Opened 12 months ago

Closed 11 months ago

#29568 closed defect (fixed)

Initialize polyhedra with backend cdd over RDF with both Vrepresentation and Hrepresentation

Reported by: gh-kliem Owned by:
Priority: major Milestone: sage-9.2
Component: geometry Keywords: inexact polyhedra, precomputed data
Cc: jipilab, gh-LaisRast Merged in:
Authors: Jonathan Kliem Reviewers: Jean-Philippe Labbé
Report Upstream: N/A Work issues:
Branch: 7e5ecaf (Commits, GitHub, GitLab) Commit: 7e5ecaf79d9a90d36e4b83cd58846dcd6a13365e
Dependencies: Stopgaps:

Status badges

Description (last modified by gh-kliem)

Setting up polyhedra with the shortest representation is not always the best choice for inexact polyhedra. Concretely, #29324 introduced a regression: 

sage: P = polytopes.buckyball(exact=False) 
sage: P+P.center()
/srv/public/kliem/sage/local/lib/python3.7/site-packages/sage/geometry/polyhedron/backend_cdd.py:142: UserWarning: This polyhedron data is numerically complicated; cdd could not convert between the inexact V and H representation without loss of data. The resulting object might show inconsistencies.
  warn("This polyhedron data is numerically complicated; cdd could not convert between the inexact V and H representation without loss of data. The resulting object might show inconsistencies.")
A 3-dimensional polyhedron in RDF^3 defined as the convex hull of 60 vertices

On the other hand 2*P always threw this warning (also before #29200).

We use that we are given both Vrepresentation and Hrepresentation to use whichever works. It works if the calculated object has the correct length of Vrepresentation and Hrepresentation (that's the standard for cdd).

Change History (6)

comment:1 Changed 12 months ago by gh-kliem

  • Branch set to public/29568
  • Commit set to aa5ef80e50468ab04c644e59f3550a35e403fe02
  • Description modified (diff)
  • Summary changed from Initialize polyhedra with backend cdd with both Vrepresentation and Hrepresentation to Initialize polyhedra with backend cdd over RDF with both Vrepresentation and Hrepresentation

New commits:

​aa5ef80initalize cdd with both Hrep and Vrep to deal with numerical inconsistency

comment:2 Changed 12 months ago by gh-kliem

  • Status changed from new to needs_review

comment:3 Changed 11 months ago by gh-kliem

  • Branch changed from public/29568 to public/29568-reb
  • Commit changed from aa5ef80e50468ab04c644e59f3550a35e403fe02 to 7e5ecaf79d9a90d36e4b83cd58846dcd6a13365e

New commits:

​7e5ecafinitalize cdd with both Hrep and Vrep to deal with numerical inconsistency

comment:4 Changed 11 months ago by jipilab

  • Authors set to Jonathan Kliem

Missing Author.

comment:5 Changed 11 months ago by jipilab

  • Reviewers set to Jean-Philippe Labbé
  • Status changed from needs_review to positive_review

Looks good to me.

comment:6 Changed 11 months ago by vbraun

  • Branch changed from public/29568-reb to 7e5ecaf79d9a90d36e4b83cd58846dcd6a13365e
  • Resolution set to fixed
  • Status changed from positive_review to closed
Note: See TracTickets for help on using tickets.