conformallab/ConformalLabpp
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

Phase 8b-Lite extension: output_uv_map named parameter (integrated layout) #14

Closed
user2595 wants to merge 2 commits from feature/output-uv-map into main
pull from: feature/output-uv-map
merge into: conformallab:main
Conversation 0 Commits 2 Files Changed 7 +296 -4

2 Commits

Author SHA1 Message Date
Tarik Moussa
868ba9d3a0 docs: add StereographicUnwrapper + CircleDomainUnwrapper to roadmap
Some checks failed
C++ Tests / test-fast (push) Successful in 1m55s
Details
C++ Tests / test-fast (pull_request) Successful in 2m23s
Details
API Docs / doc-build (pull_request) Successful in 46s
Details
C++ Tests / test-cgal (push) Has been skipped
Details
C++ Tests / test-cgal (pull_request) Failing after 10m53s
Details 
Audit found that 2 of the 4 Java-port candidates from the conformal-
mapping discussion were missing from the documentation:

* StereographicUnwrapper (266 Java LoC) — projects spherical layout
  S² → ℂ via stereographic projection + Möbius centring.  Closes the
  visualisation gap from discrete_conformal_map_spherical() which
  currently returns Point_3 on S²; downstream uses typically want a
  2-D atlas.  Suggested phase: 10b' (alternative methods, parallel
  to Hyperbolic / Quasi-isothermic).  Effort: small (~3 days).

* CircleDomainUnwrapper (570 Java LoC) — conformal map of a
  multiply-connected planar region onto a disk-with-holes (Koebe's
  general uniformization theorem 1909).  A use-case class
  conformallab++ does not currently cover (annulus, slit torus,
  fluid flow around obstacles, electrostatics with multiple
  conductors).  Suggested phase: 11c.  Effort: large (~2 weeks).

Added to all three roadmap documents:

* doc/roadmap/java-parity.md      — worth-porting table extended
* doc/roadmap/research-track.md   — Java-backlog summary extended
* doc/roadmap/phases.md           — Phase 10b' bullet + new
                                    Phase 11c block with full math
                                    context (Koebe 1909 reference,
                                    classical complex-analysis use cases).

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-22 13:12:22 +02:00
Tarik Moussa
15f6baa00b Phase 8b-Lite extension: output_uv_map named parameter for integrated layout
Some checks failed
C++ Tests / test-fast (push) Successful in 2m8s
Details
C++ Tests / test-fast (pull_request) Successful in 2m20s
Details
API Docs / doc-build (pull_request) Successful in 47s
Details
C++ Tests / test-cgal (push) Has been skipped
Details
C++ Tests / test-cgal (pull_request) Failing after 10m45s
Details 
Closes the UX gap identified in the Phase-8b-Lite design discussion:
the four classical-DCE CGAL entries now optionally run their `*_layout()`
step internally if the caller supplies `CGAL::parameters::output_uv_map(pmap)`.

Before this PR, the workflow was:

    auto res = CGAL::discrete_conformal_map_euclidean(mesh);
    // ... user has to re-set up maps, re-pin vertex, then ...
    auto layout = euclidean_layout(mesh, res.x, maps);
    // ... and copy coordinates into a property map manually.

After this PR:

    auto uv = mesh.add_property_map<Vertex_index, K::Point_2>("uv", ...).first;
    CGAL::discrete_conformal_map_euclidean(
        mesh, CGAL::parameters::output_uv_map(uv));
    // ... uv now populated for every vertex.

Coverage
────────
* `discrete_conformal_map_euclidean`  — `Point_2` per vertex.
* `discrete_conformal_map_spherical`  — `Point_3` per vertex (on S²).
* `discrete_conformal_map_hyper_ideal` — `Point_2` per vertex (Poincaré disk).

CP-Euclidean and Inversive-Distance entries do not yet support
`output_uv_map` — face-based packing has no per-vertex UV concept, and
inversive-distance needs a dedicated layout routine that uses Luo's
edge-length formula (planned follow-up).

New named parameters (`code/include/CGAL/Conformal_map/internal/parameters.h`)
─────────────────────────────────────────────────────────────────────────────
* `output_uv_map(pmap)` — write coordinates into pmap after layout.
* `normalise_layout(bool)` — apply post-layout canonical normalisation
  (PCA centroid for Euclidean, north-pole alignment for Spherical,
  Möbius centring for Hyper-ideal).

Both follow the existing Phase-8a-MVP named-parameter convention.
Chained syntax (`.output_uv_map(...).normalise_layout(true)`) is not
yet supported — pass them one at a time.

Tests (5 new in test_cgal_phase8b_lite.cpp)
───────────────────────────────────────────
* `OutputUvMap_Euclidean_PopulatesPmap`         — UVs are finite + non-trivial.
* `OutputUvMap_Spherical_PopulatesXyz`          — every output on unit S².
* `OutputUvMap_HyperIdeal_PointsInPoincareDisk` — |p|² ≤ 1 if converged.
* `OutputUvMap_Absent_DoesNotRunLayout`         — no parameter ⇒ no layout.
* `OutputUvMap_NormaliseLayout_TakesEffect`     — both raw + norm calls
  return finite UVs (named-parameter chaining limitation documented).

All five pass.  Full CGAL suite: 232/232, 0 skipped (was 227).

doc/api/tests.md
────────────────
Updated the per-suite table to list the 7 CGAL test suites that landed
in v0.9.0 (8a MVP + 9a + 9b + 8b-Lite) but had not yet been added to the
canonical table.  Total: 227 → 232.  This brings the doc back in sync
with `ctest` output and unblocks future `scripts/check-test-counts.sh`
runs.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-22 08:40:37 +02:00