/** Initialize the SSL context.
* \return pointer to SSL context object.
*/
SSL_CTX *
ssl_init(char *private_key_file, char *ca_file, int req_client_cert)
{
const SSL_METHOD *meth; /* If this const gives you a warning, you're
using an old version of OpenSSL. */
unsigned char context[128];
DH *dh;
unsigned int reps = 1;
if (!bio_err) {
if (!SSL_library_init())
return NULL;
SSL_load_error_strings();
/* Error write context */
bio_err = BIO_new_fp(stderr, BIO_NOCLOSE);
}
do_rawlog(LT_ERR, "Seeding OpenSSL random number pool.");
while (!RAND_status()) {
/* At this point, a system with /dev/urandom or a EGD file in the usual
places will have enough entropy. Otherwise, be lazy and use random numbers
until it's satisfied. */
uint32_t gibberish[4];
int n;
for (n = 0; n < 4; n++)
gibberish[n] = gen_rand32();
RAND_seed(gibberish, sizeof gibberish);
reps += 1;
}
do_rawlog(LT_ERR, "Seeded after %u %s.", reps, reps > 1 ? "cycles" : "cycle");
/* Set up SIGPIPE handler here? */
/* Create context */
meth = SSLv23_server_method();
ctx = SSL_CTX_new(meth);
/* Load keys/certs */
if (private_key_file && *private_key_file) {
if (!SSL_CTX_use_certificate_chain_file(ctx, private_key_file)) {
ssl_errordump
("Unable to load server certificate - only anonymous ciphers supported.");
}
if (!SSL_CTX_use_PrivateKey_file(ctx, private_key_file, SSL_FILETYPE_PEM)) {
ssl_errordump
("Unable to load private key - only anonymous ciphers supported.");
}
}
/* Load trusted CAs */
if (ca_file && *ca_file) {
if (!SSL_CTX_load_verify_locations(ctx, ca_file, NULL)) {
ssl_errordump("Unable to load CA certificates");
} else {
if (req_client_cert)
SSL_CTX_set_verify(ctx,
SSL_VERIFY_PEER | SSL_VERIFY_FAIL_IF_NO_PEER_CERT,
client_verify_callback);
else
SSL_CTX_set_verify(ctx, SSL_VERIFY_PEER, client_verify_callback);
#if (OPENSSL_VERSION_NUMBER < 0x0090600fL)
SSL_CTX_set_verify_depth(ctx, 1);
#endif
}
}
SSL_CTX_set_options(ctx, SSL_OP_SINGLE_DH_USE | SSL_OP_ALL);
SSL_CTX_set_mode(ctx,
SSL_MODE_ENABLE_PARTIAL_WRITE |
SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER);
/* Set up DH callback */
dh = get_dh1024();
SSL_CTX_set_tmp_dh(ctx, dh);
/* The above function makes a private copy of this */
DH_free(dh);
/* Set the cipher list to the usual default list, except that
* we'll allow anonymous diffie-hellman, too.
*/
SSL_CTX_set_cipher_list(ctx, "ALL:ADH:RC4+RSA:+SSLv2:@STRENGTH");
/* Set up session cache if we can */
strncpy((char *) context, MUDNAME, 128);
SSL_CTX_set_session_id_context(ctx, context, u_strlen(context));
/* Load hash algorithms */
OpenSSL_add_all_digests();
return ctx;
}
// NormFloat64 returns a normally distributed float64 in the range
// [-math.MaxFloat64, +math.MaxFloat64] with
// standard normal distribution (mean = 0, stddev = 1).
