void creature_t::step_toward(int px, int py, creature_t* exclude)
{
TCODMap* map = new TCODMap(current_dungeon->width, current_dungeon->height);
for (int y = 0; y < current_dungeon->height; y++)
{
for (int x = 0; x < current_dungeon->width; x++)
{
tile_t* tile = get_tile_at(current_dungeon, x, y);
bool walkable = tile_is_walkable_by(*tile, this);
if (flag & CF_INTELLIGENT)
{
// Intelligent monsters know how to open doors.
if (tile->id == TILE_DOOR_CLOSED)
{
walkable = true;
}
}
map->setProperties(x, y, tile->properties & TILE_PROP_TRANSPARENT, walkable);
}
}
for (size_t i = 0; i < current_dungeon->creatures.size(); i++)
{
creature_t* c = current_dungeon->creatures.at(i);
if (c == this || c == exclude)
continue;
if (sees(c))
{
// If this creature can see another creature, then consider that
// square as unwalkable.
map->setProperties(c->pos.x, c->pos.y, 1, 0);
}
}
TCODPath* path = new TCODPath(map, 1.0f);
if (path->compute(pos.x, pos.y, px, py))
{
if (path->size() > 0)
{
int move_x, move_y;
path->get(0, &move_x, &move_y);
try_move(move_x, move_y);
}
else
{
// No path found, walk around confused.
random_walk();
}
}
else
{
// No path found, walk around confused.
random_walk();
}
delete map;
delete path;
}
// Random walking of matrix of position created in populate
// Step 1: Move particles via 6D random walk
// Step 2: Destroy and create particles as need based on Anderson
// Step 3: check energy, end if needed.
void diffuse(h3plus& state, std::ostream& output, std::ostream& output2){
// For now, I am going to set a definite number of timesteps
// This will be replaced by a while loop in the future.
// double diff = 1.0;
// while (diff > 0.01) {
for (size_t t = 0; t < 200; t++){
double v_last = state.v_ref;
random_walk(state);
branch(state);
state.t += state.dt;
// diff = sqrt((v_last - state.v_ref) * (v_last - state.v_ref));
// Debug information for the current timestep
std::cout << std::fixed
<< state.energy << '\t'
<< state.particles.size() << '\t'
<< state.dt << '\n';
if (t % 1 == 0) {
print_visualization_data(output, state, output2);
}
state.dt = TIMESTEP * exp(-(t / 200.0)) + 0.001;
}
}
void random_walk_all(
float *turtles,
int num_turtles,
float magnitude,
float half_w,
float half_h,
float degrees
)
{
int i;
for (i = 0; i < num_turtles; ++i)
{
random_walk(turtles + (i * TURTLE_DATA_SIZE), magnitude, half_w, half_h, degrees);
}
}
void handler_ctrl(void) {
// runs at 1KHz
static uint16 t = 0;
t++;
// if (t < 512) {
// pwmWrite(X_OUT_P,tri(t));
// } else {
// pwmWrite(Y_OUT_P,tri(t));
// }
//square(t);
random_walk();
static uint16 t2 = 0;
if ((t%32 == 0)) {
t = 0;
t2++;;
random_im();
}
draw_point();
// pwmWrite(LASER,t/256);
}
// [[register]]
SEXP mcmcbas(SEXP Y, SEXP X, SEXP Rweights, SEXP Rprobinit, SEXP Rmodeldim, SEXP incint, SEXP Ralpha,SEXP method, SEXP modelprior, SEXP Rupdate, SEXP Rbestmodel, SEXP plocal,
SEXP BURNIN_Iterations, SEXP MCMC_Iterations, SEXP LAMBDA, SEXP DELTA,
SEXP Rparents)
{
SEXP RXwork = PROTECT(duplicate(X)), RYwork = PROTECT(duplicate(Y));
int nProtected = 2, nUnique=0, newmodel=0;
int nModels=LENGTH(Rmodeldim);
// Rprintf("Allocating Space for %d Models\n", nModels) ;
