1 // This file is part of the Mussa source distribution.
2 // http://mussa.caltech.edu/
3 // Contact author: Tristan De Buysscher, tristan@caltech.edu
5 // This program and all associated source code files are Copyright (C) 2005
6 // the California Institute of Technology, Pasadena, CA, 91125 USA. It is
7 // under the GNU Public License; please see the included LICENSE.txt
8 // file for more information, or contact Tristan directly.
11 // ----------------------------------------
12 // ---------- mussa_class.cc -----------
13 // ----------------------------------------
19 #include "mussa_exceptions.hpp"
20 #include "alg/mussa.hpp"
21 #include "alg/flp.hpp"
30 // set all parameters to null state
35 ana_mode = TransitiveNway;
43 // these 5 simple methods manually set the parameters for doing an analysis
44 // used so that the gui can take input from user and setup the analysis
45 // note - still need a set_append(bool, bool) method...
47 Mussa::set_name(string a_name)
49 analysis_name = a_name;
52 string Mussa::get_name()
60 if (the_seqs.size() > 0)
61 return the_seqs.size();
67 Mussa::set_window(int a_window)
72 int Mussa::get_window() const
78 Mussa::set_threshold(int a_threshold)
80 threshold = a_threshold;
81 //soft_thres = a_threshold;
84 int Mussa::get_threshold() const
90 Mussa::set_soft_thres(int sft_thres)
92 soft_thres = sft_thres;
96 Mussa::set_analysis_mode(enum analysis_modes new_ana_mode)
98 ana_mode = new_ana_mode;
101 enum Mussa::analysis_modes Mussa::get_analysis_mode() const
106 string Mussa::get_analysis_mode_name() const
111 return string("Transitive");
114 return string("Radial");
117 return string("Entropy");
120 return string("[deprecated] Recursive");
123 throw runtime_error("invalid analysis mode type");
128 const NwayPaths& Mussa::paths() const
133 // takes a string and sets it as the next seq
135 Mussa::add_a_seq(string a_seq)
140 the_seqs.push_back(aSeq);
143 const vector<Sequence>&
144 Mussa::sequences() const
149 void Mussa::load_sequence(string seq_file, string annot_file, int fasta_index,
150 int sub_seq_start, int sub_seq_end)
153 aseq.load_fasta(seq_file, fasta_index, sub_seq_start, sub_seq_end);
154 if (annot_file.size() > 0) {
155 aseq.load_annot(annot_file, sub_seq_start, sub_seq_end);
157 the_seqs.push_back(aseq);
161 Mussa::load_mupa_file(string para_file_path)
163 string file_path_base;
165 string file_data_line;
166 string param, value, annot_file;
167 int split_index, fasta_index;
168 int sub_seq_start, sub_seq_end;
169 bool seq_params, did_seq;
172 string::size_type new_index, dir_index;
177 para_file.open(para_file_path.c_str(), ios::in);
179 // if file was opened, read the parameter values
182 // need to find the path to the .mupa file
187 new_index = (para_file_path.substr(dir_index)).find("/");
188 if (new_index != string::npos)
189 dir_index += new_index + 1;
191 parsing_path = false;
194 file_path_base = para_file_path.substr(0,dir_index);
196 // setup loop by getting file's first line
197 getline(para_file,file_data_line);
198 split_index = file_data_line.find(" ");
199 param = file_data_line.substr(0,split_index);
200 value = file_data_line.substr(split_index+1);
202 while (!para_file.eof())
205 if (param == "ANA_NAME")
206 analysis_name = value;
207 else if (param == "APPEND_WIN")
209 else if (param == "APPEND_THRES")
211 else if (param == "SEQUENCE_NUM")
212 ; // ignore sequence_num now
213 else if (param == "WINDOW")
214 window = atoi(value.c_str());
215 else if (param == "THRESHOLD")
216 threshold = atoi(value.c_str());
217 else if (param == "SEQUENCE")
219 string seq_file = file_path_base + value;
220 //cout << "seq_file_name " << seq_files.back() << endl;
227 while ((!para_file.eof()) && seq_params)
229 getline(para_file,file_data_line);
230 split_index = file_data_line.find(" ");
231 param = file_data_line.substr(0,split_index);
232 value = file_data_line.substr(split_index+1);
234 if (param == "FASTA_INDEX")
235 fasta_index = atoi(value.c_str());
236 else if (param == "ANNOTATION")
237 annot_file = file_path_base + value;
