static void fill_hash_table(void)
{
int i;
for (i=0;i<TOTAL_SIZE/2;i++) {
u32 tag = random();
u32 t = tag % HASH_SIZE;
if (hash_table[t].offset == 0 ||
replace(hash_table[t].offset, i)) {
hash_table[t].offset = i;
hash_table[t].tag = tag;
}
}
print_distribution(i);
printf("\n");
for (;i<TOTAL_SIZE;i++) {
u32 tag = random();
u32 t = tag % HASH_SIZE;
if (hash_table[t].offset == 0 ||
replace(hash_table[t].offset, i)) {
hash_table[t].offset = i;
hash_table[t].tag = tag;
}
}
print_distribution(i);
}
int calibrate(int file, int width, Triplet *bias, FTriplet *scale)
{
Triplet data = {0};
MagDistribution m_dist = {{0},{0},{0}};
Triplet diameters;
int avg_diameter;
while (1) {
print_distribution (m_dist.min, m_dist.max, data, width, ' ', ' ');
read_triplet (file, XM_ADDRESS, OUT_X_L_M, &data);
update_mag_bias (data, &m_dist);
print_distribution (m_dist.min, m_dist.max, data, width, '|', '*');
move (6, 0);
refresh ();
usleep (20000);
if (getch () != ERR)
break;
}
bias->x = m_dist.bias.x;
bias->y = m_dist.bias.y;
bias->z = m_dist.bias.z;
/* Try to scale the axes (crude transform from ellipse to sphere) */
diameters.x = abs (m_dist.max.x - m_dist.min.x);
diameters.y = abs (m_dist.max.y - m_dist.min.y);
diameters.z = abs (m_dist.max.z - m_dist.min.z);
avg_diameter = (diameters.x + diameters.y + diameters.z) / 3;
scale->x = avg_diameter / (float)diameters.x;
scale->y = avg_diameter / (float)diameters.y;
scale->z = avg_diameter / (float)diameters.z;
return 1;
}
/*METHOD: CONVERT DISTRIBUTIONS INTO READ MAPPINGS - SEND TO PRINT*/
void convert_distribution(string o_file, int s_count, const map<int, string> *id2seqid, const map<int, string> *id2kmers, const map<int, int> *id2taxid, const map<int, string> *id2tandl, const taxonomy *my_taxonomy, const map<int, taxonomy *> *taxid2node, const int kmer_len, const int read_len){
/*Initialize variables for getting read mappings instead of kmer mappings */
int n_kmers = read_len - kmer_len + 1;
int seqs_read = 0;
/*Iterate over taxid2kmers in parallel*/
printf("\t>>STEP 4: CONVERTING KMER MAPPINGS INTO READ CLASSIFICATIONS:\n");
printf("\t\t%imers, with a database built using %imers\n",read_len, kmer_len);
cerr << "\t\t" << seqs_read << " sequences converted...";
int i;
#pragma omp parallel for
for(i = 1; i <= s_count; i++) {
//Get values to parse here
string curr_ks = id2kmers->find(i)->second;
//Saving values
vector<int> all_kmers;
int count_kmers = 0;
//Iterate through all of the kmer pairs
int mid, end;
string buf;
std::stringstream ss(curr_ks);
int pair_taxid, pair_count;
string curr_pair, pair_tstr;
while(ss >> buf) {
//Extract the string of this pair
curr_pair = buf;
//Split up this pair into the taxid and the number of kmers
mid = curr_pair.find(":");
end = curr_pair.find("\n");
pair_tstr = curr_pair.substr(0,mid);
pair_count = atoi(curr_pair.substr(mid+1,end-mid-1).c_str());
if (pair_tstr == "A")
pair_taxid = 0;
else
pair_taxid = atoi(pair_tstr.c_str());
//Add kmers to queue
for (int j = 0; j < pair_count; j++) {
all_kmers.push_back(pair_taxid);
count_kmers += 1;
}
}
//Process all mappings
vector<int> curr_kmers;
map<int,int> taxids_mapped;
int mapped_taxid;
int prev_kmer, next_kmer;
int prev_taxid;
for (int k = 0; k < count_kmers; k++) {
next_kmer = all_kmers[k];
curr_kmers.push_back(next_kmer);
if (curr_kmers.size() == n_kmers) {
if (prev_kmer == next_kmer) {
mapped_taxid = prev_taxid;
} else {
mapped_taxid = get_classification(&curr_kmers, my_taxonomy, taxid2node);
}
//if (mapped_taxid != 0) {
//Save to map
auto t_it = taxids_mapped.find(mapped_taxid);
if (t_it == taxids_mapped.end()){
taxids_mapped[mapped_taxid] = 1;
} else {
t_it->second += 1;
}
prev_taxid = mapped_taxid;
//Remove last element
prev_kmer = curr_kmers[0];
curr_kmers.erase(curr_kmers.begin());
}
}
//Update User
#pragma omp atomic
seqs_read += 1;
#pragma omp critical
{
cerr << "\r\t\t" << seqs_read << " sequences converted...";
print_distribution(o_file, id2seqid->find(i)->second, id2taxid->find(i)->second, id2tandl->find(i)->second, taxids_mapped);
}
}
cerr << "\r\t\t" << seqs_read << " sequences converted...\n";
}