SEXP ANS = PROTECT(allocVector(VECSXP, 15)); ++nProtected;
SEXP ANS_names = PROTECT(allocVector(STRSXP, 15)); ++nProtected;
SEXP Rprobs = PROTECT(duplicate(Rprobinit)); ++nProtected;
SEXP MCMCprobs= PROTECT(duplicate(Rprobinit)); ++nProtected;
SEXP R2 = PROTECT(allocVector(REALSXP, nModels)); ++nProtected;
SEXP shrinkage = PROTECT(allocVector(REALSXP, nModels)); ++nProtected;
SEXP modelspace = PROTECT(allocVector(VECSXP, nModels)); ++nProtected;
SEXP modeldim = PROTECT(duplicate(Rmodeldim)); ++nProtected;
SEXP counts = PROTECT(duplicate(Rmodeldim)); ++nProtected;
SEXP beta = PROTECT(allocVector(VECSXP, nModels)); ++nProtected;
SEXP se = PROTECT(allocVector(VECSXP, nModels)); ++nProtected;
SEXP mse = PROTECT(allocVector(REALSXP, nModels)); ++nProtected;
SEXP modelprobs = PROTECT(allocVector(REALSXP, nModels)); ++nProtected;
SEXP priorprobs = PROTECT(allocVector(REALSXP, nModels)); ++nProtected;
SEXP logmarg = PROTECT(allocVector(REALSXP, nModels)); ++nProtected;
SEXP sampleprobs = PROTECT(allocVector(REALSXP, nModels)); ++nProtected;
SEXP NumUnique = PROTECT(allocVector(INTSXP, 1)); ++nProtected;
SEXP Rse_m = NULL, Rcoef_m = NULL, Rmodel_m;
double *Xwork, *Ywork, *wts, *coefficients,*probs, shrinkage_m, *MCMC_probs,
SSY, yty, mse_m, *se_m, MH=0.0, prior_m=1.0, *real_model,
R2_m, RSquareFull, alpha, prone, denom, logmargy, postold, postnew;
int nobs, p, k, i, j, m, n, l, pmodel, pmodel_old, *xdims, *model_m, *bestmodel, *varin, *varout;
int mcurrent, update, n_sure;
double mod, rem, problocal, *pigamma, eps, *hyper_parameters;
double *XtX, *XtY, *XtXwork, *XtYwork, *SSgam, *Cov, *priorCov, *marg_probs;
double lambda, delta, one=1.0;
int inc=1;
int *model, *modelold, bit, *modelwork, old_loc, new_loc;
// char uplo[] = "U", trans[]="T";
struct Var *vars; /* Info about the model variables. */
NODEPTR tree, branch;
/* get dimsensions of all variables */
nobs = LENGTH(Y);
xdims = INTEGER(getAttrib(X,R_DimSymbol));
p = xdims[1];
k = LENGTH(modelprobs);
update = INTEGER(Rupdate)[0];
lambda=REAL(LAMBDA)[0];
delta = REAL(DELTA)[0];
// Rprintf("delta %f lambda %f", delta, lambda);
eps = DBL_EPSILON;
problocal = REAL(plocal)[0];
// Rprintf("Update %i and prob.switch %f\n", update, problocal);
/* Extract prior on models */
hyper_parameters = REAL(getListElement(modelprior,"hyper.parameters"));
/* Rprintf("n %d p %d \n", nobs, p); */
Ywork = REAL(RYwork);
Xwork = REAL(RXwork);
wts = REAL(Rweights);
/* Allocate other variables. */
PrecomputeData(Xwork, Ywork, wts, &XtXwork, &XtYwork, &XtX, &XtY, &yty, &SSY, p, nobs);
alpha = REAL(Ralpha)[0];
vars = (struct Var *) R_alloc(p, sizeof(struct Var));
probs = REAL(Rprobs);
n = sortvars(vars, probs, p);
for (i =n; i <p; i++) REAL(MCMCprobs)[vars[i].index] = probs[vars[i].index];
for (i =0; i <n; i++) REAL(MCMCprobs)[vars[i].index] = 0.0;
MCMC_probs = REAL(MCMCprobs);
pigamma = vecalloc(p);
real_model = vecalloc(n);
marg_probs = vecalloc(n);
modelold = ivecalloc(p);
model = ivecalloc(p);
modelwork= ivecalloc(p);
varin= ivecalloc(p);
varout= ivecalloc(p);
/* create gamma gamma' matrix */
SSgam = (double *) R_alloc(n * n, sizeof(double));
Cov = (double *) R_alloc(n * n, sizeof(double));
priorCov = (double *) R_alloc(n * n, sizeof(double));
