Delaunay tesselation failed

[vallsv]

Got a lot logs like the following i Flint, for some reason.

Any idea if it is useful for something? Or if this could be fixed in silx side?

I guess it's related to scatter rendering.

Else, i think ill filter out the message inside Flint.

2022-02-25 00:45:15,378 eh3 19785 silx.gui.plot._utils.delaunay ERROR : Delaunay tesselation failed: QH6154 Qhull precision error: Initial simplex is flat (facet 1 is coplanar with the interior point)

While executing:  | qhull d Qt Qz Qbb Qc Q12
Options selected for Qhull 2019.1.r 2019/06/21:
  run-id 468942845  delaunay  Qtriangulate  Qz-infinity-point  Qbbound-last
  Qcoplanar-keep  Q12-allow-wide  _pre-merge  _zero-centrum  Qinterior-keep
  Pgood  _max-width 4.1  Error-roundoff 3.5e-15  _one-merge 2.4e-14
  Visible-distance 6.9e-15  U-max-coplanar 6.9e-15  Width-outside 1.4e-14
  _wide-facet 4.2e-14  _maxoutside 2.8e-14

The input to qhull appears to be less than 3 dimensional, or a
computation has overflowed.

Qhull could not construct a clearly convex simplex from points:
- p164(v4): -0.46     0   2.5
- p100(v3): 1.9e-05     0     0
- p163(v2):   1.6     0   0.9
- p0(v1):  -2.5     0   2.3

The center point is coplanar with a facet, or a vertex is coplanar
with a neighboring facet.  The maximum round off error for
computing distances is 3.5e-15.  The center point, facets and distances
to the center point are as follows:

center point  -0.3469        0    1.419

facet p100 p163 p0 distance=    0
facet p164 p163 p0 distance=    0
facet p164 p100 p0 distance=    0
facet p164 p100 p163 distance=    0

These points either have a maximum or minimum x-coordinate, or
they maximize the determinant for k coordinates.  Trial points
are first selected from points that maximize a coordinate.

The min and max coordinates for each dimension are:
  0:      -2.5     1.575  difference= 4.075
  1:         0         0  difference=    0
  2:         0       2.5  difference=  2.5

If the input should be full dimensional, you have several options that
may determine an initial simplex:
  - use 'QJ'  to joggle the input and make it full dimensional
  - use 'QbB' to scale the points to the unit cube
  - use 'QR0' to randomly rotate the input for different maximum points
  - use 'Qs'  to search all points for the initial simplex
  - use 'En'  to specify a maximum roundoff error less than 3.5e-15.
  - trace execution with 'T3' to see the determinant for each point.

If the input is lower dimensional:
  - use 'QJ' to joggle the input and make it full dimensional
  - use 'Qbk:0Bk:0' to delete coordinate k from the input.  You should
    pick the coordinate with the least range.  The hull will have the
    correct topology.
  - determine the flat containing the points, rotate the points
    into a coordinate plane, and delete the other coordinates.
  - add one or more points to make the input full dimensional.

[vallsv]

Here is a related log

2022-02-25 03:45:14,854 eh3 19785 silx.gui.plot._utils.delaunay ERROR : Delaunay tesselation failed: QH6214 qhull input error: not enough points(1) to construct initial simplex (need 4)

While executing:  | qhull d Qt Qz Qbb Qc Q12
Options selected for Qhull 2019.1.r 2019/06/21:
  run-id 650441638  delaunay  Qtriangulate  Qz-infinity-point  Qbbound-last
  Qcoplanar-keep  Q12-allow-wide  _pre-merge  _zero-centrum  Qinterior-keep
  _maxoutside  0

[vasole]

Basically all those errors are telling you

• That you cannot build a triangle with aligned points
• That you cannot build a triangle with just one point (or a tetrahedral in 3D). It seems it wants a minimum of four points in the last case.

[t20100]

The data is not checked before calling delaunay.

I see 2 options:

• Change this error log to info since this is expected as it is expected to work with not enough points and return None. https://github.com/silx-kit/silx/blob/234a5f7a7b134b2c67bc3424d9e4facf9b539d45/src/silx/gui/plot/_utils/delaunay.py#L58
• Do some tests before calling delaunay like not enough points (easy), not all aligned (but this is extra computations espically when there is a lot of points... but delaunay is slow anyway), and it won't cover some cases (delaunay computation overflows).

[vallsv]

I think if we can't do much as user, it's better to switch it to debug. But it's up to you.

[t20100]

Yes, as it is implemented, it is kind of "normal" behavior: +1 to turn this long log message to debug. What about adding a short info log so one has a way for knowing that nothing is displayed.

[vallsv]

I don't know. To me it would make sense to return None or raise an error. But i don't exactly know the location nor the implication. Maybe we can take a look together with the code.

Ideally i would prefer to display the stuff with a fallback rendering, but it's probably not so important. I mean, not something we should spend a lot of time with.