_ID_MAP contains a single vector of size Nq for Nq quasars, mapping indices 1..Nq to THING_IDs
• Extension 1 LOGLAM_MAP contains a single vector of size Np for Np pixel, mapping indices 1..Np to log lambda (observed)
• Extension 2 FLUX is a Np rows x Nq columns matrix containing fluxes
• Extension 3 IVAR is a Np rows x Nq columns matrix containing ivars
• Extension 4 ANDMASK is a Np rows x Nq columns matrix containing AND masks
• Extension 5 ORMASK is a Np rows x Nq columns matrix containing OR masks