for (j=0; j < n; j++) {
for (i = 0; i < n; i++) {
SSgam[j*n + i] = 0.0;
Cov[j*n + i] = 0.0;
priorCov[j*n + i] = 0.0;
if (j == i) priorCov[j*n + i] = lambda;
}
marg_probs[i] = 0.0;
}
RSquareFull = CalculateRSquareFull(XtY, XtX, XtXwork, XtYwork, Rcoef_m, Rse_m, p, nobs, yty, SSY);
/* fill in the sure things */
for (i = n, n_sure = 0; i < p; i++) {
model[vars[i].index] = (int) vars[i].prob;
if (model[vars[i].index] == 1) ++n_sure;
}
GetRNGstate();
tree = make_node(-1.0);
/* Rprintf("For m=0, Initialize Tree with initial Model\n"); */
m = 0;
bestmodel = INTEGER(Rbestmodel);
INTEGER(modeldim)[m] = n_sure;
/* Rprintf("Create Tree\n"); */
branch = tree;
for (i = 0; i< n; i++) {
bit = bestmodel[vars[i].index];
if (bit == 1) {
if (i < n-1 && branch->one == NULL)
branch->one = make_node(-1.0);
if (i == n-1 && branch->one == NULL)
branch->one = make_node(0.0);
branch = branch->one;
}
else {
if (i < n-1 && branch->zero == NULL)
branch->zero = make_node(-1.0);
if (i == n-1 && branch->zero == NULL)
branch->zero = make_node(0.0);
branch = branch->zero;
}
model[vars[i].index] = bit;
INTEGER(modeldim)[m] += bit;
branch->where = 0;
}
/* Rprintf("Now get model specific calculations \n"); */
pmodel = INTEGER(modeldim)[m];
PROTECT(Rmodel_m = allocVector(INTSXP,pmodel));
model_m = INTEGER(Rmodel_m);
for (j = 0, l=0; j < p; j++) {
if (model[j] == 1) {
model_m[l] = j;
l +=1;}
}
SET_ELEMENT(modelspace, m, Rmodel_m);
Rcoef_m = NEW_NUMERIC(pmodel); PROTECT(Rcoef_m);
Rse_m = NEW_NUMERIC(pmodel); PROTECT(Rse_m);
coefficients = REAL(Rcoef_m);
se_m = REAL(Rse_m);
for (j=0, l=0; j < pmodel; j++) {
XtYwork[j] = XtY[model_m[j]];
for ( i = 0; i < pmodel; i++) {
XtXwork[j*pmodel + i] = XtX[model_m[j]*p + model_m[i]];
}
}
R2_m = 0.0;
mse_m = yty;
memcpy(coefficients, XtYwork, sizeof(double)*pmodel);
cholreg(XtYwork, XtXwork, coefficients, se_m, &mse_m, pmodel, nobs);
if (pmodel > 1) R2_m = 1.0 - (mse_m * (double) ( nobs - pmodel))/SSY;
SET_ELEMENT(beta, m, Rcoef_m);
SET_ELEMENT(se, m, Rse_m);
REAL(R2)[m] = R2_m;
REAL(mse)[m] = mse_m;
gexpectations(p, pmodel, nobs, R2_m, alpha, INTEGER(method)[0], RSquareFull, SSY, &logmargy, &shrinkage_m);
REAL(sampleprobs)[m] = 1.0;
REAL(logmarg)[m] = logmargy;
REAL(shrinkage)[m] = shrinkage_m;
prior_m = compute_prior_probs(model,pmodel,p, modelprior);
REAL(priorprobs)[m] = prior_m;
UNPROTECT(3);
old_loc = 0;
pmodel_old = pmodel;
nUnique=1;
INTEGER(counts)[0] = 0;
postold = REAL(logmarg)[m] + log(REAL(priorprobs)[m]);
memcpy(modelold, model, sizeof(int)*p);
/* Rprintf("model %d max logmarg %lf\n", m, REAL(logmarg)[m]); */
/* Rprintf("Now Sample the Rest of the Models \n"); */
m = 0;
while (nUnique < k && m < INTEGER(BURNIN_Iterations)[0]) {
memcpy(model, modelold, sizeof(int)*p);
pmodel = n_sure;
MH = 1.0;
if (pmodel_old == n_sure || pmodel_old == n_sure + n){
MH = random_walk(model, vars, n);
MH = 1.0 - problocal;
}
else {
if (unif_rand() < problocal) {
// random
MH = random_switch(model, vars, n, pmodel_old, varin, varout );
}
else {
// Randomw walk proposal flip bit//
MH = random_walk(model, vars, n);
}
}
branch = tree;
newmodel= 0;
for (i = 0; i< n; i++) {
bit = model[vars[i].index];
if (bit == 1) {
