Revision: 75774
http://sourceforge.net/p/brlcad/code/75774
Author: starseeker
Date: 2020-05-12 19:25:57 +0000 (Tue, 12 May 2020)
Log Message:
-----------
Just go with the simple ARB6 method, with a note about why this is hard to do
'right'
Modified Paths:
--------------
brlcad/trunk/src/libged/bot/extrude.cpp
Modified: brlcad/trunk/src/libged/bot/extrude.cpp
===================================================================
--- brlcad/trunk/src/libged/bot/extrude.cpp 2020-05-12 17:59:28 UTC (rev
75773)
+++ brlcad/trunk/src/libged/bot/extrude.cpp 2020-05-12 19:25:57 UTC (rev
75774)
@@ -77,8 +77,16 @@
return true;
}
-
-
+// The "correct" way to do this unfortunately requires an implementation of a
+// surface offsetting algorithm, which is a hard problem:
+//
+// Charlie C.L. Wang, and Yong Chen, "Thickening freeform surfaces for solid
+// fabrication", Rapid Prototyping Journal, vol.19, no.6, pp.395-406, November
+// 2013.
+//
+// Chen, Zhen and Panozzo, Daniele and Dumas, Jeremie, "Half-Space Power
+// Diagrams and Discrete Surface Offsets", IEEE Transactions on Visualization
+// and Computer Graphics, 2019
(https://github.com/geometryprocessing/voroffset)
extern "C" int
_bot_cmd_extrude(void *bs, int argc, const char **argv)
{
@@ -123,7 +131,7 @@
bu_vls_sprintf(&comb_name, "%s_extrusion.r", gb->dp->d_namep);
}
- if (db_lookup(gb->gedp->ged_wdbp->dbip, bu_vls_cstr(&comb_name),
LOOKUP_NOISY) != RT_DIR_NULL) {
+ if (db_lookup(gb->gedp->ged_wdbp->dbip, bu_vls_cstr(&comb_name),
LOOKUP_QUIET) != RT_DIR_NULL) {
bu_vls_printf(gb->gedp->ged_result_str, "Object %s already exists!\n",
bu_vls_cstr(&comb_name));
bu_vls_free(&comb_name);
return GED_ERROR;
@@ -131,32 +139,6 @@
BU_LIST_INIT(&wcomb.l);
- // Average the face normals at each vertex to get an average direction in
- // which to move each vertex for solid generation.
- vect_t *f2n = (vect_t *)bu_calloc(bot->num_faces, sizeof(vect_t), "face
normals");
- vect_t *v2n = (vect_t *)bu_calloc(bot->num_vertices, sizeof(vect_t), "vert
normals");
- int *v2fc = (int *)bu_calloc(bot->num_vertices, sizeof(int), "cnts");
- for (size_t i = 0; i < bot->num_faces; i++) {
- vect_t n;
- bot_face_normal(&n, bot, i);
- VMOVE(f2n[i],n);
- }
- // Add up all the face normal vectors at each vertex
- for (size_t i = 0; i < bot->num_faces; i++) {
- for (int j = 0; j < 3; j++) {
- VADD2(v2n[bot->faces[i*3+j]], v2n[bot->faces[i*3+j]], f2n[i]);
- v2fc[bot->faces[i*3+j]] = v2fc[bot->faces[i*3+j]] + 1;
- }
- }
- // Average based on how many faces contributed
- for (size_t i = 0; i < bot->num_vertices; i++) {
- VSCALE(v2n[i], v2n[i], (1.0/(double)v2fc[i]));
- }
- // Unitize
- for (size_t i = 0; i < bot->num_vertices; i++) {
- VUNITIZE(v2n[i]);
- }
-
// For each face, define an arb6 using shifted vertices. For each face
// vertex two new points will be constructed - an inner and outer - based
// on the original point, the local face thickness, and the avg face dir
@@ -164,90 +146,42 @@
// in which direction.
struct bu_vls prim_name = BU_VLS_INIT_ZERO;
for (size_t i = 0; i < bot->num_faces; i++) {
- point_t pnts[6];
- point_t pf[3];
- vect_t pv1[3], pv2[3];
+ point_t pnts[6];
+ point_t pf[3];
+ vect_t pv1[3], pv2[3];
+ vect_t n;
+ bot_face_normal(&n, bot, i);
+
for (int j = 0; j < 3; j++) {
- vect_t fnorm, n;
- VMOVE(fnorm, v2n[bot->faces[i*3+j]]);
- bot_face_normal(&n, bot, i);
- double vdot = VDOT(fnorm, n);
- if (vdot < 0) {
- VSCALE(fnorm, fnorm, -1);
- }
- VMOVE(pf[j], &bot->vertices[bot->faces[i*3+j]*3]);
- // TODO - try trig values here for better scaling, up to some sane
max...
