| 1 |
/* BaitFisher (version 1.2.8) a program for designing DNA target enrichment baits |
| 2 |
* BaitFilter (version 1.0.6) a program for selecting optimal bait regions |
| 3 |
* Copyright 2013-2017 by Christoph Mayer |
| 4 |
* |
| 5 |
* This source file is part of the BaitFisher-package. |
| 6 |
* |
| 7 |
* This program is free software: you can redistribute it and/or modify |
| 8 |
* it under the terms of the GNU General Public License as published by |
| 9 |
* the Free Software Foundation, either version 3 of the License, or |
| 10 |
* (at your option) any later version. |
| 11 |
* |
| 12 |
* This program is distributed in the hope that it will be useful, |
| 13 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 15 |
* GNU General Public License for more details. |
| 16 |
* |
| 17 |
* You should have received a copy of the GNU General Public License |
| 18 |
* along with BaitFisher. If not, see <http://www.gnu.org/licenses/>. |
| 19 |
* |
| 20 |
* |
| 21 |
* For any enquiries send an Email to Christoph Mayer |
| 22 |
* [email protected] |
| 23 |
* |
| 24 |
* When publishing work that is based on the results please cite: |
| 25 |
* Mayer et al. 2016: BaitFisher: A software package for multi-species target DNA enrichment probe design |
| 26 |
* |
| 27 |
*/ |
| 28 |
#ifndef CHISTOGRAM_H |
| 29 |
#define CHISTOGRAM_H |
| 30 |
|
| 31 |
// This is the CHistogram class written by Christoph Mayer, Forschungsmuseum Alexander Koenig, Bonn, Germany. |
| 32 |
// Discription: |
| 33 |
// This class allowes different binning schemes: |
| 34 |
// 1) User supplied number of bins together with lower and upper value. |
| 35 |
// IMPORTANT: The lower value is inclusive, the upper value is exclusive. |
| 36 |
// Therefore, a value equal to the upper value cannot be added |
| 37 |
// to the histrogram data. |
| 38 |
// If this is a problem, increase the upper value, or take |
| 39 |
// a different constructor. See below. |
| 40 |
// 2) Vector or range of data. From this the lower and upper bound |
| 41 |
// are determined automatically. Number fo bins or automatic binning |
| 42 |
// if the number of bins parameter has certain negative values. |
| 43 |
// 3) Discrete data. In this case the data is not binned when it is added, |
| 44 |
// but stored in a map. The binning is performed upon request. |
| 45 |
|
| 46 |
|
| 47 |
|
| 48 |
|
| 49 |
#include <iostream> |
| 50 |
#include <vector> |
| 51 |
#include <map> |
| 52 |
#include <cmath> |
| 53 |
#include <cctype> |
| 54 |
#include <iterator> |
| 55 |
#include <cstdlib> |
| 56 |
#include <algorithm> |
| 57 |
#include "statistic_functions.h" |
| 58 |
|
| 59 |
#define EPSS 0.00000000001; |
| 60 |
|
| 61 |
|
| 62 |
|
| 63 |
|
| 64 |
|
| 65 |
|
| 66 |
|
| 67 |
|
| 68 |
inline bool greater_than_pair(const std::pair<double, unsigned> &a, const std::pair<double, unsigned> &b ) |
| 69 |
{ |
| 70 |
return a.second > b.second; |
| 71 |
} |
| 72 |
