Lightcone Mode¶

Lightcone Mode opens a LightSAGE cli_lightcone HDF5 output — a flat catalogue of galaxies spanning a range of redshifts along an observer's line of sight — in the exact same Explore UI as a SAGE box: same toolbar, navigation panel (every colour-by mode), info panel, and gaussian-splat rendering.

See Launch Mode → Setting up a run for how to build a lightcone from a SAGE26 model.

Opening a lightcone¶

visage --lightcone /path/to/lightcone.h5

Or, in Launch Mode, click Open in Lightcone view when a build finishes, or pick it under LightSAGE's Load existing or the home page's Load a previous model. Or pick it from the Session Models list (see below) if you've opened it before. The page shows Opening … and reloads itself when ViSAGE is ready.

What gets loaded¶

Every SAGE field carried in the flat lightcone file is read into a full galaxy snapshot — so every Structure tab colour-by mode (stellar mass, sSFR, cold gas, B/T, …) works exactly as it does for a SAGE box. Host haloes are built from the galaxies with Type == 0 (centrals).

A lightcone is a single static point cloud (it already spans every redshift at once), not a per-snapshot simulation — so there's no background snapshot preloading and no per-model output-directory scan.

The snapshot slider is a redshift/time cut¶

Unlike a SAGE box, where the slider selects one snapshot to display, in Lightcone Mode the slider:

only spans the snapshots actually present in the cone (not the full 0–63 range of the underlying simulation)
removes galaxies nearer than the selected point (lower redshift), keeping everything farther away (higher redshift)
shows the full cone when pulled to its maximum (the default on load)

Drag it inward from the maximum to peel away the near side of the cone and see what's left at higher redshift.

Synthetic photometry (SED)¶

If the lightcone was built with photometry (see Launch Mode → Setting up a run), Lightcone Mode replaces the Box tab with a Photometry tab (the box and slice controls move to the top of the Coords tab: X/Y/Z bounds, Draw Box, Use Current View, Zoom; a thin box makes a slice, and the bounds start at the cone's own extent) — a false-colour image builder.

Photometry is its own independent layer, separate from the galaxies, with its own Visible checkbox and Opacity slider. It's hidden by default; ticking Visible shows it and hides the normal galaxies by default so the image reads cleanly (unticking it brings them back). Because it's a separate layer you can view it however you like — re-enable the galaxies (Structure tab) to overlay both, or keep them off for the photometry image on its own. Galaxies with no flux in the stacked filters are simply not drawn.

When shown, it draws the galaxies as Rvir-sized gaussian splats painted from a stack of the filters you tick in the Filters dropdown. Filters are named by survey ("SDSS g", "2MASS Ks", "IRAC 3.6 µm") and listed blue to red, observed frame first; SDSS r is the default. Each filter is tinted a representative colour — UV/u violet, g green, r/i/z and the near-/mid-IR bands (J, H, K, WISE) deepening reds — and scaled by the galaxy's brightness in that band; the ticked filters are summed into an additive false-colour composite. Tick one filter for a single-band image, or several to build a multi-band colour image; mass-to-light (M*/L) entries can be stacked the same way. A colour-swatch legend shows the current stack.

The Photometry tab only appears for lightcones that actually carry SED data (otherwise it shows a short note explaining how to produce it).

Mock images¶

Create mock image (under Image in the Photometry tab) makes a survey image of the sky around the selected galaxy, or the cone's brightest galaxy if none is selected: what any of the LightSAGE surveys would record. Only surveys whose filters the cone has are offered. Choose the field (½′ to 30′, kept under 2400 pixels a side) and whether to add the survey's sky noise. Each galaxy is drawn as its disk and bulge at its own size, inclination and colours, blurred by the survey's seeing on its pixel grid. The image opens in a Library card and is saved in the session folder as a PNG and a FITS file (one band per extension, in nanomaggies, with a WCS). Only the cone's own galaxies are on the sky, so a shallow cone looks sparser than the real sky. Galaxies beyond this cone adds those from a deeper cone in a neighbouring folder, with the same bands, built from the same box. There are no stars: SAGE26 doesn't make them. The method and survey numbers are in Lightcone photometry → Mock images.