- VSCALE(pv1[j], fnorm, bot->thickness[i] * ((bot->mode == RT_BOT_CW)
? -1 : 0));
- VSCALE(pv2[j], fnorm, -1*bot->thickness[i] * ((bot->mode ==
RT_BOT_CW) ? -1 : 0));
- }
- for (int j = 0; j < 3; j++) {
- point_t npnt1;
- point_t npnt2;
- VADD2(npnt1, pf[j], pv1[j]);
- VADD2(npnt2, pf[j], pv2[j]);
- VMOVE(pnts[j], npnt1);
- VMOVE(pnts[j+3], npnt2);
- }
+ VMOVE(pf[j], &bot->vertices[bot->faces[i*3+j]*3]);
+ bu_log("Face %zd point %d: %f %f %f\n", i, j, V3ARGS(pf[j]));
+ VSCALE(pv1[j], n, bot->thickness[i]);
+ VSCALE(pv2[j], n, -1*bot->thickness[i]);
+ }
+ for (int j = 0; j < 3; j++) {
+ point_t npnt1;
+ point_t npnt2;
+ VADD2(npnt1, pf[j], pv1[j]);
+ VADD2(npnt2, pf[j], pv2[j]);
+ VMOVE(pnts[j], npnt1);
+ VMOVE(pnts[j+3], npnt2);
+ }
- // We are in effect creating an approximation of a twisted arb6 with
this
- // shape - that means we need 8 bot faces on the 6 points.
- // To make it easier to use an arb6 shape as a guide, assign the points
in arb6 ordering:
- fastf_t pnts_array[3*6+3*3];
- /* 1 */ pnts_array[0] = pnts[4][X]; pnts_array[1] = pnts[4][Y];
pnts_array[2] = pnts[4][Z];
- /* 2 */ pnts_array[3] = pnts[3][X]; pnts_array[4] = pnts[3][Y];
pnts_array[5] = pnts[3][Z];
- /* 3 */ pnts_array[6] = pnts[0][X]; pnts_array[7] = pnts[0][Y];
pnts_array[8] = pnts[0][Z];
- /* 4 */ pnts_array[9] = pnts[1][X]; pnts_array[10] = pnts[1][Y];
pnts_array[11] = pnts[1][Z];
- /* 5 */ pnts_array[12] = pnts[5][X]; pnts_array[13] = pnts[5][Y];
pnts_array[14] = pnts[5][Z];
- /* 6 */ pnts_array[15] = pnts[2][X]; pnts_array[16] = pnts[2][Y];
pnts_array[17] = pnts[2][Z];
+ // For arb6 creation, we need a specific point order
+ fastf_t pnts_array[3*6];
+ /* 1 */ pnts_array[0] = pnts[4][X]; pnts_array[1] = pnts[4][Y];
pnts_array[2] = pnts[4][Z];
+ /* 2 */ pnts_array[3] = pnts[3][X]; pnts_array[4] = pnts[3][Y];
pnts_array[5] = pnts[3][Z];
+ /* 3 */ pnts_array[6] = pnts[0][X]; pnts_array[7] = pnts[0][Y];
pnts_array[8] = pnts[0][Z];
+ /* 4 */ pnts_array[9] = pnts[1][X]; pnts_array[10] = pnts[1][Y];
pnts_array[11] = pnts[1][Z];
+ /* 5 */ pnts_array[12] = pnts[5][X]; pnts_array[13] = pnts[5][Y];
pnts_array[14] = pnts[5][Z];
+ /* 6 */ pnts_array[15] = pnts[2][X]; pnts_array[16] = pnts[2][Y];
pnts_array[17] = pnts[2][Z];
- // Three new points are needed at the centers of what would otherwise be
- // twisted faces, to define triangles that won't overlap.
