DAP Integration Tools

The CubeGen.daptools module provides utilities for interfacing reconstructed cubes with downstream analysis tools, particularly the Data Analysis Pipeline (DAP) used by the LVM survey.

RSS Extraction from Cubes

Module: CubeGen.daptools.rss_gen

The rssp_extract() function converts a reconstructed 3D cube back into a pseudo-RSS (Row-Stacked Spectra) format. This is used when the downstream analysis pipeline expects RSS-format input rather than 3D cubes.

from CubeGen.daptools.rss_gen import rssp_extract

rssp_extract(
    name='NGC_1234',
    path='./cubes/',
    path_out='./rss/',
    basename_in='lvmCube-NAME.fits.gz',
    basename_out='lvmRSS-NAMElab.fits',
    flu16=True,
    nsplit=0,
)

How it works:

  1. Reads the input cube (flux + error extensions)

  2. Iterates over all spatial pixels (spaxels)

  3. Each non-zero spaxel becomes one “fiber” in the output RSS

  4. Constructs a SLITMAP table with RA/Dec coordinates derived from the cube’s WCS, plus pixel positions for back-referencing

  5. Writes the result as a multi-extension FITS file (see Output Data Products)

Producing RSS output from sincube

The sincube command does not have a cube2rss option – that feature is only available in pipecube. However, rssp_extract() requires only the cube file as input, so it can be called manually after sincube completes:

from CubeGen.daptools.rss_gen import rssp_extract

rssp_extract(
    name='NGC_1234',          # matches the name used in the sincube config
    path='out_cubes/',        # out_path from the sincube config
    path_out='out_rss/',
)

The function reads HDU 0 (flux) and HDU 1 (error) from lvmCube-NGC_1234.fits.gz and writes lvmRSS-NGC_1234.fits.gz to out_rss/.

Important

sincube must be run with errors: True in the config file. When errors are disabled, HDU 1 of the cube is an array of ones, and the IVAR extension of the RSS will be filled with meaningless values (1 / 1.0² = 1 everywhere).

Which RSS extensions contain real data

rssp_extract() works entirely from the cube file. Because the cube does not preserve the original per-fiber sky spectra or LSF, several output extensions cannot be reconstructed and are filled with placeholder values:

HDU

EXTNAME

Status

Content

0

ORIGINAL

Real

Primary header copied from the cube, with POSCIRA/POSCIDE set from the cube’s CRVAL1/CRVAL2. EXPOSURE is hardcoded to 000000.

1

FLUX

Real

Flux spectra extracted spaxel-by-spaxel from the cube HDU 0.

2

IVAR

Real (requires errors: True)

Inverse variance computed as 1 / error² from cube HDU 1. If the cube was generated without error propagation, this extension contains ones and should not be used.

3

MASK

Placeholder

Zero array – no mask information survives cube reconstruction.

4

WAVE

Real

Wavelength array reconstructed from the cube header keywords CRVAL3, CDELT3/CD3_3, and CRPIX3.

5

LSF

Placeholder

Ones array – the line spread function is not stored in the cube.

6

SKY_EAST

Placeholder

Ones array – sky spectra are not preserved in the cube.

7

SKY_EAST_IVAR

Placeholder

Ones array.

8

SKY_WEST

Placeholder

Ones array.

9

SKY_WEST_IVAR

Placeholder

Ones array.

10

SLITMAP

Real

Binary table. ra and dec are computed from the cube WCS via pixel_to_skycoord(). xpmm/ypmm store the pixel column/row in the cube for back-referencing. Bookkeeping columns (targettype, telescope, ifulabel, etc.) are filled with generic placeholder strings ('science', 'Sci', 'Sci1').

This is the same behaviour whether the cube was produced by sincube or pipecube – rssp_extract() is called identically in both cases and uses no information beyond the cube file itself.

Spatial tiling support:

For large cubes that exceed memory limits, the nsplit parameter enables spatial tiling. When nsplit > 1, the cube is divided into an nsplit x nsplit grid, and only the tile specified by spt=[i, j] is processed. The output filename is tagged with the tile index (e.g., lvmRSS-NGC_1234_p01.fits).

1D RSS from external spectra:

The rssp_multi1d() function creates a pseudo-RSS from a collection of independent 1D spectra (not from a cube). It reads a file list, resamples all spectra to a common wavelength grid, and writes the result in the same multi-extension format as rssp_extract(). This is useful for collecting spectra from different sources into a single DAP-compatible file.

DAP Emission Line Tools

Module: CubeGen.daptools.dap_fluxelines

These tools extract emission line measurements from DAP analysis products and map them back onto the 2D spatial grid of the original cube.

extract_values()

Reads emission line fit results from the NP_ELINES_B, NP_ELINES_R, and NP_ELINES_I extensions of a DAP output file. Extracts for each line:

  • Flux and error (flux, e_flux)

  • Velocity and error (vel, e_vel)

  • Velocity dispersion and error (disp, e_disp)

  • Equivalent width and error (EW, e_EW)

extract_PMvalues()

Reads emission line fit results from the PM_ELINES and PM_KEL extensions (parametric and kinematic emission line fits). Extracts flux, velocity, and dispersion for each modeled wavelength.

get_values()

Combines extract_values() and extract_PMvalues() to build complete 2D emission line maps. For each spaxel in the original cube grid, it matches the DAP fiber ID to the RSS SLITMAP coordinates and places the measurements at the correct spatial position.

Returns three map arrays:

  • map_dap – non-parametric line measurements (NP), shape [n_lines * 8, nx, ny] (8 quantities per line)

  • map_dap2 – parametric line measurements (PM), shape [n_models * 6, nx, ny]

  • map_dap3 – kinematic emission line measurements (PM_KEL), shape [n_models * 6, nx, ny]

dap_extract_fluxelines()

High-level wrapper that orchestrates the full extraction workflow:

  1. Reads the cube header to determine spatial dimensions and WCS

  2. Optionally handles spatial tiling (nsplit > 1)

  3. Calls get_values() for each tile or the full cube

  4. Writes the output as lvmDAPMap-<name>-flux_elines.fits.gz with three extensions: NP, PM, and PM_KEL

from CubeGen.daptools.dap_fluxelines import dap_extract_fluxelines

dap_extract_fluxelines(
    name='NGC_1234',
    out_path='./dap_maps/',
    path_cube='./cubes/',
    path_dap='./dap/',
    path_rss='./rss/',
)

The output header contains VAL_xxxx keywords that map each slice of the data cube to the corresponding emission line and measurement type (e.g., flux_Ha_6565, vel_OIII_5008).