feat(phase4): HyperIdeal Newton solver, SparseQR fallback, examples, docs

Phase 4 complete — 87 CGAL tests pass, 2 skipped.

Newton solver (phase4a):
- hyper_ideal_hessian.hpp: symmetric FD Hessian (O(ε²), PSD by convexity)
- newton_hyper_ideal(): Newton + backtracking for the HyperIdeal functional
- detail::solve_with_fallback(): optional bool* fallback_used parameter
- solve_linear_system(): public API exposing LDLT→SparseQR fallback

SparseQR fallback tests (SparseQRFallback.*):
- FullRankSystem_CorrectSolution: LDLT path, fallback_used=false
- SingularMatrix_FallbackActivated: zero-pivot → QR activated, fallback_used=true
- Euclidean_ClosedMeshNoPinConverges: gauge-mode null space handled via QR

HyperIdeal Newton tests (NewtonSolver.HyperIdeal_*):
- ConvergesTriangleAllVariable, ResultFieldsConsistent,
  ConvergesTetrahedron, SparseQRFallbackNoCrash
- Natural-target base point (b=1.0, a=0.5) — x=0 is degenerate in log-space

Pipeline tests (test_pipeline.cpp):
- End-to-end: all three geometries, mesh I/O round-trip, solve+export

Example programs (code/examples/):
- example_euclidean.cpp:   headless Euclidean pipeline
- example_hyper_ideal.cpp: headless HyperIdeal pipeline
- example_viewer.cpp:      interactive libigl viewer with jet colour map

README:
- Mathematical scope table: C++ vs Java original (18 rows)
- "For mathematicians" section: mental model, step-by-step new-functional
  guide, half-edge traversal snippets, recommended reading

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Tarik Moussa
2026-05-13 00:11:25 +02:00
parent e70689d29f
commit 3f124eb071
12 changed files with 1798 additions and 165 deletions

