simstack

An open-source, user-friendly, Python version of SIMSTACK.

As of 1/28/2017 SIMSTACK got easier to use! After setting local paths in your .bashrc (or .bash_profile), edit the parameter file (use example.cfg as a guide) to define your maps, catalogs, and pickled outputs (examples below), as well as to choose the binning and method of stacking, and run from the command line with: ../run_simstack_cmd_line.py example.cfg

The python version of the code is object oriented, with map objects (containing maps, noisemaps, beams, color corrections, etc.) and catalog objects (with RA/DEC lists for different selection criteria) collected into dictionaries.

For those familiar with the IDL version, and insist on continuing to use that implimentation of the algorithm, viero_quick_stack is a similar wrapper to the stack_in_redshift_slices function. Note, previous iPython notebooks using this version were accidentally scrubbed during a reordering of the github. If needed I can go through backups, just ask.

Getting Started

There are just a few steps needed to get started:

• Make sure all dependencies are installed (see next section)
• Set your paths in the .bashrc (or .bash_profile)
• Edit the parameter file to specify
  • paths and names of maps/catalog
  • desired binning: style and nodes
  • whether to stack in redshift bins, or all at once
  • map specific details like
    • color corrections
    • beam FWHM (solid angle if converting from MJy/sr to Jy/beam)

Dependencies

• Python 2
• Git repositories
  • simstack
  • utils
  • NDpredict *
  • torrey_cmf *
• astropy
• numpy
• pandas
• lmfit (latest version required or it will fail!)

*Optional repositories if you choose to bin masses by constant number densities.

Local Paths

In your .bashrc define the following, with edits to define your prefered directory locations:

export MAPSPATH=$MAPSPATH/data/maps/
export CATSPATH=$CATSPATH/data/catalogs/
export PICKLESPATH=$PICKLESPATH/data/pickles/

Setting up the Parameter file to run simstack from command line

It's pretty self explainatory (follow example.cfg); the only tricky part is defining the section titled [populations], which sets the the galaxy splitting scheme (and still needs work...) The two recommended options (for now) are:

• sf-qt

  • Splits the sample into star-forming and quiescent according to the Williams UVJ definition. Requires rest-frame U-J and V-J colors, labeled rf_U_V and rf_V_J, respectively.

• general

  • Populations are defined by obeying one condition (for now, ultimately would like no limit on conditions). To ignore condition set it to False.
    Each line defines one population, and takes the form
    • population_key = population_index condition_key greater_than less_than equal_to Ignore conditions by setting them to be False.
      For example:
      • sf = 1 CLASS False False 1 the population 'sf' is given the label 1 in the sfg column of the catalog, and are defined as catalog objects whose CLASS == 1
      • sb = 2 lssfr -8.0 False False the population 'sb' is given the label 2 in the sfg column of the catalog, and are defined as objects whose lssfr < -8.0

  Conditions are set in reverse order, so if, say, a galaxy obeys the conditions of both sf and agn, but agn = 3 and sf = 1, then it will be defined as agn.

Code Output

The command-line simstack routine writes pickled objects to custom-named folders. Tools to read pickles and interpret the objects are under constant developement and improving all the time.
For the moment, an iPython Notebook is provided as an example of how to use the following functions:

• from simstack import PickledStacksReader
  • the workhorse. Uses the info stored in the configuration file (which is rewritten to the output directory) to read in the stacks, determines if they are bootstraps, and if they are then calculates the uncertainties.
• from simstack import measure_cib

Example

The configuration file example.cfg is provided to test simstack. Example maps and catalog called by the configuration file can be downloaded from http://www.astro.caltech.edu/viero/simstack/cosmos/

• maps -- *.cutout.fits
• catalog -- COSMOS2015_Laigle+_Simplified_v1.1.csv

To run, comfirm that the maps and catalog are in the right directories, and simply call: ./run_simstack_cmd_line.py example.cfg

Future Work

Future projects include adding priors from SIMSTACK, and moving from a least-squares solution to marginalized probabilities using a full MCMC.

SEDSTACK is an attempt to fit entire SEDs to the full set of maps -- rather than fitting for flux densities one wavelength at a time -- in order to dig still deeper into the confusion-dominated data. Outstanding issues are dealing with systematic offsets, covariance between maps, careful consideration of beam and pixel size differences and how the affect the chi-squared, and using actual beams instead of Gaussian approximations.

Hopefully, with community input (which I encourage!) the code will grow in utility.

People

• Marco Viero (KIPAC/Stanford)
• Lorenzo Moncelsi (Caltech)
• Jason Sun (Caltech)
• Dongwoo Chung (KIPAC/Stanford)

License, Credits etc

This is work in progress! If you use any of the code or ideas here in your research, please cite us as Viero et al. (2013b).

All content Copyright 2015 the authors. The code is available for use under the MIT license.