238 else if (param == "SEQ_START")
239 sub_seq_start = atoi(value.c_str());
240 else if (param == "SEQ_END")
242 sub_seq_end = atoi(value.c_str());
244 //ignore empty lines or that start with '#'
245 else if ((param == "") || (param == "#")) {}
246 else seq_params = false;
248 load_sequence(seq_file, annot_file, fasta_index, sub_seq_start,
252 //ignore empty lines or that start with '#'
253 else if ((param == "") || (param == "#")) {}
256 clog << "Illegal/misplaced mussa parameter in file\n";
257 clog << param << "\n";
262 getline(para_file,file_data_line);
263 split_index = file_data_line.find(" ");
264 param = file_data_line.substr(0,split_index);
265 value = file_data_line.substr(split_index+1);
272 soft_thres = threshold;
273 //cout << "nway mupa: analysis_name = " << analysis_name
274 // << " window = " << window
275 // << " threshold = " << threshold << endl;
277 // no file was loaded, signal error
280 throw mussa_load_error("Config File: " + para_file_path + " not found");
286 Mussa::analyze(int w, int t, enum Mussa::analysis_modes the_ana_mode, double new_ent_thres)
288 time_t t1, t2, begin, end;
289 double seqloadtime, seqcomptime, nwaytime, savetime, totaltime;
293 ana_mode = the_ana_mode;
294 ent_thres = new_ent_thres;
305 if (the_seqs.size() < 2) {
306 throw mussa_analysis_error("you need to have at least 2 sequences to "
309 //cout << "nway ana: seq_num = " << the_seqs.size() << endl;
312 seqloadtime = difftime(t2, t1);
317 seqcomptime = difftime(t2, t1);
321 the_paths.setup(window, threshold);
324 nwaytime = difftime(t2, t1);
329 savetime = difftime(t2, t1);
332 totaltime = difftime(end, begin);
335 //cout << "seqload\tseqcomp\tnway\tsave\ttotal\n";
336 //cout << seqloadtime << "\t";
337 //cout << seqcomptime << "\t";
338 //cout << nwaytime << "\t";
339 //cout << savetime << "\t";
340 //cout << totaltime << "\n";
346 vector<int> seq_lens;
347 vector<FLPs> empty_FLP_vector;
349 string save_file_string;
351 empty_FLP_vector.clear();
352 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
354 all_comps.push_back(empty_FLP_vector);
355 for(vector<Sequence>::size_type i2 = 0; i2 < the_seqs.size(); i2++)
356 all_comps[i].push_back(dummy_comp);
358 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
359 seq_lens.push_back(the_seqs[i].size());
361 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
362 for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
364 //cout << "seqcomping: " << i << " v. " << i2 << endl;
365 all_comps[i][i2].setup(window, threshold);
366 all_comps[i][i2].seqcomp(the_seqs[i].get_seq(), the_seqs[i2].get_seq(), false);
367 all_comps[i][i2].seqcomp(the_seqs[i].get_seq(),the_seqs[i2].rev_comp(),true);
374 vector<string> some_Seqs;
376 the_paths.set_soft_thres(soft_thres);
378 if (ana_mode == TransitiveNway) {
379 the_paths.trans_path_search(all_comps);
381 else if (ana_mode == RadialNway) {
382 the_paths.radiate_path_search(all_comps);
384 else if (ana_mode == EntropyNway)
386 //unlike other methods, entropy needs to look at the sequence at this stage
388 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
390 some_Seqs.push_back(the_seqs[i].get_seq());
393 the_paths.setup_ent(ent_thres, some_Seqs); // ent analysis extra setup
394 the_paths.entropy_path_search(all_comps);
397 // old recursive transitive analysis function
398 else if (ana_mode == RecursiveNway)
399 the_paths.find_paths_r(all_comps);
401 the_paths.simple_refine();
407 string save_name, save_path, create_dir_cmd, flp_filepath;
409 ostringstream append_info;
410 int dir_create_status;
413 // not sure why, but gotta close file each time since can't pass file streams
415 save_name = analysis_name;
417 // gotta do bit with adding win & thres if to be appended
421 append_info << "_w" << window;
422 save_name += append_info.str();
428 append_info << "_t" << threshold;
429 save_name += append_info.str();
432 //#include <stdlib.h>
433 // ******* use appropriate for os ------- 1 of 4
434 // the additions for osX make it more sane where it saves the analysis
435 // will come up with a cleaner sol'n later...