// To produce a different normal distribution, callers can
// adjust the output using:
//
// sample = NormFloat64() * desiredStdDev + desiredMean
//
real_t rstdnorm_zig(void) {
for (;;) {
int32_t j = gen_rand32(); // Possibly negative
int32_t i = j & 0x7F;
real_t x = (real_t)(j) * (real_t)(wn[i]);
if (absInt32(j) < kn[i]) {
// This case should be hit better than 99% of the time.
return x;
}
if (i == 0) {
// This extra work is only required for the base strip.
for (;;) {
x = -log(runif()) * (1.0 / rn);
real_t y = -log(runif());
if (y+y >= x*x) {
break;
}
}
if (j > 0) {
return rn + x;
}
return -rn - x;
}
if (fn[i]+runif()*(fn[i-1]-fn[i]) < exp(-.5*x*x)) {
return x;
}
}
errx(EXIT_FAILURE,"Unreachable: %s (%s:%d)",__func__,__FILE__,__LINE__);
}
char get_random_card(char maxv) {
#ifdef USE_MERSENNE_TWISTER
#ifdef USE_BUFFER
if (bp >= BUFFER_SIZE) {
fill_array32((uint32_t*) buffer, BUFFER_SIZE/4);
bp = 0;
}
return ((buffer[bp++] & 127) >> 1) % maxv;
#else
return (char)(gen_rand32() % maxv);
#endif
#else
return (char) (rand() % maxv);
#endif
}
int Book::GetMove( Shogi* pshogi, MOVE& move, TYPE type ){
/************************************************
定跡から指し手を探す。
************************************************/
unsigned i;
int flag;
list<BLIST>::iterator ib;
list<MLIST>::iterator im;
int cnt;
uint64 hash = pshogi->GetHash();
MOVE mtemp;
i = (unsigned)( hash & BOOK_HASH_MASK );
// 局面が既にあるか調べる。
flag = 0;
for( ib = blist[i].begin() ; ib != blist[i].end() ; ++ib ){
if( (*ib).hash == hash ){
flag = 1;
break;
}
}
if( flag == 0 )
return 0;
// 指し手を決定
switch( type ){
// 評価値を用いる場合
case TYPE_EVAL:
{
list<MLIST>* pmlist = &(*ib).mlist;
int vmax = (*pmlist->begin()).val - 1;
for( im = pmlist->begin() ; im != pmlist->end() ; ++im ){
mtemp.Import( (*im).mv );
if( (*im).val > vmax && pshogi->IsLegalMove( mtemp ) ){
vmax = (*im).val;
move = mtemp;
}
}
return 1;
}
break;
// 出現頻度を用いる場合
case TYPE_FREQ:
default:
list<MLIST>* pmlist = &(*ib).mlist;
cnt = (int)( gen_rand32() % (*ib).cnt );
for( im = pmlist->begin() ; im != pmlist->end() ; ++im ){
cnt -= (*im).cnt;
if( cnt < 0 ){
mtemp.Import( (*im).mv );
if( pshogi->IsLegalMove( mtemp ) ){
move = mtemp;
return 1;
}
}
}
break;
}
return 0;
}
int main(int argcs, char* pArgs[])
{
// Variable declarations.