HD galaxies (a checkbox under the field size) draws galaxies big enough on the sky to show structure from their HD models instead of smooth profiles: the selected galaxy and up to 12 more, the brightest at least a PSF across. Each is seen from the observer at its angular size, with HD's three colour channels scaled to the galaxy's own magnitude in each of the survey's three bands, then blurred and given noise as the rest. Optical surveys only (SDSS, DES, Pan-STARRS1, LSST): HD's colours are optical. HD sizes follow the observed size–mass relation (Shen et al. 2003), so an HD galaxy can look smaller than the same galaxy's profile, which uses SAGE's disk radius.

Spectra¶

Make spectrum (in the Photometry tab) makes the selected galaxy's spectrum as it would be observed: f_ν in µJy against observed wavelength, with rest wavelength along the top. It is made the way the cone's photometry was, with the same settings recorded in the file (IMF, recycling, metallicity, dust and its infrared glow, emission lines), so the catalogue's own magnitudes, marked as circles, sit on it. The stars' light without dust and the disk and bulge (when the cone has them) are drawn too. It passes through the intergalactic medium and starts at the rest-frame Lyman limit. The plot opens in a Library card and is saved in sage_library/spectra/ as a PNG and a CSV table (rest and observed wavelength, the flux of each part, and the rest-frame L_ν). A card shows what is being done and how far it has got, with a Cancel button. The first spectrum at each stellar metallicity takes about a minute while FSPS builds its stellar populations; later ones in the same session are immediate. Spectra are made in a separate worker process, so the viewer stays responsive while one builds. The cone needs its star-formation histories (SFHMassDisk/SFHMassBulge) and the run's scale-factor list on disk.

Camera¶

Resetting the camera frames the cone horizontally, centred and zoomed in, showing it end to end (rather than the cubic-box framing used for a SAGE box) — this is also the initial view on load, framed to your window once the page has opened.

The go-to-centre button (▶ centre icon) doesn't use a box centre in Lightcone Mode — instead it stands you at the observer (the coordinate origin) looking outward along the cone, with the sky spread horizontally, as if gazing out into the lightcone.

Session models¶

The Launch-Mode dropdown (SAGE-logo button, top-left of the toolbar) lists every box and lightcone opened so far this session under Session Models, with a box or telescope icon marking the kind and the currently active one highlighted. Click any entry to jump straight back to it — a quick relaunch on the same port — so loading a SAGE box after a lightcone (or vice versa) never loses track of what you had open.

This list is persisted across relaunches in ~/.visage/session_models.json.

Things that don't apply in Lightcone Mode¶

A lightcone has no periodic box and isn't tied to a single SAGE parameter file, so a few Explore-mode features that assume a box don't apply:

Multi-model overlay / side-by-side comparison (a lightcone is always the sole loaded model)
The .par-derived cosmology display (cosmology is read from the lightcone's copied SAGE header instead)

Survey catalogues¶

The Survey view in the Photometry tab selects the galaxies a real survey would from the cone, and exports them as FITS or CSV with Petrosian and fibre magnitudes in every band with disk and bulge photometry. python -m visage.sed.survey does the same from the command line. A cone that stops short of a survey's depth gives a lower limit on its counts: GAMA needs z ≈ 0.5.

Exports also give each galaxy's group and group_n, from the group finder of Van Kempen et al. (2024, 2025), and the note scores those groups against the cone's true haloes.

Mock spectroscopy¶

python -m visage.sed.spectroscopy --input cone.h5 --survey sdss_main --output spectra.fits

Make fibre spectra in the Survey view runs the same thing in its own process. This writes SDSS Main fibre spectra and what's measured from them. The first extension is the catalogue; FLUX, LOGLAM and SIGMA hold the spectra. About 2,000 galaxies take a few minutes. A fibre mixes SAGE26's disk and bulge, so it can't show gradients within either; noise comes from real faint SDSS spectra, sky lines included, without each galaxy's own shot noise; and Balmer lines aren't corrected for stellar absorption.