|
| 73 |
|
| 74 |
|
| 75 |
|
| 76 |
inline void add_or_count(std::map<double, unsigned> &m, double &x) |
| 77 |
{ |
| 78 |
std::map<double,unsigned>::iterator it; |
| 79 |
|
| 80 |
it = m.find(x); |
| 81 |
if (it == m.end() ) |
| 82 |
{ |
| 83 |
m[x] = 1; |
| 84 |
} |
| 85 |
else |
| 86 |
{ |
| 87 |
++it->second; |
| 88 |
} |
| 89 |
} |
| 90 |
|
| 91 |
|
| 92 |
|
| 93 |
|
| 94 |
class CHistogram |
| 95 |
|
| 96 |
{ |
| 97 |
private: |
| 98 |
double a; // Inclusive range minimum |
| 99 |
double b; // Exclusive range maximum |
| 100 |
unsigned bins; |
| 101 |
double step; |
| 102 |
unsigned entries; |
| 103 |
|
| 104 |
std::vector<unsigned> data; |
| 105 |
|
| 106 |
bool discrete_data; // is determined in initialisation. Switched on with -10. |
| 107 |
std::map<double, unsigned> discrete_hist_data; |
| 108 |
|
| 109 |
private: |
| 110 |
bool add_intern_non_discrete(double x) |
| 111 |
{ |
| 112 |
// Be careful: If x==b this is allowed, but it places |
| 113 |
// this hit into the |
| 114 |
|
| 115 |
if (x < a || x >= b) |
| 116 |
{ |
| 117 |
if (x == b) |
| 118 |
std::cerr << "Warning: CHistrogram::add_intern_non_discrete data.siz\n" |
| 119 |
<< " Data points that are added must be exclusive the upper bound of the given range.\n\n"; |
| 120 |
return false; |
| 121 |
} |
| 122 |
|
| 123 |
++entries; |
| 124 |
|
| 125 |
unsigned bin = (x-a)/step; |
| 126 |
// std::cerr << "HH " << bin << " " << data.size() << std::endl; |
| 127 |
|
| 128 |
++data[bin]; |
| 129 |
return true; |
| 130 |
} |
| 131 |
|
| 132 |
// Adds the value to the discete data map. |
| 133 |
// Number of bins as well as a and b are not changed. |
| 134 |
bool add_intern_discrete_no_update(double x) |
| 135 |
{ |
| 136 |
++entries; |
| 137 |
add_or_count(discrete_hist_data, x); |
| 138 |
return true; |
| 139 |
} |
| 140 |
|
| 141 |
void update_a_b_bins_discrete_data() |
| 142 |
{ |
| 143 |
// Remember: The discrete data is stored in an automatically sorted map. |
| 144 |
// Therefore, the bounds are determined from the first and last element. |
| 145 |
|
| 146 |
// Determine a, b and bins: |
| 147 |
a = discrete_hist_data.begin()->first; |
| 148 |
b = discrete_hist_data.rbegin()->first; |
| 149 |
|
| 150 |
bins = discrete_hist_data.size(); |
| 151 |
} |
| 152 |
|
| 153 |
|
| 154 |
public: |
| 155 |
void init_with_vector(const std::vector<double> &vals, int bins_param=0) |
| 156 |
{ |
| 157 |
init_with_range(vals.begin(), vals.end(), vals.size(), bins_param); |
| 158 |
} |
| 159 |
|
| 160 |
|
| 161 |
template <typename T> |
| 162 |
void init_with_range(T it_beg, // pointer or iterator to first element |
| 163 |
T it_end, // pointer or iterator to element behind last element |
| 164 |
unsigned num_range_elements, // number of elements in range |
| 165 |
int bins_param=0) // number of bins or if <=0 specifies the method |
| 166 |
// to determine this number |
| 167 |
{ |
| 168 |
double mi, ma, me, sd; |