if (branch->one != NULL) branch = branch->one;
else newmodel = 1;
}
else {
if (branch->zero != NULL) branch = branch->zero;
else newmodel = 1.0;
}
pmodel += bit;
}
if (pmodel == n_sure || pmodel == n + n_sure) MH = 1.0/(1.0 - problocal);
if (newmodel == 1) {
new_loc = nUnique;
PROTECT(Rmodel_m = allocVector(INTSXP,pmodel));
model_m = INTEGER(Rmodel_m);
for (j = 0, l=0; j < p; j++) {
if (model[j] == 1) {
model_m[l] = j;
l +=1;}
}
Rcoef_m = NEW_NUMERIC(pmodel); PROTECT(Rcoef_m);
Rse_m = NEW_NUMERIC(pmodel); PROTECT(Rse_m);
coefficients = REAL(Rcoef_m);
se_m = REAL(Rse_m);
for (j=0, l=0; j < pmodel; j++) {
XtYwork[j] = XtY[model_m[j]];
for ( i = 0; i < pmodel; i++) {
XtXwork[j*pmodel + i] = XtX[model_m[j]*p + model_m[i]];
}
}
R2_m = 0.0;
mse_m = yty;
memcpy(coefficients, XtYwork, sizeof(double)*pmodel);
cholreg(XtYwork, XtXwork, coefficients, se_m, &mse_m, pmodel, nobs);
if (pmodel > 1) R2_m = 1.0 - (mse_m * (double) ( nobs - pmodel))/SSY;
prior_m = compute_prior_probs(model,pmodel,p, modelprior);
gexpectations(p, pmodel, nobs, R2_m, alpha, INTEGER(method)[0], RSquareFull, SSY, &logmargy, &shrinkage_m);
postnew = logmargy + log(prior_m);
}
else {
new_loc = branch->where;
postnew = REAL(logmarg)[new_loc] + log(REAL(priorprobs)[new_loc]);
}
MH *= exp(postnew - postold);
// Rprintf("MH new %lf old %lf\n", postnew, postold);
if (unif_rand() < MH) {
if (newmodel == 1) {
new_loc = nUnique;
insert_model_tree(tree, vars, n, model, nUnique);
INTEGER(modeldim)[nUnique] = pmodel;
SET_ELEMENT(modelspace, nUnique, Rmodel_m);
SET_ELEMENT(beta, nUnique, Rcoef_m);
SET_ELEMENT(se, nUnique, Rse_m);
REAL(R2)[nUnique] = R2_m;
REAL(mse)[nUnique] = mse_m;
REAL(sampleprobs)[nUnique] = 1.0;
REAL(logmarg)[nUnique] = logmargy;
REAL(shrinkage)[nUnique] = shrinkage_m;
REAL(priorprobs)[nUnique] = prior_m;
UNPROTECT(3);
++nUnique;
}
old_loc = new_loc;
postold = postnew;
pmodel_old = pmodel;
memcpy(modelold, model, sizeof(int)*p);
}
else {
if (newmodel == 1) UNPROTECT(3);
}
INTEGER(counts)[old_loc] += 1;
for (i = 0; i < n; i++) {
/* store in opposite order so nth variable is first */
real_model[n-1-i] = (double) modelold[vars[i].index];
REAL(MCMCprobs)[vars[i].index] += (double) modelold[vars[i].index];
}
// Update SSgam = gamma gamma^T + SSgam
F77_NAME(dsyr)("U", &n, &one, &real_model[0], &inc, &SSgam[0], &n);
m++;
}
for (i = 0; i < n; i++) {
REAL(MCMCprobs)[vars[i].index] /= (double) m;
}
// Rprintf("\n%d \n", nUnique);
// Compute marginal probabilities
mcurrent = nUnique;
compute_modelprobs(modelprobs, logmarg, priorprobs,mcurrent);
compute_margprobs(modelspace, modeldim, modelprobs, probs, mcurrent, p);
// Now sample W/O Replacement
// Rprintf("NumUnique Models Accepted %d \n", nUnique);
INTEGER(NumUnique)[0] = nUnique;
if (nUnique < k) {
update_probs(probs, vars, mcurrent, k, p);
update_tree(modelspace, tree, modeldim, vars, k,p,n,mcurrent, modelwork);
for (m = nUnique; m < k; m++) {
for (i = n; i < p; i++) {
INTEGER(modeldim)[m] += model[vars[i].index];
}
branch = tree;
for (i = 0; i< n; i++) {
pigamma[i] = 1.0;
bit = withprob(branch->prob);
/* branch->done += 1; */
if (bit == 1) {
for (j=0; j<=i; j++) pigamma[j] *= branch->prob;
if (i < n-1 && branch->one == NULL)
