static void compute_sum_dcov_thread(int ni,int ki,int di,
const float *v,const float *mu_old,const float *p,
float *mu,float *sigma,float *w,
int n_thread) {
long n=ni,d=di,k=ki;
compute_sum_dcov_t t= {
n,k,d,
v,mu_old,p,
fvec_new(n_thread*d*k), /* mu */
fvec_new(n_thread*d*k), /* sigma */
fvec_new(n_thread*k), /* w */
n_thread
};
compute_tasks(n_thread,n_thread,&compute_sum_dcov_task_fun,&t);
/* accumulate over n's */
long i;
fvec_cpy(mu,t.mu,k*d);
fvec_cpy(sigma,t.sigma,k*d);
fvec_cpy(w,t.w,k);
for(i=1; i<n_thread; i++) {
fvec_add(mu,t.mu+i*d*k,d*k);
fvec_add(sigma,t.sigma+i*d*k,d*k);
fvec_add(w,t.w+i*k,k);
}
free(t.mu);
free(t.sigma);
free(t.w);
}
/*
* A simple test for the binary embedding
*/
int test_embed_tfidf()
{
int i, j, n, err = 0;
string_t strs[10];
config_set_string(&cfg, "features.vect_norm", "none");
config_set_string(&cfg, "features.tfidf_file", TEST_TFIDF);
unlink(TEST_TFIDF);
char *test_file = getenv("TEST_FILE");
idf_create(test_file);
test_printf("Testing TFIDF embedding");
input_config("lines");
n = input_open(test_file);
input_read(strs, n);
/* Compute IDF manually */
config_set_string(&cfg, "features.vect_embed", "bin");
fvec_t *w = fvec_zero();
for (i = 0, err = 0; i < n; i++) {
fvec_t *fv = fvec_extract(strs[i].str, strs[i].len);
fvec_add(w, fv);
fvec_destroy(fv);
}
fvec_invert(w);
fvec_mul(w, n);
fvec_log2(w);
if (!idf_check(w)) {
err++;
test_error("(%d) internal idf values seem to be wrong", i);
}
/* Invert w for multiplying out IDFs */
fvec_invert(w);
config_set_string(&cfg, "features.vect_embed", "tfidf");
for (i = 0, err = 0; i < n; i++) {
fvec_t *fv = fvec_extract(strs[i].str, strs[i].len);
fvec_times(fv, w);
/* Check if rest tf */
double d = 0;
for (j = 0; j < fv->len; j++)
d += fv->val[j];
err += fabs(d - 1.0) > 1e-6;
fvec_destroy(fv);
}
test_return(err, n);
fvec_destroy(w);
input_free(strs, n);
input_close();
idf_destroy();
unlink(TEST_TFIDF);
return err;
}
void
aubio_pitchmcomb_spectral_pp (aubio_pitchmcomb_t * p, fvec_t * newmag)
{
fvec_t *mag = (fvec_t *) p->scratch;
fvec_t *tmp = (fvec_t *) p->scratch2;
uint_t j;
uint_t length = mag->length;
/* copy newmag to mag (scracth) */
for (j = 0; j < length; j++) {
mag->data[j] = newmag->data[j];
}
fvec_min_removal (mag); /* min removal */
fvec_alpha_normalise (mag, p->alpha); /* alpha normalisation */
/* skipped *//* low pass filtering */
/** \bug fvec_moving_thres may write out of bounds */
fvec_adapt_thres (mag, tmp, p->win_post, p->win_pre); /* adaptative threshold */
fvec_add (mag, -p->threshold); /* fixed threshold */
{
aubio_spectralpeak_t *peaks = (aubio_spectralpeak_t *) p->peaks;
uint_t count;
/* return bin and ebin */
count = aubio_pitchmcomb_quadpick (peaks, mag);
for (j = 0; j < count; j++)
peaks[j].mag = newmag->data[peaks[j].bin];
/* reset non peaks */
for (j = count; j < length; j++)
peaks[j].mag = 0.;
p->peaks = peaks;
p->count = count;
}
}
/* Accumulate information for PCA for n input vectors */
void pca_online_accu (struct pca_online_s * pca, const float * v, long n)
{
int d = pca->d;
float * cov = fvec_new (d*(long)d);
float * mu = fvec_new (d);
fmat_sum_rows (v, d, n, mu);
fmat_mul_tr (v, v, d, d, n, cov);
fvec_add (pca->mu, mu, d);
fvec_add (pca->cov, cov, d*(long)d);