uint64_t i = 0, y = 0, q = 0, k = 0;
uint64_t line_length = 0;
uint64_t seq_min_id = 0, seq_max_id = 0;
uint64_t is_sample_repeat = 0;
char * line_ptr = NULL;
const uint64_t MAX_HEADER_LENGTH = 1024;
const uint64_t ymax = 100;
const uint64_t SAMPLE_SIZE = 50;
char ** db_seq = NULL;
char ** seq_headers = NULL;
uint64_t * seq_len = NULL;
uint64_t * seq_alloc = NULL;
struct node * sequences = NULL;
//int num_buckets = 0;
uint64_t current_seq = 0;
uint64_t num_seq_total = 0;
uint64_t num_seq_selected = 0;
uint64_t total_db_length = 0;
uint64_t local_seq_length = 0;
uint64_t line_return = 0;
uint64_t seq_len_i=0, seq_i_bin=0;
uint64_t seq_min = 0, seq_avg = 0, seq_max = 0;
double seq_avg_d = 0.0, delta_x = 0.0;
double variance = 0.0, std_dev = 0.0;
double s_variance = 0.0, s_std_dev = 0.0;
double s_mean_d = 0.0;
uint64_t s_mean = 0;
uint64_t random_prot = 0, seed = 0;
uint64_t seq_id = 0, seq_len_remaining = 0;
uint64_t num_seq_alloc = 0;
//uint64_t sample_query[SAMPLE_SIZE];
uint64_t * sample_query = NULL;
uint64_t sample_size = 0;
uint64_t got_newline = 0;
uint64_t current_space = 0;
char line[MAX_HEADER_LENGTH];
FILE * fasta_db_in = NULL;
FILE * fasta_db_out = NULL;
uint64_t filter_length=0;
uint64_t has_filter = 0;
uint64_t sequence_count = 0;
uint64_t begin_range = 0;
uint64_t end_range = 0;
uint64_t has_range = 0;
uint64_t target_letters = 0;
const char* const short_options = "hi:o:s:f:b:e:t:";
const struct option long_options[] = {
{ "help", 0, NULL, 'h' },
{ "input" , 1, NULL, 'i' },
{ "output", 1, NULL, 'o' },
{ "seqs", 1, NULL, 's' },
{ "filter", 0, NULL, 'f' },
{ "begin", 0, NULL, 'b' },
{ "end", 0, NULL, 'e' },
{ "total",0,NULL,'t'},
{ NULL, 0, NULL, 0 } };
int next_option = 0;
int got_in = 0;
int got_out = 0;
int got_num_seqs = 0;
int got_filter = 0;
int got_begin = 0;
int got_end = 0;
int got_total = 0;
do
{
next_option = getopt_long( argcs, pArgs, short_options, long_options, NULL );
switch( next_option )
{
case 'h':
print_usage();
exit(0);
case 'i':
fasta_db_in = fopen( optarg, "r");
got_in = 1;
break;
case 'o':
fasta_db_out = fopen( optarg, "w" );
got_out = 1;
break;
case 's':
sample_size = atoi( optarg );
got_num_seqs = 1;
if ( sample_size < 1 )
{
printf("Error: Sample size is non-positive.\n");
fflush(stdout);
exit(1);
}
break;
case 'f':
filter_length = atoi(optarg);
got_filter = 1;
has_filter = 1;
printf("Filtering on sequences of length %d\n", filter_length);
break;
case 'b':
begin_range = atoi(optarg);
got_begin = 1;
has_range = 1;
break;
case 'e':
end_range = atoi(optarg);
got_end = 1;
has_range = 1;
break;
case 't':
target_letters = atoi(optarg);
got_total = 1;
printf("Targeting to sample up to %d total letters.\n",target_letters);
break;
case '?':
printf("Error: Invalid command line argument!\n");
print_usage();
exit(1);
case -1:
break;
default:
printf("Error: Something strange happened.\n");
fflush(stdout);
exit(1);
}
}
while ( next_option != -1 );
if ( !got_in )
{
printf("Error: No input specified!\n");
print_usage();
exit(1);
}
if ( !got_out )
{
printf("Error: No output specified!\n");
print_usage();
exit(1);
}
// Initialize the database array to have 50,000 sequences each with an intial length of 250.
printf("Initialize the sequence array... ");
num_seq_alloc = 50000;
db_seq = (char**) malloc( 50000 * sizeof(char*));
if ( db_seq == NULL )
{ printf("\nError: Could not allocate db_seq.\n"); exit(1); }
for ( i=0; i < 50000; ++i )
{
db_seq[i] = NULL;
db_seq[i] = (char*) malloc( 250 * sizeof(char) );
if ( db_seq[i] == NULL )
{ printf("\nError: Could not allocate db_seq[%llu].\n",i); exit(1); }
}
printf("Done\n");
// Initialize the headers for the database sequences.