branch->one = make_node(vars[i+1].prob);
if (i == n-1 && branch->one == NULL)
branch->one = make_node(0.0);
branch = branch->one;
}
else {
for (j=0; j<=i; j++) pigamma[j] *= (1.0 - branch->prob);
if (i < n-1 && branch->zero == NULL)
branch->zero = make_node(vars[i+1].prob);
if (i == n-1 && branch->zero == NULL)
branch->zero = make_node(0.0);
branch = branch->zero;
}
model[vars[i].index] = bit;
INTEGER(modeldim)[m] += bit;
}
REAL(sampleprobs)[m] = pigamma[0];
pmodel = INTEGER(modeldim)[m];
/* Now subtract off the visited probability mass. */
branch=tree;
for (i = 0; i < n; i++) {
bit = model[vars[i].index];
prone = branch->prob;
if (bit == 1) prone -= pigamma[i];
denom = 1.0 - pigamma[i];
if (denom <= 0.0) {
if (denom < 0.0) {
warning("neg denominator %le %le %le !!!\n", pigamma, denom, prone);
if (branch->prob < 0.0 && branch->prob < 1.0)
warning("non extreme %le\n", branch->prob);}
denom = 0.0;}
else {
if (prone <= 0) prone = 0.0;
if (prone > denom) {
if (prone <= eps) prone = 0.0;
else prone = 1.0;
/* Rprintf("prone > 1 %le %le %le %le !!!\n", pigamma, denom, prone, eps);*/
}
else prone = prone/denom;
}
if (prone > 1.0 || prone < 0.0)
Rprintf("%d %d Probability > 1!!! %le %le %le %le \n",
m, i, prone, branch->prob, denom, pigamma);
/* if (bit == 1) pigamma /= (branch->prob);
else pigamma /= (1.0 - branch->prob);
if (pigamma > 1.0) pigamma = 1.0; */
branch->prob = prone;
if (bit == 1) branch = branch->one;
else branch = branch->zero;
/* Rprintf("%d %d \n", branch->done, n - i); */
/* if (log((double) branch->done) < (n - i)*log(2.0)) {
if (bit == 1) branch = branch->one;
else branch = branch->zero;
}
else {
branch->one = NULL;
branch->zero = NULL;
break; } */
}
/* Now get model specific calculations */
PROTECT(Rmodel_m = allocVector(INTSXP, pmodel));
model_m = INTEGER(Rmodel_m);
for (j = 0, l=0; j < p; j++) {
if (model[j] == 1) {
model_m[l] = j;
l +=1;}
}
SET_ELEMENT(modelspace, m, Rmodel_m);
for (j=0, l=0; j < pmodel; j++) {
XtYwork[j] = XtY[model_m[j]];
for ( i = 0; i < pmodel; i++) {
XtXwork[j*pmodel + i] = XtX[model_m[j]*p + model_m[i]];
}
}
PROTECT(Rcoef_m = allocVector(REALSXP,pmodel));
PROTECT(Rse_m = allocVector(REALSXP,pmodel));
coefficients = REAL(Rcoef_m);
se_m = REAL(Rse_m);
mse_m = yty;
memcpy(coefficients, XtYwork, sizeof(double)*pmodel);
cholreg(XtYwork, XtXwork, coefficients, se_m, &mse_m, pmodel, nobs);
/* olsreg(Ywork, Xwork, coefficients, se_m, &mse_m, &pmodel, &nobs, pivot,qraux,work,residuals,effects,v,betaols); */
if (pmodel > 1) R2_m = 1.0 - (mse_m * (double) ( nobs - pmodel))/SSY;
SET_ELEMENT(beta, m, Rcoef_m);
SET_ELEMENT(se, m, Rse_m);
REAL(R2)[m] = R2_m;
REAL(mse)[m] = mse_m;
gexpectations(p, pmodel, nobs, R2_m, alpha, INTEGER(method)[0], RSquareFull, SSY, &logmargy, &shrinkage_m);
REAL(logmarg)[m] = logmargy;
REAL(shrinkage)[m] = shrinkage_m;
REAL(priorprobs)[m] = compute_prior_probs(model,pmodel,p, modelprior);
if (m > 1) {
rem = modf((double) m/(double) update, &mod);
if (rem == 0.0) {
mcurrent = m;
compute_modelprobs(modelprobs, logmarg, priorprobs,mcurrent);
compute_margprobs(modelspace, modeldim, modelprobs, probs, mcurrent, p);
if (update_probs(probs, vars, mcurrent, k, p) == 1) {