| 169 |
|
| 170 |
discrete_data = false; |
| 171 |
|
| 172 |
if (num_range_elements == 0) // Fill in some dummy values so that everything is well defined. Pratically this does not make sense, but should prevent other classes using this data from crashing. |
| 173 |
{ |
| 174 |
a = 0; |
| 175 |
b = 1; |
| 176 |
data.push_back(0); |
| 177 |
bins = 1; |
| 178 |
step = 1; |
| 179 |
|
| 180 |
return; |
| 181 |
} |
| 182 |
|
| 183 |
mi = ma = 0; // This code silences the used uninitialized warning of some compilers. It is not necessary, since the two variables are initialized in |
| 184 |
// range_min_max as the name suggests. |
| 185 |
|
| 186 |
range_min_max(it_beg, it_end, mi, ma); |
| 187 |
|
| 188 |
// Special case: Only one category and automatic number of bins, but not discrete data |
| 189 |
// In principle this is a special case of discrete data but not explicitly mentioned. |
| 190 |
if (bins_param != -10 && bins_param <= 0 && mi== ma) |
| 191 |
bins_param = 1; |
| 192 |
|
| 193 |
|
| 194 |
|
| 195 |
// Formulas for auto bins number: http://en.wikipedia.org/wiki/Histogram |
| 196 |
|
| 197 |
if (bins_param == 0) // Sturges formula -- can perform poorly if n < 30. |
| 198 |
{ |
| 199 |
bins = ceil(log(num_range_elements)/log(2)+1); |
| 200 |
} |
| 201 |
else if (bins_param == -1) // Scotts formula |
| 202 |
{ |
| 203 |
range_mean_sd(it_beg, it_end, me, sd); |
| 204 |
bins = 3.5*sd/pow(num_range_elements, 1./3.); |
| 205 |
if (bins < 1) |
| 206 |
bins = 1; |
| 207 |
} |
| 208 |
else if (bins_param == -2) // sqrt formula |
| 209 |
{ |
| 210 |
bins = sqrt(num_range_elements); |
| 211 |
} |
| 212 |
else if (bins_param == -3) // Freedman-Diaconis' formula // Not implemented |
| 213 |
{ |
| 214 |
bins = 1; |
| 215 |
} |
| 216 |
else if (bins_param == -4) // minimization of risk function // Not implemented |
| 217 |
{ |
| 218 |
bins = 1; |
| 219 |
} |
| 220 |
else if (bins_param == -10) // Each value gets its own bin |
| 221 |
{ |
| 222 |
discrete_data = true; |
| 223 |
} |
| 224 |
else if (bins_param > 0) |
| 225 |
{ |
| 226 |
bins = bins_param; |
| 227 |
} |
| 228 |
else |
| 229 |
{ |
| 230 |
std::cerr << "Error: the init_param parameter is negative, but its value does not specify a method. This needs to be corrected." << std::endl; |
| 231 |
exit(-133); |
| 232 |
} |
| 233 |
|
| 234 |
// We have minimum (mi), maximum (ma) and number of bins. |
| 235 |
// Now we determine the step size as well as lower and upper bound (a,b). |
| 236 |
|
| 237 |
if (!discrete_data) // Add non discrete values: |
| 238 |
{ |
| 239 |
if (bins > 1) |
| 240 |
{ |
| 241 |
step = (ma-mi)/(bins-1.0); |
| 242 |
a = mi - 0.5*step; |
| 243 |
b = ma + 0.5*step; |
| 244 |
} |
| 245 |
else |
| 246 |
{ |
| 247 |
step = 1.0; // fantasy value |
| 248 |
a = mi - 0.5*step; |
| 249 |
b = ma + 0.5*step; |
| 250 |
} |
| 251 |
|
| 252 |
/* std::cerr << "--a " << a << std::endl; */ |
| 253 |