- // ARB6 face 1,2,3,4
- point_t pnts_avg[3] = {VINIT_ZERO, VINIT_ZERO, VINIT_ZERO};
- VADD2(pnts_avg[0], pnts[0], pnts_avg[0]);
- VADD2(pnts_avg[0], pnts[1], pnts_avg[0]);
- VADD2(pnts_avg[0], pnts[3], pnts_avg[0]);
- VADD2(pnts_avg[0], pnts[4], pnts_avg[0]);
- VSCALE(pnts_avg[0], pnts_avg[0], 0.25);
- /* 7 */ pnts_array[18] = pnts_avg[0][X]; pnts_array[19] =
pnts_avg[0][Y]; pnts_array[20] = pnts_avg[0][Z];
+ bu_vls_sprintf(&prim_name, "%s.arb6.%zd", gb->dp->d_namep, i);
- // ARB6 face 2,3,5,6
- VADD2(pnts_avg[1], pnts[0], pnts_avg[1]);
- VADD2(pnts_avg[1], pnts[2], pnts_avg[1]);
- VADD2(pnts_avg[1], pnts[3], pnts_avg[1]);
- VADD2(pnts_avg[1], pnts[5], pnts_avg[1]);
- VSCALE(pnts_avg[1], pnts_avg[1], 0.25);
- /* 8 */ pnts_array[21] = pnts_avg[1][X]; pnts_array[22] =
pnts_avg[1][Y]; pnts_array[23] = pnts_avg[1][Z];
+ // For arb6 creation we need to move a couple points the array.
- // ARB6 face 1,5,6,4
- VADD2(pnts_avg[2], pnts[1], pnts_avg[2]);
- VADD2(pnts_avg[2], pnts[2], pnts_avg[2]);
- VADD2(pnts_avg[2], pnts[4], pnts_avg[2]);
- VADD2(pnts_avg[2], pnts[5], pnts_avg[2]);
- VSCALE(pnts_avg[2], pnts_avg[2], 0.25);
- /* 9 */ pnts_array[24] = pnts_avg[2][X]; pnts_array[25] =
pnts_avg[2][Y]; pnts_array[26] = pnts_avg[2][Z];
-
- int faces_array[3*14];
- /* 1 */ faces_array[0] = 1; faces_array[1] = 0; faces_array[2] = 4;
- /* 2 */ faces_array[3] = 3; faces_array[4] = 2; faces_array[5] = 5;
-
- /* 3 */ faces_array[6] = 1; faces_array[7] = 2; faces_array[8] = 6;
- /* 4 */ faces_array[9] = 2; faces_array[10] = 3; faces_array[11] = 6;
- /* 5 */ faces_array[12] = 3; faces_array[13] = 0; faces_array[14] = 6;
- /* 6 */ faces_array[15] = 0; faces_array[16] = 1; faces_array[17] = 6;
- /* 7 */ faces_array[18] = 4; faces_array[19] = 5; faces_array[20] = 7;
- /* 8 */ faces_array[21] = 5; faces_array[22] = 2; faces_array[23] = 7;
- /* 9 */ faces_array[24] = 2; faces_array[25] = 1; faces_array[26] = 7;
- /* 10 */ faces_array[27] = 1; faces_array[28] = 4; faces_array[29] = 7;
- /* 11 */ faces_array[30] = 4; faces_array[31] = 0; faces_array[32] = 8;
- /* 12 */ faces_array[33] = 5; faces_array[34] = 4; faces_array[35] = 8;
- /* 13 */ faces_array[36] = 3; faces_array[37] = 5; faces_array[38] = 8;
- /* 14 */ faces_array[39] = 0; faces_array[40] = 3; faces_array[41] = 8;
-
- bu_vls_sprintf(&prim_name, "%s.tarb6.%zd", gb->dp->d_namep, i);
-
- mk_bot(gb->gedp->ged_wdbp, bu_vls_cstr(&prim_name), RT_BOT_SOLID,
RT_BOT_CCW, 0, 9, 14, (fastf_t *)pnts_array, (int *)faces_array, NULL, NULL);
- (void)mk_addmember(bu_vls_cstr(&prim_name), &(wcomb.l), NULL,
DB_OP_UNION);
+ mk_arb6(gb->gedp->ged_wdbp, bu_vls_cstr(&prim_name), pnts_array);
+ (void)mk_addmember(bu_vls_cstr(&prim_name), &(wcomb.l), NULL,
DB_OP_UNION);
}
// Write the comb
@@ -255,9 +189,6 @@
bu_vls_free(&comb_name);
bu_vls_free(&prim_name);
- bu_free(f2n, "face normals");
- bu_free(v2n, "vert normals");
- bu_free(v2fc, "cnts");
return GED_OK;
}
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
_______________________________________________
BRL-CAD Source Commits mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-commits