// Rprintf("Updating Model Tree %d \n", m);
update_tree(modelspace, tree, modeldim, vars, k,p,n,mcurrent, modelwork);
}
}}
UNPROTECT(3);
}
}
compute_modelprobs(modelprobs, logmarg, priorprobs,k);
compute_margprobs(modelspace, modeldim, modelprobs, probs, k, p);
SET_VECTOR_ELT(ANS, 0, Rprobs);
SET_STRING_ELT(ANS_names, 0, mkChar("probne0"));
SET_VECTOR_ELT(ANS, 1, modelspace);
SET_STRING_ELT(ANS_names, 1, mkChar("which"));
SET_VECTOR_ELT(ANS, 2, logmarg);
SET_STRING_ELT(ANS_names, 2, mkChar("logmarg"));
SET_VECTOR_ELT(ANS, 3, modelprobs);
SET_STRING_ELT(ANS_names, 3, mkChar("postprobs"));
SET_VECTOR_ELT(ANS, 4, priorprobs);
SET_STRING_ELT(ANS_names, 4, mkChar("priorprobs"));
SET_VECTOR_ELT(ANS, 5,sampleprobs);
SET_STRING_ELT(ANS_names, 5, mkChar("sampleprobs"));
SET_VECTOR_ELT(ANS, 6, mse);
SET_STRING_ELT(ANS_names, 6, mkChar("mse"));
SET_VECTOR_ELT(ANS, 7, beta);
SET_STRING_ELT(ANS_names, 7, mkChar("mle"));
SET_VECTOR_ELT(ANS, 8, se);
SET_STRING_ELT(ANS_names, 8, mkChar("mle.se"));
SET_VECTOR_ELT(ANS, 9, shrinkage);
SET_STRING_ELT(ANS_names, 9, mkChar("shrinkage"));
SET_VECTOR_ELT(ANS, 10, modeldim);
SET_STRING_ELT(ANS_names, 10, mkChar("size"));
SET_VECTOR_ELT(ANS, 11, R2);
SET_STRING_ELT(ANS_names, 11, mkChar("R2"));
SET_VECTOR_ELT(ANS, 12, counts);
SET_STRING_ELT(ANS_names, 12, mkChar("freq"));
SET_VECTOR_ELT(ANS, 13, MCMCprobs);
SET_STRING_ELT(ANS_names, 13, mkChar("probs.MCMC"));
SET_VECTOR_ELT(ANS, 14, NumUnique);
SET_STRING_ELT(ANS_names, 14, mkChar("n.Unique"));
setAttrib(ANS, R_NamesSymbol, ANS_names);
UNPROTECT(nProtected);
// Rprintf("Return\n");
PutRNGstate();
return(ANS);
}
/*-------------------------------------------------------------------*/
void loop() {
random_walk();
broadcast();
color_code();
}
/* Launches one of those fancy effects. */
void launch_effect(int effect)
{
switch (effect) {
case 0:
/* Lights all the layers one by one. */
load_bar(1000);
break;
case 1:
/* A pixel bouncing randomly around. */
boing_boing(150, 500, 0x03, 0x01);
break;
case 2:
/* Randomly fill and empty the cube. */
fill(0x00);
random_filler(100, 1, 500, 1);
random_filler(100, 1, 500, 0);
break;
case 3:
/* Send voxels randomly back and forth the z-axis. */
send_voxels_rand_z(150, 500, 2000);
break;
case 4:
/* Spinning spiral */
spiral(1, 75, 1000);
break;
case 5:
/* A coordinate bounces randomly around the cube. For every position
* the status of that voxel is toggled. */
boing_boing(150, 500, 0x03, 0x02);
break;
case 6:
/* Random raindrops */
rain(40, 1000, 500, 500);
break;
case 7:
/* Snake: a snake randomly bounce around the cube. */
boing_boing(150, 500, 0x03, 0x03);
break;
case 8:
/* Spinning plane */
spinning_plane(1, 50, 1000);
break;
case 9:
/* Set x number of random voxels, delay, unset them. x increases
* from 1 to 32 and back to 1. */
random_2(48);
break;
case 10:
/* Set then unset all 64 voxels in a random order. */
random_filler2(200, 1);
delay_ms(2000);
random_filler2(200, 0);
delay_ms(1000);
break;
case 11:
/* Bounce a plane up and down all the directions. */
fly_plane('z', 1, 1000);
delay_ms(2000);
fly_plane('y', 1, 1000);
delay_ms(2000);
fly_plane('x', 1, 1000);
delay_ms(2000);
fly_plane('z', 0, 1000);