/* std::cerr << "--b " << b << std::endl; */ |
| 254 |
/* std::cerr << "--bins " << bins << std::endl; */ |
| 255 |
/* std::cerr << "--step " << step << std::endl; */ |
| 256 |
|
| 257 |
|
| 258 |
data = std::vector<unsigned>(bins, 0); |
| 259 |
|
| 260 |
while (it_beg != it_end) |
| 261 |
{ |
| 262 |
if (!add_intern_non_discrete(*it_beg) ) |
| 263 |
{ |
| 264 |
std::cerr << "Internal error: Value out of range when building the histgram.\n" << std::endl; |
| 265 |
exit(-144); |
| 266 |
} |
| 267 |
++it_beg; |
| 268 |
} |
| 269 |
} |
| 270 |
else // Add dicrete values |
| 271 |
{ |
| 272 |
step = 0; |
| 273 |
// Add all values to map: |
| 274 |
while (it_beg != it_end) |
| 275 |
{ |
| 276 |
add_intern_discrete_no_update(*it_beg); |
| 277 |
++it_beg; |
| 278 |
} |
| 279 |
update_a_b_bins_discrete_data(); |
| 280 |
} |
| 281 |
|
| 282 |
|
| 283 |
} |
| 284 |
|
| 285 |
// Constructor with: lower and upper value as well as the number of |
| 286 |
// equidistant bins. |
| 287 |
// Here we do not allow to determine the number of bins automatically for the simple reason: |
| 288 |
// The number of necessary bins can only be determined if the number of data points is known |
| 289 |
// which is not assumed here. |
| 290 |
CHistogram(double a_param, double b_param, int bins_param): |
| 291 |
a(a_param), b(b_param), bins(bins_param), entries(0), data(bins_param, 0) |
| 292 |
{ |
| 293 |
if (bins_param < 1) |
| 294 |
{ |
| 295 |
std::cerr << "Critical error in constructor of CHistogram. Parameter bins_param must be positive here.\n" << std::endl; |
| 296 |
exit(-143); |
| 297 |
} |
| 298 |
|
| 299 |
discrete_data = false; |
| 300 |
step = (b-a)/bins; |
| 301 |
|
| 302 |
// Adds no data. |
| 303 |
} |
| 304 |
|
| 305 |
// Constructor with: range or vector as well as the bins_param |
| 306 |
// if bins_param < 1: then the parameter specifies the method to determine the number |
| 307 |
// of bins automatically. |
| 308 |
// 0: Sturges formula -- can perform poorly if n < 30. |
| 309 |
// -1: Scotts formula |
| 310 |
// -2: sqrt formula |
| 311 |
// -3: Freedman-Diaconis' formula // Not implemented |
| 312 |
// -4: minimization of risk function // Not implemented |
| 313 |
// if bins_param > 1: this specifies the actual number of bins |
| 314 |
|
| 315 |
CHistogram(const std::vector<double> &vals, int bins_param=0):entries(0) |
| 316 |
{ |
| 317 |
init_with_range(vals.begin(), vals.end(), vals.size(), bins_param); |
| 318 |
} |
| 319 |
|
| 320 |
CHistogram(const std::vector<float> &vals, int bins_param=0):entries(0) |
| 321 |
{ |
| 322 |
init_with_range(vals.begin(), vals.end(), vals.size(), bins_param); |
| 323 |
} |
| 324 |
|
| 325 |
CHistogram(const std::vector<unsigned> &vals, int bins_param=0):entries(0) |
| 326 |
{ |
| 327 |
init_with_range(vals.begin(), vals.end(), vals.size(), bins_param); |
| 328 |
} |
| 329 |
|
| 330 |
template <typename T> |
| 331 |
CHistogram(T it_beg, |
| 332 |
T it_end, |