printf("Initialize the sequence header array... ");
seq_headers = (char**) malloc( 50000 * sizeof(char*));
if ( seq_headers == NULL )
{ printf("\nError: Could not allocate seq_headers.\n"); exit(1); }
for ( i=0; i < 50000; ++i )
{
seq_headers[i] = NULL;
seq_headers[i] = (char*) malloc( MAX_HEADER_LENGTH * sizeof(char) );
if ( seq_headers[i] == NULL )
{ printf("\nError: Could not allocate seq_headers[%llu].\n",i); exit(1); }
}
printf("Done\n");
// Initialize the allocated length of the database sequences.
printf("Initialize the sequence allocated length array... ");
seq_alloc = (uint64_t*) malloc( 50000 * sizeof(uint64_t));
if ( seq_alloc == NULL )
{ printf("\nError: Could not allocate seq_alloc.\n"); exit(1); }
for ( i=0 ; i < 50000; ++i )
{
seq_alloc[i] = 250;
}
printf("Done\n");
// Initialize the sequence length of the database sequences.
printf("Initialize the sequence length array... ");
seq_len = (uint64_t*) malloc( 50000 * sizeof(uint64_t));
if ( seq_len == NULL )
{ printf("\nError: Could not allocate seq_len.\n"); exit(1); }
for ( i=0 ; i < 50000; ++i )
{
seq_len[i] = 0;
}
printf("Done\n");
// Read in the file one line at a time.
printf("Reading in the database... ");
current_seq = -1;
if ( (line_ptr = fgets( line, MAX_HEADER_LENGTH-1, fasta_db_in )) == NULL )
{ printf("Error: Could not get the first line of file.\n"); exit(1); }
while ( line_ptr != NULL )
{
// Check if this is a new sequence.
if ( line[0] == '>' ) // Start of a new sequence.
{
++current_seq;
// Check that there is enough memory. Allocate more if needed.
if ( current_seq >= num_seq_alloc )
{
num_seq_alloc += (uint64_t)50000;
db_seq = (char**) realloc( (void*) db_seq, num_seq_alloc * sizeof(char*) );
if ( db_seq == NULL )
{ printf("Error: Reallocation of db_seq failed.\n"); exit(1); }
for ( i=(num_seq_alloc-50000); i < num_seq_alloc; ++i )
{
db_seq[i] = NULL;
db_seq[i] = (char*) malloc( 250 * sizeof(char) );
if ( db_seq[i] == NULL )
{ printf("\nError: Could not allocate db_seq[%llu].\n",i); exit(1); }
}
seq_headers = (char**) realloc( (void*) seq_headers, num_seq_alloc * sizeof(char*) );
if ( seq_headers == NULL )
{ printf("Error: Reallocation of seq_headers failed.\n"); exit(1); }
for ( i=(num_seq_alloc-50000); i < num_seq_alloc; ++i )
{
seq_headers[i] = NULL;
seq_headers[i] = (char*) malloc( MAX_HEADER_LENGTH * sizeof(char) );
if ( seq_headers[i] == NULL )
{ printf("\nError: Could not allocate seq_headers[%llu].\n",i); exit(1); }
}
seq_alloc = (uint64_t*) realloc( (void*) seq_alloc, num_seq_alloc * sizeof(uint64_t) );
if ( seq_alloc == NULL )
{ printf("Error: Could not reallocate seq_alloc.\n"); exit(1); }
for ( i=(num_seq_alloc-50000); i < num_seq_alloc; ++i )
{ seq_alloc[i] = 250; }
seq_len = (uint64_t*) realloc( (void*) seq_len, num_seq_alloc * sizeof(uint64_t) );
if ( seq_len == NULL )
{ printf("Error: Could not reallocate seq_len.\n"); exit(1); }
for ( i=(num_seq_alloc-50000); i < num_seq_alloc; ++i )
{ seq_len[i] = 0; }
}
// check that we got a new line character.
got_newline = 0;
for ( i=0; i < MAX_HEADER_LENGTH; ++i )
{
if ( line[i] == '\n' )
got_newline = 1;
}
// Copy the header information.