delay_ms(2000);
fly_plane('y', 0, 1000);
delay_ms(2000);
fly_plane('x', 0, 1000);
delay_ms(2000);
break;
case 12:
/* Fade in and out at low framerate. */
blinky2();
break;
case 13:
/* Random walk */
random_walk(4, 250, 1000);
break;
case 14:
/* Spinning square */
spinning_square(1, 50, 1000);
break;
}
}
std::pair<ptrobot2D_test_world::point_type, bool> ptrobot2D_test_world::random_back_walk(const ptrobot2D_test_world::point_type& p_u) const {
return random_walk(p_u);
};
int main() {
GraphTraits::out_edge_iterator out_i, out_end;
GraphTraits::edge_descriptor e;
boost::graph_traits<Graph>::adjacency_iterator ai;
boost::graph_traits<Graph>::adjacency_iterator ai_end;
boost::minstd_rand gen;
std::map<uint64_t, uint32_t> m;
std::map<int_type, int_type> id_map;
size_t count = 0;
uint64_t *v_array = (uint64_t *)malloc(sizeof(uint64_t)*(n_count + n_count*3));
// Create graph with 100 nodes and edges with probability 0.05
double prob = n_count*n_count;
prob = n_count*3/prob;
Graph g(ERGen(gen, n_count, prob), ERGen(), n_count);
std::cout << "done generating\n\n";
IndexMap index = get(boost::vertex_index, g);
cow_trie_p my_map = cow_trie_alloc(0, 0);
my_map->ref_count = 1;
my_map->child_bitmap = 0;
my_map->value_bitmap = 0;
uint32_t id = 0;
std::cout << "beginning vertex iteration\n\n";
std::pair<vertex_iter, vertex_iter> vp;
for (vp = vertices(g); vp.first != vp.second; ++vp.first) {
Vertex v = *vp.first;
boost::tie(out_i, out_end) = out_edges(v, g);
uint32_t deg = out_degree(v, g);
uint32_t *succs = new uint32_t[deg];
val_t my_val = cow_trie_create_val(0, deg, succs, 0, 0);
int_type my_id = cow_trie_insert(my_map, my_val, &my_map);
id_map[id] = my_id;
++id;
}
std::cout << "vertex iteration done, beginning edge iteration\n\n";
boost::graph_traits<Graph>::edge_iterator ei, ei_end;
boost::tie(ei, ei_end) = edges(g);
std::cout << "id before is" << id << "\n\n";
for (ei; ei != ei_end; ++ei) {
val_t my_val = cow_trie_create_val(1, 0, 0, id_map[index[source(*ei, g)]], id_map[index[target(*ei, g)]]);
int_type my_id = cow_trie_insert(my_map, my_val, &my_map);
m[index[source(*ei, g)] << 32 | index[target(*ei, g)]] = id;
++id;
}
std::cout << "id is\n\n" << id << "\n\n";
std::cout << "done edge iterating\n\n";
std::cout << "putting in successors\n\n";
int p = 0;
for (vp = vertices(g); vp.first != vp.second; ++vp.first) {
Vertex v = *vp.first;
val_t my_val;
cow_trie_lookup(my_map, p, &my_val);
uint32_t *succs = edge_list_values(my_val._.node.succs);
boost::tie(out_i, out_end) = out_edges(v, g);
int eindex = 0;
for (out_i; out_i != out_end; ++out_i) {
e = *out_i;
Vertex src = source(e, g), targ = target(e, g);
uint32_t id = m[index[src] << 32 | index[targ]];
succs[eindex] = id_map[id];
++eindex;
}
++p;
}
val_t v;
cow_trie_lookup(my_map, 0, &v);
Vertex vtest = *vertices(g).first;
std::cout << "beginning random traversal\n\n";
float t1 = random_walk(vtest, index, g);
printf("time for theirs %f", t1);
float t2 = random_walk_c(v, my_map);
printf("time for ours %f", t2);
std::cout << "end random traversal\n\n";
}
void creature_t::update()
{
if (flag & CF_SPELLCASTER)
{
if (spell_timer > 0)
spell_timer--;
}
if (get_active_effects_for(this) & EF_CONFUSE)
{