| 333 |
int bins_param=0):entries(0) |
| 334 |
{ |
| 335 |
init_with_range(it_beg, it_end, std::distance(it_beg, it_end), bins_param); |
| 336 |
} |
| 337 |
|
| 338 |
double get_lower() |
| 339 |
{ |
| 340 |
return a; |
| 341 |
} |
| 342 |
|
| 343 |
double get_upper() |
| 344 |
{ |
| 345 |
return b; |
| 346 |
} |
| 347 |
|
| 348 |
unsigned get_bins() |
| 349 |
{ |
| 350 |
return bins; |
| 351 |
} |
| 352 |
|
| 353 |
double get_binsize() |
| 354 |
{ |
| 355 |
return step; |
| 356 |
} |
| 357 |
|
| 358 |
// This function should be reimplemented in a cleaner way!!! |
| 359 |
// The use of the data field to store disc |
| 360 |
const std::vector<unsigned> &get_histogram_data() |
| 361 |
{ |
| 362 |
if (discrete_data) // data needs to be prepared: |
| 363 |
{ |
| 364 |
data.clear(); |
| 365 |
|
| 366 |
std::map<double, unsigned>::iterator it_beg, it_end; |
| 367 |
|
| 368 |
it_beg = discrete_hist_data.begin(); |
| 369 |
it_end = discrete_hist_data.end(); |
| 370 |
|
| 371 |
while (it_beg != it_end) |
| 372 |
{ |
| 373 |
data.push_back(it_beg->second); |
| 374 |
++it_beg; |
| 375 |
} |
| 376 |
} |
| 377 |
return data; |
| 378 |
} |
| 379 |
|
| 380 |
void get_bin_coords(std::vector<double> &v) |
| 381 |
{ |
| 382 |
/* std::cerr << "a " << a << std::endl; */ |
| 383 |
/* std::cerr << "b " << b << std::endl; */ |
| 384 |
/* std::cerr << "bins " << bins << std::endl; */ |
| 385 |
/* std::cerr << "step " << step << std::endl; */ |
| 386 |
|
| 387 |
|
| 388 |
unsigned i; |
| 389 |
|
| 390 |
v.clear(); |
| 391 |
|
| 392 |
if (!discrete_data) |
| 393 |
{ |
| 394 |
for (i=0; i<bins; ++i) |
| 395 |
{ |
| 396 |
v.push_back(a+(i+0.5)*step); |
| 397 |
} |
| 398 |
} |
| 399 |
else |
| 400 |
{ |
| 401 |
std::map<double, unsigned>::iterator it_beg, it_end; |
| 402 |
|
| 403 |
it_beg = discrete_hist_data.begin(); |
| 404 |
it_end = discrete_hist_data.end(); |
| 405 |
|
| 406 |
while (it_beg != it_end) |
| 407 |
{ |
| 408 |
v.push_back(it_beg->first); |
| 409 |
++it_beg; |
| 410 |
} |
| 411 |
} |
| 412 |
} |
| 413 |
|
| 414 |
|
| 415 |
//XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
| 416 |
// old code - needs revision. See method to compute coordinates |
| 417 |
/* void print_bin_coord_hist_data(std::ostream &os, bool normal=true) */ |
| 418 |
/* { */ |
| 419 |
|
| 420 |
/* unsigned i; */ |
| 421 |
/* double tmp; */ |
| 422 |
|
| 423 |
/* // Not very elegant to change this for os - should be changed */ |
| 424 |
|
| 425 |
/* os.setf(std::ios::fixed); */ |
| 426 |
/* os.precision(6); */ |
| 427 |
|
| 428 |
/* if (!discrete_data) */ |
| 429 |
/* { */ |
| 430 |
/* for (i=0; i<bins; ++i) */ |
| 431 |
/* { */ |
| 432 |
/* if (normal) */ |
| 433 |
/* tmp = (double) data[i]/entries/step; */ |
| 434 |
/* else */ |
| 435 |
/* tmp = (double) data[i]; */ |
| 436 |
/* os << (a+(i+0.5)*step) << "\t"; */ |
| 437 |
/* os << tmp << std::endl; */ |
| 438 |
/* } */ |
| 439 |
/* } */ |
| 440 |
/* else */ |
