slosar/foretfusion

Code for merging BOSS and eBOSS data for Lya-forest

ForetFusion

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

• Searching for quasars ======= Code:

  mpirun -np 4 run_catalog.py

The calculations and main functions are located at:

 qso_catalog.py

Python libraries you may need: fitsio, pandas, mpi4py, healpy. I am using Python 3.5.2.

• It goes through the spAll-v5_10_0.fits file and filter all qsos within BOSS, EBOSS (filtering conditions can be easily modified). (the full 12GB file can be read, however to save some time we are using a subset, subset_spAll, with only the columns we are interested on)

Filter Quasars

• Filter the Quasars that satisfy the the bit condition:

  ('Quasars with Bit condition in EBOSS_TARGET1 =', 201742)

  ('Quasars with Bit condition in EBOSS_TARGET0 =', 54108)

  ('Quasars with Bit condition in BOSS_TARGET1 =' , 466098)

  Total qsos = 3008000

• Filter the Quasars that satisfy the condition 'CLASS== "QSO" & (OBJTYPE=="QSO" | ''OBJTYPE=="NA".) & THING_ID != -1'

  ('Quasars with General condition in EBOSS_TARGET1 =', 160854)

  ('Quasars with General condition in EBOSS_TARGET0 =', 38164)

  ('Quasars with General condition in BOSS_TARGET1 =' , 285071)

  Total qsos = 610184

• Filtering by using both conditions, we have:

  ('Quasars with Both condition in EBOSS_TARGET1 =', 160854)

  ('Quasars with Both condition in EBOSS_TARGET0 =', 38164)

  ('Quasars with Both condition in BOSS_TARGET1 = ' , 285071)

  Total qsos = 484089

• The names of the files that contain the 484089 spectra corresponding to these quasars are located at SpAll_files.csv, (use Qsos.write_names = True)

Compute healpix

• Once filtered and selected the Qsos, we compute the healpix given RA and DEC. Then group quasar observations given by healpix. i.e.

Coadd spectra with same THING_ID

• Withing a healpix, finds the repeated objects with the same THING_ID (i.e. self.rep_thid = 4), Coadd them (average ivar*flux.) and compute the chisq with respect to zero flux:

• Eliminate those observations with Chisq< 2 (trim_chisq = 2) to neglect noise, i.e.

Write Fits files for each healpix.

• Write a file for each healpix that contains THING_ID's. For each THING_ID coadd 'flux'. For example, considering the Healpix 24985, there are three different THING_ID with their corresponding coadd, ivar, and_mask and or_mask:

Write Master file

• The master file contains the information from all accepted spectra: healpix, thing_id, mjd, plate, fiberid, etc..

THING_ID distribution

• The figure shows the distribution of THING_ID. The x-axis displays the number spectra per THING_ID, while the y-axis the number of Accepted spectra within a THING_ID. Blue rectangles represent the total of THING_IDs, and the top label the total of spectra. For example, let's consider the case with #Specs per THING_ID = 18. there are two THING_IDs with 18 spectra, one has only 16 accepted and the other has all accepted (18), given a total of 34 accepted spectra.

with Chisq > 0.5

<<<<<<< HEAD

=======

Additional info may be found here

upstream/master