strncpy( seq_headers[current_seq], line, MAX_HEADER_LENGTH-1 );
current_space = MAX_HEADER_LENGTH;
while ( !got_newline )
{
// Grab the next line
line_ptr = fgets( line, MAX_HEADER_LENGTH-1, fasta_db_in );
// Allocate more space
seq_headers[current_seq] = (char*)realloc( (void*)seq_headers[current_seq], current_space+MAX_HEADER_LENGTH);
// Append the end.
strncpy( &(seq_headers[current_seq][current_space]), line, MAX_HEADER_LENGTH-1 );
current_space += MAX_HEADER_LENGTH;
for ( i=0; i < MAX_HEADER_LENGTH; ++i )
{
if ( line[i] == '\n' )
got_newline = 1;
}
}
}
else
{
// Any other line should have nucleotides or amino acids on it. Put them into the sequence array.
line_length = strlen(line);
if ( line_length > (MAX_HEADER_LENGTH-2) )
{
printf("ERROR: Line length for sequence line is too long: %llu\n",line_length);
fflush(stdout);
exit(1);
}
// Check that the current line can be written to the sequence array. Add more memory if not.
if ( seq_len[current_seq] + line_length >= seq_alloc[current_seq] )
{
//printf("current_seq=%llu seq_alloc= %llu \n",current_seq, seq_alloc[current_seq]);
//fflush(stdout);
//printf("seq_alloc[current_seq] + 250 = %llu\n", seq_alloc[current_seq] + 250 );
//fflush(stdout);
// Allocate additional space for the sequence.
//printf(" Realloc size requested is seq_alloc[%llu] + 250 = %llu + 250 = %llu\n",current_seq,seq_alloc[current_seq],seq_alloc[current_seq] + 250);
//printf(" REALLOC SIZE - (seq_alloc[current_seq] + 250 ) * (uint64_t)(sizeof(char)) = %llu \n",(seq_alloc[current_seq] + 250 ) * (uint64_t)(sizeof(char)));
fflush(stdout);
db_seq[current_seq] = (char*) realloc( (void*) db_seq[current_seq], (seq_alloc[current_seq] + 250 ) );
if ( db_seq[current_seq] == NULL )
{ printf("Error: Could not reallocate db_seq[%llu].\n",current_seq); exit(1); }
seq_alloc[current_seq] += 250;
}
if ( line[line_length-1] == '\n' )
{ line[line_length-1] = '\0'; }
// Checked length already so assume its safe to pop the current line into the sequence.
// Start from the end of the sequence ( squash the terminator already there )
strcpy( &(db_seq[current_seq][seq_len[current_seq]]), line );
seq_len[current_seq] += (line_length-1);
}
// Read in the next line for next iteration.
line_ptr = fgets( line, MAX_HEADER_LENGTH-1, fasta_db_in );
}
num_seq_total = current_seq+1;
// Close the file.
fclose(fasta_db_in);
printf("Done.\n");
printf("Total number of sequences in the database is %llu.\n", num_seq_total);
if ( sample_size >= num_seq_total )
{
printf("Error: The sample size is too large. <%llu>\n",sample_size);
fflush(stdout);
exit(1);
}
sample_query = (uint64_t*) malloc(sizeof(uint64_t) * sample_size);
if ( sample_query == NULL )
{ printf("\nError: Could not allocate sample_query.\n"); exit(1); }
for ( i=0; i < sample_size; i++ )
sample_query[i] = -1; // Initialize to a false sequence identity.