pca->n += n;
free (cov);
free (mu);
}
float *fmat_center_columns(int d,int n,float *v)
{
assert(n>0);
float *accu=fvec_new_cpy(v,d);
long i;
for(i=1;i<n;i++)
fvec_add(accu,v+i*d,d);
fvec_div_by(accu,d,n);
for(i=0;i<n;i++)
fvec_sub(v+i*d,accu,d);
return accu;
}
/*
* Internal: Allocates and extracts a feature vector from a string
* without postprocessing but blended n-grams
* @param x String of bytes (with space delimiters)
* @param l Length of sequence
* @return feature vector
*/
fvec_t *fvec_extract_intern(char *x, int l)
{
int blend;
cfg_int i, n;
/* Get config */
config_lookup_bool(&cfg, "features.ngram_blend", &blend);
config_lookup_int(&cfg, "features.ngram_len", &n);
/* Extract n-grams */
fvec_t *fv = fvec_extract_intern2(x, l, n);
/* Blended n-grams */
for (i = 1; blend && i < n; i++) {
fvec_t *fx = fvec_extract_intern2(x, l, i);
fvec_add(fv, fx);
fvec_destroy(fx);
}
return fv;
}
/*
* A stres test for the addition of feature vectors
*/
int test_stress_add()
{
int i, j, err = 0;
fvec_t *fx, *fy, *fz;
char buf[STR_LENGTH + 1];
test_printf("Stress test for addition of feature vectors");
/* Create empty vector */
fz = fvec_extract("aa0bb0cc", 8, "zero");
for (i = 0; i < NUM_VECTORS; i++) {
/* Create random key and string */
for (j = 0; j < STR_LENGTH; j++)
buf[j] = rand() % 10 + '0';
buf[j] = 0;
/* Extract features */
fx = fvec_extract(buf, strlen(buf), "test");
/* Add fx to fz */
fy = fvec_add(fz, fx);
fvec_destroy(fz);
err += fabs(fvec_norm2(fy) - 1.4142135623) > 1e-7;
/* Substract fx from fz */
fz = fvec_sub(fy, fx);
fvec_sparsify(fz);
/* Clean up */
fvec_destroy(fy);
fvec_destroy(fx);
}
fvec_destroy(fz);
test_return(err, i);
return err;
}
/*
* A simple static test for the addition of feature vectors
*/
int test_static_add()
{
int i, err = 0;
fvec_t *fx, *fy, *fz;
test_printf("Addition of feature vectors");
for (i = 0; test_add[i].x; i++) {
/* Extract features */
fx = fvec_extract(test_add[i].x, strlen(test_add[i].x), "test");
fy = fvec_extract(test_add[i].y, strlen(test_add[i].y), "test");
/* Add test vectors */
fz = fvec_add(fx, fy);
err += fabs(fvec_norm1(fz) - test_add[i].res) > 1e-7;
fvec_destroy(fz);
fvec_destroy(fx);
fvec_destroy(fy);
}
test_return(err, i);
return err;
}
void fmat_add_to_columns(int d,int n,float *v,const float *avg) {
long i;
for(i=0;i<n;i++)
fvec_add(v+i*d,avg,d);
}
void
fvec_min_removal (fvec_t * v)
{
smpl_t v_min = fvec_min (v);
fvec_add (v, - v_min );
}
/**
* Print shared n-grams for each cluster
* @param c Clustering structure
* @param fa Array of feature vectors
* @param file Output file
*/
void export_shared_ngrams(cluster_t *c, farray_t *fa, const char *file)
{
assert(c && fa && file);
int i, j, k;
double shared;
FILE *f;
char *name = NULL;
config_lookup_float(&cfg, "cluster.shared_ngrams", &shared);
if (shared <= 0.0)
return;
if (verbose > 0)
printf("Exporting shared n-grams with minimum ratio %4.2f.\n",
shared);
if (!(f = fopen(file, "a"))) {
error("Could not create file '%s'.", file);
return;
}
/* Print incremental header */
fprintf(f, "# ---\n# Shared n-grams for %s\n", fa->src);
fprintf(f, "# Minimum ratio of shared n-grams: %4.2f (%2.0f%%)\n",
shared, shared * 100);
fprintf(f, "# ---\n# <cluster> <ratio> <hash> <ngram>\n");
