ConKer

*** DEV ***

Setup

Clone this repository to your local machine:

git clone https://github.com/mishtak00/conker.git

Get inside the conker directory and install requirements:

cd conker
pip install -r requirements.txt

Help

Get all available cmd line arguments with explanations:

python driver.py --help

python calibrate.py --help

Calibration

For best binning results, please always run a calibration for any particular combination of randoms file, grid spacing and correlation scanning range:

python calibrate.py random0_DR12v5_CMASS_South.fits --scan 60 140

Note that a Correlator will automatically search for and load the correct calibration file for the current run from the calibration folder. Changing calibration filenames manually will result in undefined behavior when a Correlator goes to read them.

If Correlator cannot open an existing calibration file, then it will default to calibration file "calib_ref_random0_DR12v5_CMASS_South_gs_5.npy" that comes with this repo.

Rules on input and randoms catalogs

There are 4 arguments for specifying the inputs. All inputs have to be in the 'data' folder.

• Positional (obligatory) argument for input catalog 1

python driver.py filename1.fits

• Optional argument for input catalog 0. Defining this argument requests a cross-correlation between input catalogs 1 and 0. Omitting this will cause ConKer to calculate an auto-correlation based solely on catalog 1.

-f0 filename0.fits OR --file0 filename0.fits

• Optional argument for randoms catalog. Defining this argument requests that the randoms grid (pre-convolution background) be calculated from histogramming and factorizing this randoms catalog. Omitting this will cause ConKer to calculate the background grid from input catalog 1.
• It's a good idea to save the randoms grid for a certain catalog at a specific grid spacing so that it's only calculated once. Saved in the output folder and has to be manually put in the 'data' folder if it's to be used as input for future correlations.

--randoms_file filenameR.fits --save_randoms

Optional argument for specifying a pre-made randoms grid. Putting this argument requests that the randoms grid be loaded from the 'data' folder and used as is. This randoms grid has to have been made from the same catalog as the input catalog and its grid spacing must be the same as the one in the parameters file for current run. Undefined behavior if otherwise.

--randoms_grid gridR_filenameR_gs_5.npy

Add below argument to request that the randoms catalog be factorized:

--factorize_randoms

Outputs

The requested correlation and its related separation distance will be saved in .npy format in a folder named 'out_f1_FILENAME1_...'. If it's a scan of a diagonal correlation, a plot will be saved in the 'plots' folder.

Rules on boundaries

• All boundaries will be decided from input catalog 0, i.e. they will override the boundaries from randoms catalog and input catalog 1.
• In the case of a separate randoms catalog, only those randoms within file 0 boundaries will be considered in the grid making processes.
• In the case of a cross-correlation, i.e. 2 different input catalogs, the 2 input catalogs will be made to have the same boundaries by overriding file 1 boundaries with those from file 0.

Various runs

Run a nondiagonal 3rd order auto-correlation on unweighted filename1 using randoms from filenameR, save them to an unfactorized randoms grid, fill up the 2D correlation function with axes from 60 to 145 Mpch-1 at steps of grid_spacing (which defaults at 5 Mpch-1):

python driver.py filename1.fits -v -fR filenameR.fits --save_randoms -r0 5 --scan 60 145 --nondiag -n 3

Run the above but with randoms from a premade grid:

python driver.py filename1.fits -v -gR gridR_filenameR_gs_5.npy -r0 5 --scan 60 145 --nondiag -n 3

Run diagonal 2pcf scan on weighted filename1 from 60 to 145 Mpch-1 (both inclusive) in steps of grid_spacing, with weighted randoms read from filenameR, also factorize and save them, with progress logs on std out (-v):

python driver.py filename1.fits -fR filenameR.fits --factorize_randoms --save_randoms -r0 5 --scan 60 145 -v -w1 -wR

Run diagonal 5pcf scan at K0 radius = 5 Mpch-1 and K1 radius from 60 (incusive) to 145 (exclusive) Mpch-1 in steps of grid_spacing, saves just the resulting correlation and separation arrays in the output folder:

python driver.py filename1.fits -fR filenameR.fits -v -r0 5 --scan 60 145 -n 5

Run single 2pcf calculation at K1 radius = 110 and K0 radius = 5:

python driver.py mock_cmassDR9_north_3001.fits -v -r0 5 -r1 110