View File

@@ -0,0 +1,147 @@
// example_hyper_ideal.cpp
//
// conformallab++ — Hyper-ideal discrete conformal map (headless example)
//
// Demonstrates the full library pipeline for the HYPER-IDEAL discrete conformal
// functional (Springborn 2020). The hyper-ideal functional operates in
// hyperbolic geometry: vertices have "horoball radii" (DOF b_i) and edges have
// "intersection lengths" (DOF a_e). The energy is strictly convex, so Newton
// converges globally from any valid starting point.
//
// Pipeline:
// 1. Load (or synthesise) a triangle mesh
// 2. Set up HyperIdeal maps + assign all vertex and edge DOFs
// 3. Choose equilibrium base point (b=1.0, a=0.5) and set natural targets
// 4. Perturb and solve with Newton
// 5. Print DOF values at equilibrium
// 6. Save result mesh
//
// Build (requires -DWITH_CGAL=ON):
// cmake -S code -B build -DWITH_CGAL=ON
// cmake --build build --target example_hyper_ideal
// ./build/examples/example_hyper_ideal [input.off] [output.off]
#include "conformal_mesh.hpp"
#include "mesh_builder.hpp"
#include "mesh_io.hpp"
#include "hyper_ideal_functional.hpp"
#include "newton_solver.hpp"
#include <iostream>
#include <string>
#include <vector>
#include <cmath>
using namespace conformallab;
int main(int argc, char* argv[])
{
// ── Step 1: obtain mesh ───────────────────────────────────────────────
ConformalMesh mesh;
std::string input_path = (argc > 1) ? argv[1] : "";
std::string output_path = (argc > 2) ? argv[2] : "/tmp/conformallab_hyper_ideal_out.off";
if (input_path.empty()) {
std::cout << "[example_hyper_ideal] No input file — using make_triangle().\n";
mesh = make_triangle();
} else {
std::cout << "[example_hyper_ideal] Loading mesh from: " << input_path << "\n";
try { mesh = load_mesh(input_path); }
catch (const std::exception& e) {
std::cerr << "Error loading mesh: " << e.what() << "\n";
return 1;
}
}
std::cout << "[example_hyper_ideal] Mesh: "
<< mesh.number_of_vertices() << " vertices, "
<< mesh.number_of_faces() << " faces.\n";
// ── Step 2: set up functional maps ────────────────────────────────────
auto maps = setup_hyper_ideal_maps(mesh);
int n = assign_all_dof_indices(mesh, maps);
std::cout << "[example_hyper_ideal] DOFs: " << n
<< " (" << mesh.number_of_vertices() << " vertex + "
<< mesh.number_of_edges() << " edge).\n";
// ── Step 3: choose equilibrium base point and set natural targets ─────
//
// x = 0 is degenerate for the HyperIdeal functional (log-space).
// We pick a valid base point (b_i = b_base, a_e = a_base), evaluate
// the gradient there, and absorb it into the target angles so that
// G(xbase) = 0. This makes xbase the equilibrium x*.
//
// In a real application you would set theta_v / theta_e to the desired
// hyperbolic angle targets (e.g. from a reference mesh).
const double b_base = 1.0; // horoball radii at equilibrium
const double a_base = 0.5; // edge-length DOFs at equilibrium
const auto sz = static_cast<std::size_t>(n);
std::vector<double> xbase(sz, 0.0);
for (auto v : mesh.vertices()) {
int iv = maps.v_idx[v];
if (iv >= 0) xbase[static_cast<std::size_t>(iv)] = b_base;
}
for (auto e : mesh.edges()) {
int ie = maps.e_idx[e];
if (ie >= 0) xbase[static_cast<std::size_t>(ie)] = a_base;
}
// G = Σβ theta_target; absorb G(xbase) into targets so G(xbase) = 0
auto G0 = evaluate_hyper_ideal(mesh, xbase, maps, false).gradient;
for (auto v : mesh.vertices()) {
int iv = maps.v_idx[v];
if (iv >= 0) maps.theta_v[v] += G0[static_cast<std::size_t>(iv)];
}
for (auto e : mesh.edges()) {
int ie = maps.e_idx[e];
if (ie >= 0) maps.theta_e[e] += G0[static_cast<std::size_t>(ie)];
}
// ── Step 4: perturb and solve ─────────────────────────────────────────
const double perturb = 0.25;
std::vector<double> x0 = xbase;
for (auto& v : x0) v += perturb;
double g_start = 0.0;
for (double v : G0) g_start = std::max(g_start, std::abs(v));
std::cout << "[example_hyper_ideal] Starting Newton from perturbation +" << perturb
<< " (G at xbase = " << g_start << ").\n";
auto result = newton_hyper_ideal(mesh, x0, maps, /*tol=*/1e-9, /*max_iter=*/200);
// ── Step 5: report ────────────────────────────────────────────────────
if (result.converged) {
std::cout << "[example_hyper_ideal] Converged in " << result.iterations
<< " iterations. ||G||_inf = " << result.grad_inf_norm << "\n";
} else {
std::cout << "[example_hyper_ideal] Did NOT converge after " << result.iterations
<< " iterations. ||G||_inf = " << result.grad_inf_norm << "\n";
}
std::cout << "[example_hyper_ideal] DOF values at equilibrium:\n";
for (auto v : mesh.vertices()) {
int iv = maps.v_idx[v];
if (iv < 0) continue;
std::cout << " v" << v
<< " b = " << result.x[static_cast<std::size_t>(iv)]
<< " (expected " << b_base << ")\n";
}
for (auto e : mesh.edges()) {
int ie = maps.e_idx[e];
if (ie < 0) continue;
std::cout << " e" << e
<< " a = " << result.x[static_cast<std::size_t>(ie)]
<< " (expected " << a_base << ")\n";
}
// ── Step 6: write output mesh ─────────────────────────────────────────
try {
save_mesh(output_path, mesh);
std::cout << "[example_hyper_ideal] Mesh saved to: " << output_path << "\n";
} catch (const std::exception& e) {
std::cerr << "Warning: could not write output: " << e.what() << "\n";
}
return result.converged ? 0 : 1;
}