/* Compute shared n-grams per cluster */
for (i = 0; i < c->num; i++) {
fvec_t *s = fvec_zero();
for (j = 0, k = 0; j < c->len; j++) {
if (c->cluster[j] != i)
continue;
/* Clone and binarize */
fvec_t *x = fvec_clone(fa->x[j]);
fvec_bin(x);
if (k == 0)
name = cluster_get_name(c, j);
/* Merge n-grams in cluster */
fvec_t *y = fvec_add(s, x);
fvec_destroy(s);
fvec_destroy(x);
s = y;
k++;
}
/* Check for empty cluster */
if (k == 0)
continue;
fvec_div(s, k);
/* Output shared n-grams */
for (j = 0; j < s->len; j++) {
if (s->val[j] < shared)
continue;
fprintf(f, "%s %6.4f %.16llx ", name, s->val[j],
(long long unsigned int) s->dim[j]);
/* Lookup feature */
fentry_t *fe = ftable_get(s->dim[j]);
if (!fe)
error("Oops. Feature not in lookup table.");
/* Print feature */
fprintf(f, "\"");
for (k = 0; k < fe->len; k++) {
if (isprint(fe->data[k]) || fe->data[k] == '%')
fprintf(f, "%c", fe->data[k]);
else
fprintf(f, "%%%.2x", fe->data[k]);
}
fprintf(f, "\"\n");
}
fvec_destroy(s);
}
fclose(f);
}
/* estimate the GMM parameters */
static void gmm_compute_params (int n, const float * v, const float * p,
gmm_t * g,
int flags,
int n_thread)
{
long i, j;
long d=g->d, k=g->k;
float * vtmp = fvec_new (d);
float * mu_old = fvec_new_cpy (g->mu, k * d);
float * w_old = fvec_new_cpy (g->w, k);
fvec_0 (g->w, k);
fvec_0 (g->mu, k * d);
fvec_0 (g->sigma, k * d);
if(0) {
/* slow and simple */
for (j = 0 ; j < k ; j++) {
double dtmp = 0;
for (i = 0 ; i < n ; i++) {
/* contribution to the gaussian weight */
dtmp += p[i * k + j];
/* contribution to mu */
fvec_cpy (vtmp, v + i * d, d);
fvec_mul_by (vtmp, d, p[i * k + j]);
fvec_add (g->mu + j * d, vtmp, d);
/* contribution to the variance */
fvec_cpy (vtmp, v + i * d, d);
fvec_sub (vtmp, mu_old + j * d, d);
fvec_sqr (vtmp, d);
fvec_mul_by (vtmp, d, p[i * k + j]);
fvec_add (g->sigma + j * d, vtmp, d);
}
g->w[j] = dtmp;
}
} else {
/* fast and complicated */
if(n_thread<=1)
compute_sum_dcov(n,k,d,v,mu_old,p,g->mu,g->sigma,g->w);
else
compute_sum_dcov_thread(n,k,d,v,mu_old,p,g->mu,g->sigma,g->w,n_thread);
}
if(flags & GMM_FLAGS_1SIGMA) {
for (j = 0 ; j < k ; j++) {
float *sigma_j=g->sigma+j*d;
double var=fvec_sum(sigma_j,d)/d;
fvec_set(sigma_j,d,var);
}
}
long nz=0;
for(i=0; i<k*d; i++)
if(g->sigma[i]<min_sigma) {
g->sigma[i]=min_sigma;
nz++;
}
if(nz) printf("WARN %ld sigma diagonals are too small (set to %g)\n",nz,min_sigma);
for (j = 0 ; j < k ; j++) {
fvec_div_by (g->mu + j * d, d, g->w[j]);
fvec_div_by (g->sigma + j * d, d, g->w[j]);
}
assert(finite(fvec_sum(g->mu, k*d)));
fvec_normalize (g->w, k, 1);
printf ("w = ");
fvec_print (g->w, k);
double imfac = k * fvec_sum_sqr (g->w, k);
printf (" imfac = %.3f\n", imfac);
free (vtmp);
free (w_old);
free (mu_old);
}
void vlad_compute(int k, int d, const float *centroids, int n, const float *v,int flags, float *desc)
{
int i,j,l,n_quantile,i0,i1,ai,a,ma,ni;
int *perm ;
float un , diff;
float *tab,*u,*avg,*sum,*mom2,*dists;
int *hist,*assign;
if(flags<11 || flags>=13)
{
assign=ivec_new(n);
nn(n,k,d,centroids,v,assign,NULL,NULL);
if(flags==6 || flags==7)
{
n_quantile = flags==6 ? 3 : 1;
fvec_0(desc,k*d*n_quantile);
perm = ivec_new(n);
tab = fvec_new(n);
ivec_sort_index(assign,n,perm);