// Fill in the histogram array.
seq_min = (uint64_t)1<<52;
seq_max = 0;
for ( i=0; i < num_seq_total; ++i )
{
seq_avg += seq_len[i];
if ( seq_min > seq_len[i] )
{
seq_min = seq_len[i];
seq_min_id = i;
}
if ( seq_max < seq_len[i] )
{
seq_max = seq_len[i];
seq_max_id = i;
}
}
seq_avg_d = ( (double)seq_avg )/( (double) num_seq_total );
seq_avg = (uint64_t)seq_avg_d;
// Compute the standard deviation and variance.
for ( i=0; i < num_seq_total; i++ )
{
variance += ( (double)seq_len[i] - seq_avg_d ) * ( (double)seq_len[i] - seq_avg_d );
}
variance = variance * (1./( (double) (num_seq_total-1) ));
std_dev = sqrt(variance);
printf("Average sequence length is %lf ( or %llu ).\n", seq_avg_d, seq_avg );
printf("Minimum sequence length is %llu.\n", seq_min);
printf("Maximum sequence length is %llu.\n", seq_max);
printf("\n");
printf("Standard deviation is %lf.\n",std_dev);
printf("Variance is %lf.\n",variance);
if ( has_range )
{
if ( !got_begin )
{
begin_range = seq_min;
}
if ( !got_end )
{
end_range = seq_max;
}
printf("Filtering on sequences in the range of [%d , %d]\n",begin_range,end_range);
}
// Sort the sequences according to lengths.
printf("Initialize the sequence node array... ");
sequences = (struct node *) malloc( num_seq_total * sizeof(struct node));
if ( sequences == NULL )
{ printf("\nError: Could not allocate sequences node array.\n"); exit(1); }
sequence_count = 0;
double window = 0.0;
uint64_t lower_bound = 0;
uint64_t upper_bound = 0;
uint64_t midpoint = 0;
uint64_t min_seqs = 5 * sample_size;
if ( has_range )
midpoint = ( end_range - begin_range )/2;
if ( has_filter )
{
lower_bound = filter_length - ( (uint64_t) window * filter_length);
upper_bound = filter_length + ( (uint64_t) window * filter_length);
}
else if ( has_range )
{
lower_bound = begin_range - ( (uint64_t) window * midpoint);
upper_bound = end_range + ( (uint64_t) window * midpoint);
}
else
{
lower_bound = seq_min;
upper_bound = seq_max;
}
while ( sequence_count < min_seqs )
{
sequence_count = 0;
for ( i=0; i < num_seq_total; ++i )
{
if ( has_filter || has_range )
{
if ( seq_len[i] >= lower_bound &&
seq_len[i] <= upper_bound )
{
sequences[sequence_count].id = i;
sequences[sequence_count].length = seq_len[i];
++sequence_count;
}
}
else
{
sequences[sequence_count].id = i;
sequences[sequence_count].length = seq_len[i];
++sequence_count;
}
}
printf("Attempting to sample %llu sequences in the range of [%llu , %llu]. Found %llu in the range.\n",sample_size,lower_bound,upper_bound,sequence_count);
if ( (has_filter || has_range) && sequence_count < min_seqs )
{
window = window + 0.005;
lower_bound = filter_length - ( (uint64_t) (window * filter_length));
upper_bound = filter_length + ( (uint64_t) (window * filter_length));
if ( window > 0.06 ) // Ok give up.
{
fprintf(stderr,"Error, window is getting too big so I'm giving up.\n");
fflush(stderr);
exit(1);
}
printf("Adding 0.005 to the filter length window, now at %lf : ( %llu , %llu )\n",window,lower_bound,upper_bound);
}
if ( !(has_filter || has_range) && sequence_count == 0 )
{
fprintf(stderr,"ERROR: Apparently the database is empty.\n");
fflush(stderr);
exit(1);
}
}
printf("Done\n");
if ( !(has_filter || has_range) )
{
printf("Starting the sort algorithm... ");
// Sort the nodes by sequence length.
qsort ( sequences , num_seq_total, sizeof(struct node), compare_length );
printf("Done\n");
}
else if ( has_filter )
{
printf("Found %llu sequences of, or around, length %llu in the database.\n",sequence_count, filter_length);
}
else
{
printf("Found %llu sequences in the range of [%llu , %llu] in the database.\n",sequence_count,lower_bound,upper_bound);
}
//for (i=0;i<num_seq_total;i++)
// {
// printf("%llu %llu\n",sequences[i].length, sequences[i].id);
// }
// Initialize the psuedo-random number generator.
