Specific error obtained with Qhull

[Rekyt]

Here is an issue that aims to be a list of specific errors due to Qhull to better manage #38 in the future.

I obtained the following error in Qhull:

Error: Received error code 5 from qhull. Qhull error:
QH6271 qhull precision error (qh_check_dupridge): wide merge (1122660224434 times wider) due to duplicate ridge with nearly coincident points (0.027) between f102531 and f102515, merge dist 0.093, while processing p2261

• Ignore error with option 'Q12'
• To be fixed in a later version of Qhull
• Vertex distance 0.027 is greater than 100 times maximum distance 8.3e-14
  Please report to bradb@shore.net with steps to reproduce and all output
  ERRONEOUS FACET:
• f102531
  • flags: top new seen mergehorizon dupridge mergeridge2
  • normal: 0.05806 0.1744 -0.5788 0.7752 -0.174
  • offset: 0.1009689
  • vertices: p2261(v1312) p2266(v1046) p2272(v799) p2262(v289) p2265(v160)
  • neighboring facets: f12145 f12159 f102533 f102515 f102532
  • ridges:
  • r74938 tested
    vertices: p2266(v1046) p2272(v799) p2262(v289) p2265(v160)
    between f102531 and f12145
  • r97773
    vertices: p2261(v13

When running the inst/new_benchmark.R script with the following parameter: 50 sites, 5 traits, 200 species, code at c9d85d8 commit.

[Bisaloo]

Somewhat smaller reprexes:

Trimmed down inst/new_benchmark.R

pkgload::load_all()  # Need last version of fundiversity
source(here::here("inst", "benchmark_fcts", "make_data.R"))

n_sites   = c(5e1)

n_traits  = c(5)

n_species = c(2e2)

set.seed(20210906)
given_data = make_data(n_species, n_traits, n_sites)

fd_fric_intersect(given_data[["traits"]], given_data[["site_sp"]])

Actual reprex

given_data <- dget("https://gist.githubusercontent.com/Bisaloo/f3a8042be8a55a887e1f9a2bbcce9a8d/raw/dea950268d971229cd1b91bda47c45f31d418acb/fundiversity_qhull_error.txt")

library(fundiversity)
fd_fric_intersect(given_data[["traits"]], given_data[["site_sp"]])
#> Error in convhulln(ps, "n FA"): Received error code 5 from qhull. Qhull error:
#> QH6271 qhull precision error (qh_check_dupridge): wide merge (798797635164 times wider) due to duplicate ridge with nearly coincident points (0.0055) between f159315 and f159297, merge dist 0.028, while processing p3702
#> - Ignore error with option 'Q12'
#> - To be fixed in a later version of Qhull
#> - Vertex distance 0.0055 is greater than 100 times maximum distance 3.5e-14
#>   Please report to bradb@shore.net with steps to reproduce and all output
#> ERRONEOUS FACET:
#> - f159315
#>     - flags: top new seen mergehorizon dupridge mergeridge2
#>     - normal:   0.04571 -0.002063  -0.0352 -0.08176    0.995
#>     - offset: -0.8682543
#>     - vertices: p3702(v2235) p6412(v1056) p2282(v632) p3731(v13) p4214(v3)
#>     - neighboring facets: f35106 f159297 f159316 f159325
#>     - ridges:
#>      - r36029 tested
#>            vertices: p6412(v1056) p2282(v632) p3731(v13) p4214(v3)
#>            between f159315 and f35106
#>      - r113711
#>            vertices: p3702(v2235) p2282(v632) p

^{Created on 2021-10-26 by the reprex package (v2.0.1.9000)}

I feel like this can be trimmed down even more but it's a good start.