| 441 |
/* { */ |
| 442 |
/* std::map<double, unsigned>::iterator it_beg, it_end; */ |
| 443 |
|
| 444 |
/* it_beg = discrete_hist_data.begin(); */ |
| 445 |
/* it_end = discrete_hist_data.end(); */ |
| 446 |
|
| 447 |
/* while (it_beg != it_end) */ |
| 448 |
/* { */ |
| 449 |
/* if (normal) */ |
| 450 |
/* tmp = (double) it_beg->second/entries/step; */ |
| 451 |
/* else */ |
| 452 |
/* tmp = (double) it_beg->second; */ |
| 453 |
/* os << it_beg->first << "\t" << it_beg->second << std::endl; */ |
| 454 |
|
| 455 |
/* ++it_beg; */ |
| 456 |
/* } */ |
| 457 |
/* } */ |
| 458 |
/* } */ |
| 459 |
|
| 460 |
|
| 461 |
// Returns false if out of range error, else true. |
| 462 |
bool add(double x) |
| 463 |
{ |
| 464 |
if (discrete_data) |
| 465 |
{ |
| 466 |
add_intern_discrete_no_update(x); |
| 467 |
update_a_b_bins_discrete_data(); |
| 468 |
return true; // Can never fail |
| 469 |
} |
| 470 |
else |
| 471 |
{ |
| 472 |
return add_intern_non_discrete(x); |
| 473 |
} |
| 474 |
} |
| 475 |
|
| 476 |
bool add(std::vector<double> v) |
| 477 |
{ |
| 478 |
int i, n=v.size(); |
| 479 |
|
| 480 |
if (!discrete_data) // Non-discrete data: |
| 481 |
{ |
| 482 |
for (i=0; i<n; ++i) |
| 483 |
if (!add_intern_non_discrete(v[i])) |
| 484 |
return false; |
| 485 |
} |
| 486 |
else // discrete data: |
| 487 |
{ |
| 488 |
for (i=0; i<n; ++i) |
| 489 |
{ |
| 490 |
add_intern_discrete_no_update(v[i]); |
| 491 |
} |
| 492 |
update_a_b_bins_discrete_data(); |
| 493 |
} |
| 494 |
return true; |
| 495 |
} |
| 496 |
|
| 497 |
template<typename T> // T is some iterator or pointer type |
| 498 |
bool add(T b, T e) // defines range |
| 499 |
{ |
| 500 |
if (!discrete_data) // Discrete data: |
| 501 |
{ |
| 502 |
while (b != e) |
| 503 |
{ |
| 504 |
if (!add_intern_non_discrete(*b)) |
| 505 |
return false; |
| 506 |
++b; |
| 507 |
} |
| 508 |
} |
| 509 |
else |
| 510 |
{ |
| 511 |
while (b != e) |
| 512 |
{ |
| 513 |
add_intern_discrete_no_update(*b); |
| 514 |
++b; |
| 515 |
} |
| 516 |
update_a_b_bins_discrete_data(); |
| 517 |
} |
| 518 |
return true; |
| 519 |
} |
| 520 |
|
| 521 |
bool is_discrete_data() |
| 522 |
{ |
| 523 |
return discrete_data; |
| 524 |
} |
| 525 |
|
| 526 |
unsigned num_entries() |
| 527 |
{ |
| 528 |
return entries; |
| 529 |
} |
| 530 |
|
| 531 |
//XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX |
| 532 |
// old code - needs revision. |
| 533 |
/* bool get_dominant_category_and_num(double &cat, unsigned &num) */ |
| 534 |
/* { */ |
| 535 |
/* if (discrete_data) */ |
| 536 |
/* { */ |
| 537 |
/* std::map<double, unsigned>::iterator it_beg, it_end; */ |
| 538 |
|
| 539 |
/* it_beg = discrete_hist_data.begin(); */ |
| 540 |
/* it_end = discrete_hist_data.end(); */ |
| 541 |
|
| 542 |
/* if (it_beg == it_end) */ |
| 543 |
/* { */ |
| 544 |
/* return false; */ |
| 545 |
/* } */ |
| 546 |
|
| 547 |
/* cat = it_beg->first; */ |
| 548 |
/* num = it_beg->second; */ |