random_walk();
}
else if (identity == IDENT_SHOPKEEPER)
{
update_shopkeeper(this);
}
else if (flag & CF_NEUTRAL)
{
random_walk();
}
else if ((flag & CF_HIDES) && (!(get_active_effects_for(this) & EF_HIDDEN)) && !player.sees(this))
{
hide_creature(this);
}
else if (flag & CF_PLAYER_ALLY)
{
creature_t* closest_enemy = get_closest_enemy();
if (closest_enemy && sees(closest_enemy))
{
if ((flag & CF_SPELLCASTER) && spell_timer == 0)
{
creature_cast_spell(this, closest_enemy);
}
else
{
step_toward(closest_enemy->pos.x, closest_enemy->pos.y, closest_enemy);
}
}
else
{
_regular_walk_around(this);
}
}
else if (sees(&player))
{
if (seen_player == 0)
{
// Shout should depend on if the creature is intelligent etc.
if ((flag & CF_SPEAKS)&& !(flag & CF_SPELLCASTER))
{
shout("");
}
seen_player = 1;
}
forget_player_timer = FORGETFULNESS;
if ((flag & CF_SPELLCASTER) && spell_timer == 0)
{
creature_cast_spell(this, &player);
}
else
{
bool move_towards = true;
if (flag & CF_INTELLIGENT)
{
if (_try_zap_wand(this))
{
move_towards = false;
}
}
if (move_towards)
{
step_toward(player.pos.x, player.pos.y, nullptr);
}
}
// Stop being interested in whatever noise it heard.
heard_noise = 0;
}
else
{
// If the spell timer has reached 0 while the creature is outside player view, reset
// it for a random value to make things more varied when a caster enters view.
if (spell_timer == 0)
{
spell_timer = random(0, spell_timer_duration);
}
if (seen_player && forget_player_timer > 0)
{
forget_player_timer--;
step_toward(player.pos.x, player.pos.y, nullptr);
}
else
{
seen_player = 0;
if (heard_noise)
{
step_toward(noise_location.x, noise_location.y, nullptr);
if (pos.x == noise_location.x && pos.y == noise_location.y)
{
// Found the position of the noise.
heard_noise = 0;
}
}
else
{
// If the player has allies and I can see one, attack it.
creature_t* closest_enemy = get_closest_enemy();
if (closest_enemy && sees(closest_enemy))
{
step_toward(closest_enemy->pos.x, closest_enemy->pos.y, closest_enemy);
}
else
{
_regular_walk_around(this);
}
}
}
}
ap -= speed;
}
void DropletCustomOne::DropletMainLoop()
{
switch ( state )
{
case COLLABORATING:
if ( get_32bit_time() - collab_time > COLLABORATE_DURATION )
{
change_state ( LEAVING );
}
break;
case LEAVING:
if ( !is_moving(NULL) )
{
change_state ( SAFE );
}
break;
case SAFE:
if ( check_safe () )
{
change_state ( SEARCHING );
}
// If you start in the red region then try to get out before you die
else
{
random_walk ();
}
break;
case SEARCHING:
random_walk ();
if ( !check_safe() )
{
change_state ( WAITING );
}
break;
case START_DELAY:
{
uint32_t curr_time = get_32bit_time ();
if ( curr_time - start_delay_time > START_DELAY_TIME )
change_state ( SAFE );
}
break;
case WAITING:
if ( get_32bit_time() - heartbeat_time > HEART_RATE )
{
heartbeat_time = get_32bit_time ();
send_heartbeat ();
}
// Checks incoming messages and updates group size.
// There is a chance the state can be changed to COLLABORATING in
// this function if the droplet sees a GO message.
update_group_size ();
if ( get_32bit_time() - voting_time > HEART_RATE &&
state == WAITING )
{
voting_time = get_32bit_time ();
check_votes ();
}
break;
default:
break;
}
}