i0=0;
for(i=0;i<k;i++)
{
i1=i0;
while(i1<n && assign[perm[i1]]==i)
{
i1++;
}
if(i1==i0) continue;
for(j=0;j<d;j++)
{
for(l=i0;l<i1;l++)
{
tab[l-i0]=v[perm[l]*d+j];
}
ni=i1-i0;
fvec_sort(tab,ni);
for(l=0;l<n_quantile;l++)
{
desc[(i*d+j)*n_quantile+l]=(tab[(l*ni+ni/2)/n_quantile]-centroids[i*d+j])*ni;
}
}
i0=i1;
}
free(perm);
free(tab);
}
else if(flags==5)
{
fvec_0(desc,k*d);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
desc[assign[i]*d+j]+=v[i*d+j];
}
}
}
else if(flags==8 || flags==9)
{
fvec_0(desc,k*d);
u = fvec_new(d);
for(i=0;i<n;i++)
{
fvec_cpy(u,v+i*d,d);
fvec_sub(u,centroids+assign[i]*d,d);
un=(float)sqrt(fvec_norm2sqr(u,d));
if(un==0) continue;
if(flags==8)
{
fvec_div_by(u,d,un);
} else if(flags==9)
{
fvec_div_by(u,d,sqrt(un));
}
fvec_add(desc+assign[i]*d,u,d);
}
free(u);
}
else if(flags==10)
{
fvec_0(desc,k*d);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
desc[assign[i]*d+j]+=v[i*d+j];
}
}
for(i=0;i<k;i++)
{
fvec_normalize(desc+i*d,d,2.0);
}
}
else if(flags==13)
{
fvec_0(desc,k*d);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
desc[assign[i]*d+j]+=(float)sqr(v[i*d+j]-centroids[assign[i]*d+j]);
}
}
}
else if(flags==14)
{
avg = fvec_new_0(k*d);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
avg[assign[i]*d+j]+=v[i*d+j]-centroids[assign[i]*d+j];
}
}
hist=ivec_new_histogram(k,assign,n);
for(i=0;i<k;i++)
{
if(hist[i]>0)
{
for(j=0;j<d;j++)
{
avg[i*d+j]/=hist[i];
}
}
}
free(hist);
fvec_0(desc,k*d);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
desc[assign[i]*d+j]+=(float)(sqr(v[i*d+j]-centroids[assign[i]*d+j]-avg[assign[i]*d+j]));
}
}
fvec_sqrt(desc,k*d);
free(avg);
}
else if(flags==15)
{
fvec_0(desc,k*d*2);
sum = desc;
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
sum[assign[i]*d+j]+=v[i*d+j]-centroids[assign[i]*d+j];
}
}
hist = ivec_new_histogram(k,assign,n);
mom2 = desc+k*d;
for(i=0;i<n;i++)
{
ai=assign[i];
for(j=0;j<d;j++)
{
mom2[ai*d+j]+=(float)(sqr(v[i*d+j]-centroids[ai*d+j]-sum[ai*d+j]/hist[ai]));
}
}
fvec_sqrt(mom2,k*d);
free(hist);
}
else if(flags==17)
{
fvec_0(desc,k*d*2);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
diff=v[i*d+j]-centroids[assign[i]*d+j];
if(diff>0)
{
desc[assign[i]*d+j]+=diff;
}
else
{
desc[assign[i]*d+j+k*d]-=diff;
}
}
}
}
else
{
fvec_0(desc,k*d);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
desc[assign[i]*d+j]+=v[i*d+j]-centroids[assign[i]*d+j];
}
}
if(flags==1)
{
hist=ivec_new_histogram(k,assign,n);
/* printf("unbalance factor=%g\n",ivec_unbalanced_factor(hist,k)); */
for(i=0;i<k;i++)
{
for(j=0;j<d;j++)
{
desc[i*d+j]/=hist[i];
}
}
free(hist);
}
if(flags==2)
{
for(i=0;i<k;i++)
{
fvec_normalize(desc+i*d,d,2.0);
}
}
if(flags==3 || flags==4)
{
assert(!"not implemented");
}
if(flags==16)
{
hist=ivec_new_histogram(k,assign,n);
for(i=0;i<k;i++)
{
if(hist[i]>0)
{
fvec_norm(desc+i*d,d,2);
fvec_mul_by(desc+i*d,d,sqrt(hist[i]));
}
}
free(hist);
}
}
free(assign);
}
else if(flags==11 || flags==12)
{
ma=flags==11 ? 4 : 2;
assign=ivec_new(n*ma);
dists=knn(n,k,d,ma,centroids,v,assign,NULL,NULL);
fvec_0(desc,k*d);
for(i=0;i<n;i++)
{
for(j=0;j<d;j++)
{
for(a=0;a<ma;a++)
{
desc[assign[ma*i+a]*d+j]+=v[i*d+j]-centroids[assign[ma*i+a]*d+j];
}
}
}
free(dists);
free(assign);
}
}