[Rekyt]

Not actually a reprex but:

> fd_benchmark_multicore(fd_fric_intersect)
Running with:
  n_sites n_traits n_species
1     100        4       200
 Error: Received error code 5 from qhull. Qhull error:
QH6271 qhull precision error (qh_check_dupridge): wide merge (10000524017918 times wider) due to duplicate ridge with nearly coincident points (0.002) between f2144 and f2123, merge dist 0.59, while processing p118
- Ignore error with option 'Q12'
- To be fixed in a later version of Qhull
- Vertex distance 0.002 is greater than 100 times maximum distance 5.9e-14
Please report to bradb@shore.net with steps to reproduce and all output
ERRONEOUS FACET:
- f2144
- flags: top new seen mergehorizon dupridge mergeridge2
- normal: 0.1284 -0.5988 0.3437 0.7119
- offset: 0.1695765
- vertices: p118(v130) p128(v108) p347(v52) p125(v0)
- neighboring facets: f1022 f2123 f2145
- ridges:
- r778
vertices: p128(v108) p347(v52) p125(v0)
between f2144 and f1022
- r1071
vertices: p118(v130) p128(v108) p347(v52)
between f1022 and f2144
- r1069
vertices

[Rekyt]

Another case with fd_fric_intersect()

fd_fric_intersect 
Running with:
  n_sites n_traits n_species
1      50        4       500
Error: Received error code 5 from qhull. Qhull error:
QH6271 qhull precision error (qh_check_dupridge): wide merge (6453987413504 times wider) due to duplicate ridge with nearly coincident points (0.0038) between f3040 and f3036, merge dist 0.35, while processing p539
- Ignore error with option 'Q12'
- To be fixed in a later version of Qhull
- Vertex distance 0.0038 is greater than 100 times maximum distance 5.5e-14
  Please report to bradb@shore.net with steps to reproduce and all output
ERRONEOUS FACET:
- f3040
    - flags: top new seen mergehorizon dupridge mergeridge2
    - normal:   -0.7599   0.2833   0.4993   0.3051
    - offset: -0.4838983
    - vertices: p539(v191) p540(v146) p543(v137) p132(v74)
    - neighboring facets: f2198 f3041 f3036
    - ridges:
     - r1112
           vertices: p540(v146) p543(v137) p132(v74)
           between f3040 and f2198
     - r1614
           vertices: p539(v191) p543(v137) p132(v74)
           between f2198 and f3040
     - r1616
           ve

[mluerig]

I get similar errors with fd_fric . An easy solution might be to allow additional options to be passed through to convhulln? Didn't help in my case, but this would make this error message more meaningful if we could actually try their suggestion...

[Rekyt]

Dear @mluerig, thanks for your feedback!

Indeed, we could try to follow convhulln suggestions by letting the user passing additional options.
Could you share a reprex and the error you got and maybe the dataset that caused the error?

[mluerig]

I currently can’t share the data as it is not yet published, but I’m happy to do so once it is. Here is the error-message though:

You may be able to avoid the following error by adding  "Qbb" to the "options" argument of convhulln()
For example: options="FA  Qt Qbb"
If this does not work, please add to the issue at
https://github.com/davidcsterratt/geometry/issues/58

Error in doTryCatch(return(expr), name, parentenv, handler) : 
  Received error code 8 from qhull. Qhull error:
QH6271 qhull topology error (qh_check_dupridge): wide merge (60568613.7x wider) due to dupridge between f247851 and f247847 (vertex dist 4.4e-11), merge dist 1.8e-05, while processing p344
- Allow error with option 'Q12'
ERRONEOUS FACET:
- f247851
    - flags: bottom newfacet dupridge mergeridge2
    - normal:   -0.3481   0.6381   0.6792 -0.02768  0.07535  0.06125  -0.0136
    - offset: -0.0009814611
    - vertices: p344(v219) p1(v193) p187(v181) p367(v114) p695(v110) p392(v73) p492(v49)
    - neighboring facets: f197505 f247646 f247925 f247844 f248096 f247847 f247848
    - ridges:
     - r381 tested simplicialbot
           vertices: p1(v193) p187(v181) p367(v114) p695(v110) p392(v73) p492(v49)
           between f197505 and f247851
     - r682 simplicialtop simplicialbot
           vertices: p344(v219) p187(v181) p367(v114) p695(v110) p392(v73) p492(v49)
           between f247851 and f247646
     - r683 simplicialtop simplicial

[Rekyt]

Dear @mluerig,

Thank you for sharing the error!
I'll probably work on letting the user pass options to convhulln as an easy fix.