Search ¶
Neighborhood: radius is in meters along the major axis of the
search ellipsoid (rotation azimuth, dip, rake and ratios semi/major,
minor/major), or of the variogram's anisotropy when no ellipsoid is given.
octant takes at most max_samples / 8 samples from each octant around
the target, split along the ellipsoid's axes; with 2D data (one elevation)
the sectors are the ellipse's four quadrants, max_samples / 4 each.
sectors instead splits the plane of the ellipsoid's major and
semi-major axes into that many equal angles, the first starting at the
major axis, and takes at most max_per_sector samples from each
(max_samples / sectors rounded up by default).
high_grade, a HighGrade or (threshold, radius), restricts samples
above threshold to targets within radius, measured in the same
ellipsoid unless the HighGrade has its own, in estimation and
cross-validation alike; give each pass its own to vary it by pass. The threshold is always in data
units: simulators compare it with the data values and the simulated
values of nodes, so it picks the same samples as in estimation. Estimators also take a sequence
of searches as passes: targets one leaves unestimated go to the next.
soft lets samples of another domain inform a target strictly within a
distance in the same ellipsoid: one distance for every pair of domains,
or a dict {(target_domain, sample_domain): distance}, one way; pairs not
listed are hard.
target_slope or target_efficiency (at most one) calibrate the kriging
estimators per target: each takes the fewest samples from min_samples
whose slope of regression or kriging efficiency reaches it, and
max_samples where none does. With octant the fewer samples are taken
from the sectors in turn.
length_unit
property
¶
length_unit
Length unit of the radius and of the variogram's ranges used with it; None when not declared.
rotation
property
¶
rotation
Azimuth, dip and rake of the search ellipsoid; None when the search follows the variogram's anisotropy.
soft
property
¶
soft
The soft distance, or the dict of distances by domain pair; None when domain boundaries are hard.
target_efficiency
property
¶
target_efficiency
Kriging efficiency the search is calibrated to; None when it is not.
target_slope
property
¶
target_slope
Slope of regression the search is calibrated to; None when it is not.
from_json
staticmethod
¶
from_json(text)
Reads to_json output; raises InvalidInput on another class's JSON
or a newer format.
used in