Data organization

CAMELS data is organized in a hierarchical way.

Type folders

At the highest level, CAMELS contains different type folders:

  • Sims. This folder contains the raw output from the simulations.

  • FOF_Subfind. This folder contains the Subfind halo and subhalo catalogues.

  • Rockstar. This folder contains the Rockstar halo and subhalo catalogues.

  • Parameters. This folder contains the value of the cosmological, astrophysical, and initial random seed.

  • Caesar. This folder contains the CAESAR halo and subhalo catalogues.

  • AHF. This folder contains the AMIGA Halo Finder (AHF) halo and subhalo catalogues.

  • Pk. This folder contains the power spectra measurements.

  • Bk. This folder contains the bispectra measurements.

  • PDF. This folder contains the PDF measurements.

  • VIDE_Voids. This folder contains the void catalogues.

  • Lya. This folder contains the Lyman-\(\alpha\) spectra.

  • X-rays. This folder contains the X-rays files.

  • Profiles. This folder contains the halo radial profiles.

  • Photometry. This folder contains the photometric catalogues.

  • CMD. This folder contains the CAMELS Multifield Dataset (CMD).

  • SCSAM. This folder contains the data from CAMELS-SAM.

  • SubLink. This folder contains the SubLink merger trees.

  • SubLink_gal. This folder contains the SubLink gal merger trees.

When possible, we have tried to organize the data inside the type folders in a self-similar way.

Suite folders

Inside the considered type folder, there are different suite folders:

Folder

Description

IllustrisTNG

Data from the simulations of the IllustrisTNG suite

IllustrisTNG_DM

N-body counterparts of the IllustrisTNG simulations

SIMBA

Data from the simulations of the SIMBA suite

SIMBA_DM

N-body counterparts of the SIMBA simulations

Astrid

Data from the simulations of the Astrid suite

Astrid_DM

N-body counterparts of the Astrid simulations

Magneticum

Data from the simulations of the Magneticum suite

Magneticum_DM

N-body counterparts of the Magneticum simulations

Swift-EAGLE

Data from the simulations of the Swift-EAGLE suite

Swift-EAGLE_DM

N-body counterparts of the Swift-EAGLE simulations

Ramses

Data from the simulations of the Ramses suite

Ramses_DM

N-body counterparts of the Ramses simulations

CROCODILE

Data from the simulations of the CROCODILE suite

CROCODILE_DM

N-body counterparts of the CROCODILE simulations

Enzo

Data from the simulations of the Enzo suite

Enzo_DM

N-body counterparts of the Enzo simulations

Obsidian

Data from the simulations of the Obsidian suite

Obsidian_DM

N-body counterparts of the Obsidian simulations

Note

For some data products some suite folders may be missing. For instance, Lyman-\(\alpha\) spectra are not generated from N-body simulations. Thus, in that case, only the IllustrisTNG and SIMBA suite folders are present.

Generation folders

Inside a suite folder there are different generation folders:

  • L25n256. This folder contains the data from the simulations with \(256^3\) dark matter particles (plus \(256^3\) initial gas elements if hydrodynamic) in a periodic box of \(25~h^{-1}{\rm Mpc}\) side.

  • L50n512. This folder contains the data from the simulations with \(512^3\) dark matter particles (plus \(512^3\) initial gas elements if hydrodynamic) in a periodic box of \(50~h^{-1}{\rm Mpc}\) side.

  • L100n1024. This folder contains the data from the simulations with \(1024^3\) dark matter particles (plus \(1024^3\) initial gas elements if hydrodynamic) in a periodic box of \(100~h^{-1}{\rm Mpc}\) side.

Note

Note that not all simulations, and its associated data, are available. For instance, as of April 2024, for most of the suites, L25n256 is pretty complete, while L50n512 only contains IllustrisTNG simulations and there are no simulations in the L100n1024 volume for any suite.

Set folders

Inside a volume folder there are several set folders:

  • SB. This folder contains the data from the simulations of the SB set. Typically, we will add a number of this set to identify the number of parameters varied. For instance, SB28 in the case of IllustrisTNG to denote that 28 parameters are varied in a sobol sequence. Inside this folder, there are subfolders named SBN_X, where N is the number of parameters varied (e.g. 28 in the case of IllustrisTNG), and XX ranges from 0 to a multiple of 2 (e.g. 2048 for IllustrisTNG-SB28, 256 for Ramses-SB5…etc).

  • LH. This folder contains the data from the simulations of the LH set. Inside this folder, there are subfolders named LH_X, where where X ranges from 0 to 999.

  • 1P. This folder contains the data from the simulations of the 1P set. Inside this folder, there are subfolders named 1P_pX_, where X ranges from 1 to N while Y goes from n2 (-2) to 2. Where N is the number of parameters (e.g. 28 in IllustrisTNG, 6 in Astrid…etc).

  • CV. This folder contains the data from the simulations of the CV set. Inside this folder, there are subfolders named CV_X, where X ranges from 0 to 26.

  • EX. This folder contains the data from the simulations of the EX set. Inside this folder, there are subfolders named EX_X, where X ranges from 0 to 3.

  • BE. This folder contains the data from the simulations of the BE set. Inside this folder, there are subfolders named BE_X, where X ranges from 0 to 26.

Note

The IllustrisTNG suite folder contains a new set of zoom-in simulations under the simulation folder zoom. See CAMELS-zoomGZ for more details.

As can be seen, the name of the folder can be used to identify the simulation set and its parameters.

Note

The numeric scheme of the 1P set labels was chosen to help the user to identify which parameter and its variation is the one considered. This may be more useful than just listing the simulations from 0 to, e.g., 65. We refer the reader to Parameters for the actual value of the parameters in each simulation.

Actual data

Finally, inside a set folder the user can find the associated data for that particular simulation. We note that these folders can contain multiple files, e.g. the power spectra of the considered simulations at all redshifts.

The image below shows an scheme with the generic data structure for the case of the power spectra:

Generic data structure

Warning

There are some data products that are organized in a different way to the one outlined above. For instance, the CAMELS Multifield Dataset (CMD) follows a different data structure. In these cases we describe in detail the structure of those data products.