| 549 |
|
| 550 |
/* while (it_beg != it_end) */ |
| 551 |
/* { */ |
| 552 |
/* if (it_beg->second > num) */ |
| 553 |
/* { */ |
| 554 |
/* cat = it_beg->first; */ |
| 555 |
/* num = it_beg->second; */ |
| 556 |
/* } */ |
| 557 |
/* ++it_beg; */ |
| 558 |
/* } */ |
| 559 |
/* } */ |
| 560 |
/* else // Non discrete data: */ |
| 561 |
/* { */ |
| 562 |
/* double max_key; */ |
| 563 |
/* unsigned max_value; */ |
| 564 |
|
| 565 |
/* unsigned i, n; */ |
| 566 |
|
| 567 |
/* n = data.size(); */ |
| 568 |
/* i = 0; */ |
| 569 |
|
| 570 |
/* if (n == 0) */ |
| 571 |
/* { */ |
| 572 |
/* return false; */ |
| 573 |
/* } */ |
| 574 |
|
| 575 |
/* max_key = a + step/2; */ |
| 576 |
/* max_value = data[0]; */ |
| 577 |
|
| 578 |
/* while (i != n) */ |
| 579 |
/* { */ |
| 580 |
/* if (data[i] > num) */ |
| 581 |
/* { */ |
| 582 |
/* max_key = a + (i + 0.5)*step; */ |
| 583 |
/* max_value = data[i]; */ |
| 584 |
/* } */ |
| 585 |
/* ++i; */ |
| 586 |
/* } */ |
| 587 |
/* } */ |
| 588 |
/* return true; */ |
| 589 |
/* } */ |
| 590 |
|
| 591 |
/* // Should be tested before it is used. */ |
| 592 |
/* void dominant_category(double &value, double &prop) */ |
| 593 |
/* { */ |
| 594 |
/* double cat; */ |
| 595 |
/* unsigned num; */ |
| 596 |
|
| 597 |
/* if (get_dominant_category_and_num(cat, num)) */ |
| 598 |
/* { */ |
| 599 |
/* value = cat; */ |
| 600 |
/* prop = (double)num/entries; */ |
| 601 |
/* } */ |
| 602 |
/* else */ |
| 603 |
/* { */ |
| 604 |
/* value = 0; */ |
| 605 |
/* prop = -1; */ |
| 606 |
/* } */ |
| 607 |
/* } */ |
| 608 |
|
| 609 |
|
| 610 |
|
| 611 |
void categories_sorted_by_dominance(std::vector<std::pair<double, unsigned> > &cat_val_pair_vec) |
| 612 |
{ |
| 613 |
if (discrete_data) |
| 614 |
{ |
| 615 |
std::map<double, unsigned>::iterator it_beg, it_end; |
| 616 |
|
| 617 |
it_beg = discrete_hist_data.begin(); |
| 618 |
it_end = discrete_hist_data.end(); |
| 619 |
|
| 620 |
if (it_beg == it_end) |
| 621 |
{ |
| 622 |
return; |
| 623 |
} |
| 624 |
|
| 625 |
while (it_beg != it_end) |
| 626 |
{ |
| 627 |
double cat; |
| 628 |
unsigned num; |
| 629 |
|
| 630 |
cat = it_beg->first; |
| 631 |
num = it_beg->second; |
| 632 |
|
| 633 |
cat_val_pair_vec.push_back(std::make_pair(cat, num)); |
| 634 |
|
| 635 |
++it_beg; |
| 636 |
} |
| 637 |
} |
| 638 |
else // Non discrete data: |
| 639 |
{ |
| 640 |
double key; |
| 641 |
unsigned value; |
| 642 |
|
| 643 |
unsigned i, n; |
| 644 |
|
| 645 |
n = data.size(); |
| 646 |
i = 0; |
| 647 |
|
| 648 |
if (n == 0) |
| 649 |
{ |
| 650 |
return; |
| 651 |
} |
| 652 |
|
| 653 |
while (i != n) |
| 654 |
{ |
| 655 |
key = a + step/2; |
| 656 |
value = data[0]; |
| 657 |
|
| 658 |
cat_val_pair_vec.push_back(std::make_pair(key, value)); |
| 659 |
++i; |
| 660 |
} |
| 661 |
} |
| 662 |
|
| 663 |
sort(cat_val_pair_vec.begin(), cat_val_pair_vec.end(), greater_than_pair); |
| 664 |
} |
| 665 |
|
| 666 |
}; |
| 667 |
|
| 668 |
#endif |