galpy.df.StreamTrack.cov¶
- StreamTrack.cov(tp, basis='galcenrect', ro=None, vo=None, use_physical=None, quantity=None)[source]¶
Return the 6x6 covariance matrix of the particle distribution at tp.
- Parameters:
tp (float or array) – Progenitor time coordinate(s).
basis (str, optional) –
Output basis for the covariance matrix. One of: -
"galcenrect"(default):(x, y, z, vx, vy, vz)ingalactocentric Cartesian. Stored natively by the fit.
"galcencyl":(R, vR, vT, z, vz, phi)galactocentric cylindrical."galsky":(ll, bb, dist, pmll, pmbb, vlos)heliocentric Galactic sky + distance + kinematics."sky":(ra, dec, dist, pmra, pmdec, vlos)heliocentric equatorial sky + distance + kinematics."customsky":(phi1, phi2, dist, pmphi1, pmphi2, vlos)sky rotated bycustom_sky_transform. Requirescustom_sky_transformto have been set.
ro (float or Quantity, optional) – Distance scale for physical-units output (overrides the stored value). Mirrors the per-call override on the mean accessors.
vo (float or Quantity, optional) – Velocity scale for physical-units output (overrides the stored value).
use_physical (bool, optional) – Override the object-wide default (
_roSetand_voSet) for whether to scale the covariance to physical units. WhenFalse, the covariance is returned in galpy internal units for galactocentric bases; sky bases retain the prior behavior of running the Jacobian with the storedro/vodefaults (output remains unitless in the internal sense, useful only as a relative shape).quantity (bool, optional) – Not supported: a covariance 6x6 has heterogeneous units across entries (e.g. kpc² for position-position, kpc·km/s for position-velocity, deg² for sky-sky) and can’t be wrapped as a single astropy
Quantity. PassingTrueraisesNotImplementedError.the (Non-galcenrect bases are computed by analytical Jacobian of)
no (coord chain evaluated at the track mean (C' = J · C · Jᵀ) —)
rotations (finite differences. Pure rotations (e.g. the tangent)
a (between sky frames) are inverted via transpose; lbd↔XYZ is)
by (non-rotation transform so its 6x6 block is inverted)
numpy.linalg.inv.
(kpc (Units match the mean-track accessors)
km/s
deg
mas/yr)
otherwise. (when physical output is on; galpy internal units)