GPROBE(1)

💖 GPROBE(1) — like, the manual page but make it make sense 💖

okay so bestie, welcome to gprobe, a User Commands girly from section 1 of the manual. She's a demo, a show-off, a main character — her whole thing is flexing what the gage library can do. No cap. 🎀

✨ NAME ✨

gprobe — shows off the functionality of the gage library 🫶

✨ SYNOPSIS ✨

gprobe -i <nin> -k <kind> [-v <verbosity>] -q <query> [-gmc
<min gradmag>] [-ofs] [-seed <N>] [-c <bool>] [-k00 <kern00>] [-k11
<kern11>] [-k22 <kern22>] [-rn] [-ssn <SS #>] [-ssr <scale range>]
[-kssb <kernel>] [-kssr <kernel>] [-ssu] [-sso] [-ssnd] [-ssnb <bias>]
[-ssf <SS read/save format>] [-s <sclX sclY sxlZ>] [-ssp <pos>] [-pg
<nrrd>] [-pi <nrrd>] [-pp <pos ...>] [-eps <epsilon>] [-psi <p>] [-t
<type>] [-o <nout>]

✨ DESCRIPTION ✨

So here's the vibe: gprobe shows off the functionality of the gage library. It calls gageProbe() to go interrogate all different kinds of volumes, and in return it learns a bunch of measured or derived quantities. Like, it asks the volume questions and the volume ANSWERS. Iconic behavior honestly. 🌸

✨ OPTIONS ✨

Every single one of these, bestie. We're not skipping. Periodt. 💅

• -i <nin> — the input volume. Your source material, your muse.
• -k <kind> — the “kind” of volume, and it's gotta be one of "scalar", "vector", "tensor", or "dwi". Pick your fighter.
• -v <verbosity> — verbosity level (int); default: "1". Crank it up if you want her to yap.
• -q <query> — the quantity (scalar, vector, or matrix) to learn by probing (string). This is THE question you're asking.
• -gmc <min gradmag> — for curvature-based queries, use zero when gradient magnitude is below this (double); default: "0.0". Below this threshold? Zero. Done.
• -ofs — if only per-axis spacing is available, use that to contrive full orientation info. Making it work with what you've got, honestly relatable.
• -seed <N> — RNG seed; mostly for debugging (unsigned int); default: "42". Yes, 42. Of course.
• -c <bool> — clamp positions as part of probing (bool); default: "false". Not clamped by default, she's free-range.
• -k00 <kern00> — kernel for gageKernel00; default: "tent". ⛺
• -k11 <kern11> — kernel for gageKernel11; default: "cubicd:1,0".
• -k22 <kern22> — kernel for gageKernel22; default: "cubicdd:1,0".
• -rn — renormalize kernel weights at each new sample location. Important tea: “Accurate” kernels don't need this, and doing it always makes things go slower. Always. Slower. You've been warned. 😭
• -ssn <SS #> — how many scale-space samples to evaluate, or, 0 to turn-off all scale-space behavior (unsigned int); default: "0". Zero means she's OFF, okay?
• -ssr <scale range> — range of scales in scale-space (2 doubles); default: "nan nan". Two doubles, one vibe.
• -kssb <kernel> — blurring kernel, to sample scale space; default: "dgauss:1,5".
• -kssr <kernel> — kernel for reconstructing from scale space samples "hermite". 🧚
• -ssu — do uniform samples along sigma, and not (by default) samples according to the effective diffusion scale. Uniform sigma sampling, as a treat.
• -sso — if not using "-ssu", use pre-computed optimal sigmas when possible. Optimal. Efficient. Slay.
• -ssnd — normalize derivatives by scale. Derivative normalization arc.
• -ssnb <bias> — bias on scale-based derivative normalization (double); default: "0.0".
• -ssf <SS read/save format> — a printf(3)-style format (including a "%u") for the filenames from which to read pre-blurred volumes computed for the stack, if they exist and match the stack parameters, and where to save them if they had to be re-computed. Leave this as empty string to disable this. (string); default: "".
• -s <sclX sclY sxlZ> — scaling factor for resampling on each axis (>1.0: supersampling); use "-ssp" (and "-psi") to specify scale position of sampling (3 doubles); default: "1 1 1". Note the third one is literally spelled sxlZ in the synopsis — we don't fix, we report. 🫡
• -ssp <pos> — when using scale-space, scale-position at which to probe (double); default: "0".
• -pg <nrrd> — overrides "-s": a 2-D nrrd which specifies origin and direction vectors for sampling grid; default: "".
• -pi <nrrd> — overrides "-pv": probes at this list of 3-vec or 4-vec positions; default: "". (Yes it says -pv. Yes that's what the page says. Moving on. 💅)
• -pp <pos ...> — overrides "-pi": only sample at this specified point (3..4 doubles); default: "nan nan".
• -eps <epsilon> — if non-zero, and if query is a scalar, and if using "-pp" to query at a single point, then do epsilon offset probes to calculate discrete differences, to find the numerical gradient and hessian (for debugging) (double); default: "0". That's a lot of conditions bestie. Read it twice.
• -psi <p> — whether the probe location specification (by any of the four previous flags) are in index space (bool); default: "false". Index space? Bool. Default: no.
• -t <type> — type of output volume (type); default: "float".
• -o <nout> — output volume (string); default: "-". And - means stdout, iykyk.

✨ SEE ALSO ✨

Go read her friends, they're all iconic: ilk(1), miter(1), mrender(1), nrrdSanity(1), overrgb(1), puller(1), tend(1), unu(1), vprobe(1) 🦋

and that's the whole page, bestie!! April 2021, User Commands, GPROBE(1) — mwah, go probe something gorgeous 💖✨🎀🫶