//srand48(time(NULL));
//random number = drand48();
//mt_seed();
//srand(time(NULL));
seed = time(NULL);
//init_gen_rand(1384375234u);
//seed = 1384375254u;
init_gen_rand(seed);
printf("Seed used = %u\n", seed);
uint64_t letters_collected = 0;
uint64_t samples_taken = 0;
for ( y=0; y < sample_size ; ++y )
{
is_sample_repeat = 1;
while ( is_sample_repeat )
{
// Assume that it is not a repeat.
is_sample_repeat = 0;
// Fill the data sample.
random_prot = gen_rand32();
//printf("Got random number -> %llu\n",random_prot);
random_prot = random_prot % sequence_count;
//printf("Modulo number of sequences -> %llu\n", random_prot);
if ( random_prot >= sequence_count || random_prot < 0 )
{
printf("Error : random_prot = %llu\n", random_prot);
exit(1);
}
// check if sample is a repeat.
for (k=0; k < y; ++k )
{
if ( random_prot == sample_query[k] )
is_sample_repeat = 1;
}
}
sample_query[y] = random_prot;
++samples_taken;
letters_collected += sequences[random_prot].length ;
if ( got_total )
{
if ( letters_collected > target_letters )
y = sample_size; // exit early
}
}
// possibly reset sample size
sample_size = samples_taken;
if ( got_total )
{
printf("Got a sample with %d letters and %d sequences.\n",letters_collected,sample_size);
}
// Compute relevant statistics.
seq_min = (uint64_t)1<<52;
seq_max = 0;
seq_avg = 0;
for ( y=0; y < sample_size; ++y )
{
q = sample_query[y];
seq_avg += sequences[q].length;
if ( seq_min > sequences[q].length )
{
seq_min = sequences[q].length;
seq_min_id = q;
}
if ( seq_max < sequences[q].length )
{
seq_max = sequences[q].length;
seq_max_id = q;
}
}
seq_avg_d = ( (double)seq_avg )/( (double) sample_size );
seq_avg = (uint64_t)seq_avg_d;
// Compute the standard deviation and variance.
for ( y=0; y < sample_size; y++ )
{
q = sample_query[y];
variance += ( (double)sequences[q].length - seq_avg_d ) *
( (double)sequences[q].length - seq_avg_d );
}
variance = variance * (1./( (double) (sample_size) ));
std_dev = sqrt(variance);
printf(" Average sequence length is %lf ( or %llu ).\n", seq_avg_d, seq_avg );
printf(" Minimum sequence length is %llu.\n", seq_min);
printf(" Maximum sequence length is %llu.\n", seq_max);
printf(" Standard deviation is %lf.\n",std_dev);
printf(" Variance is %lf.\n\n",variance);
for ( i=0; i < sample_size; ++i )
{
// Retrieve sequence id.
seq_id = sample_query[i];
seq_id = sequences[seq_id].id;
// Write the sequence header.
fprintf(fasta_db_out,"%s", seq_headers[seq_id]);
// Write out the amino acids.
seq_len_remaining = seq_len[seq_id];
for ( y=0; y < seq_len[seq_id]; y+=60 , seq_len_remaining-=60 )
{
if ( seq_len_remaining >= 60 )
{
strncpy( line, &( db_seq[seq_id][y] ), 60 );
line[60] = '\0';
}
else
{
strncpy( line, &( db_seq[seq_id][y] ), seq_len_remaining );
line[seq_len_remaining] = '\0';
}
fprintf(fasta_db_out,"%s\n", line);
}
}
// Close the file.
fclose(fasta_db_out);
// Clean out the memory.
for ( i=0; i < num_seq_total; ++i )
{
free(db_seq[i]);
free(seq_headers[i]);
}
free(db_seq);
free(seq_headers);
free(seq_len);
free(seq_alloc);
free(sequences);
free(sample_query);
return 0;
}