436 create_dir_cmd = "mkdir " + save_name; //linux
437 //create_dir_cmd = "mkdir " + file_path_base + save_name; //osX
439 dir_create_status = system( (const char*) create_dir_cmd.c_str());
440 //cout << "action: " << dir_create_status << endl;
442 // save sequence and annots to a special mussa file
444 // ******** use appropriate for OS ---------- 2 of 4
445 save_path = save_name + "/" + save_name + ".museq"; //linux
446 //save_path = file_path_base + save_name + "/" + save_name + ".museq"; //osX
448 save_file.open(save_path.c_str(), ios::out);
449 save_file << "<Mussa_Sequence>" << endl;
452 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
454 the_seqs[i].save(save_file);
457 //save_file.open(save_path.c_str(), ios::app);
458 save_file << "</Mussa_Sequence>" << endl;
461 // save nway paths to its mussa save file
463 // ******** use appropriate for OS -------- 3 of 4
464 save_path = save_name + "/" + save_name + ".muway"; //linux
465 //save_path = file_path_base + save_name + "/" + save_name + ".muway"; //os X
466 the_paths.save(save_path);
468 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
469 for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
472 append_info << "_sp_" << i << "v" << i2;
473 // ******** use appropriate for OS --------- 4 of 4
475 save_path = save_name + "/" + save_name + append_info.str() + ".flp";
477 //save_path = file_path_base + save_name + "/" + save_name + append_info.str() + ".flp";
478 all_comps[i][i2].save(save_path);
483 Mussa::save_muway(string save_path)
485 the_paths.save(save_path);
489 Mussa::load(string ana_file)
492 string::size_type start_index, end_index;
493 string file_path_base, a_file_path, ana_path;
497 ostringstream append_info;
498 vector<FLPs> empty_FLP_vector;
501 //cout << "ana_file name " << ana_file << endl;
504 end_index = ana_path.size()-1;
505 if (ana_path[end_index] == '/') {
508 start_index = ana_path.rfind('/', end_index);
509 if (start_index == string::npos) {
513 // skip the / we found
516 analysis_name = ana_path.substr(start_index, end_index-start_index+1);
517 //cout << " ana_name " << analysis_name << endl;
518 file_path_base = ana_path.substr(0, start_index) + analysis_name
519 + "/" + analysis_name;
520 a_file_path = file_path_base + ".muway";
521 //cout << " loading museq: " << a_file_path << endl;
522 the_paths.load(a_file_path);
524 int seq_num = the_paths.sequence_count();
526 a_file_path = file_path_base + ".museq";
528 // this is a bit of a hack due to C++ not acting like it should with files
529 for (i = 1; i <= seq_num; i++)
532 //cout << "mussa_class: loading museq frag... " << a_file_path << endl;
533 tmp_seq.load_museq(a_file_path, i);
534 the_seqs.push_back(tmp_seq);
537 empty_FLP_vector.clear();
538 for(i = 0; i < seq_num; i++)
540 all_comps.push_back(empty_FLP_vector);
541 for(i2 = 0; i2 < seq_num; i2++)
542 all_comps[i].push_back(dummy_comp);
545 for(i = 0; i < seq_num; i++)
547 for(i2 = i+1; i2 < seq_num; i2++)
550 append_info << "_sp_" << i << "v" << i2;
551 //cout << append_info.str() << endl;
552 a_file_path = file_path_base + append_info.str() + ".flp";
553 all_comps[i][i2].load(a_file_path);
554 //cout << "real size = " << all_comps[i][i2].size() << endl;
565 save_file.open(analysis_name.c_str(), ios::out);
567 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
568 save_file << the_seqs[i].get_seq() << endl;
570 save_file << window << endl;
572 //note more complex eventually since analysis_name may need to have
573 //window size, threshold and other stuff to modify it...
574 the_paths.save_old(analysis_name);
579 Mussa::load_old(char * load_file_path, int s_num)
582 string file_data_line;
583 int i, space_split_i, comma_split_i;
584 vector<int> loaded_path;
585 string node_pair, node;
589 the_paths.setup(0, 0);
590 save_file.open(load_file_path, ios::in);
592 // currently loads old mussa format
595 for(i = 0; i < seq_num; i++)
597 getline(save_file, file_data_line);
598 a_seq.set_seq(file_data_line);
599 the_seqs.push_back(a_seq);
603 getline(save_file, file_data_line);
604 window = atoi(file_data_line.c_str());
607 while (!save_file.eof())
610 getline(save_file, file_data_line);
611 if (file_data_line != "")
612 for(i = 0; i < seq_num; i++)
614 space_split_i = file_data_line.find(" ");
615 node_pair = file_data_line.substr(0,space_split_i);
616 //cout << "np= " << node_pair;
617 comma_split_i = node_pair.find(",");
618 node = node_pair.substr(comma_split_i+1);
619 //cout << "n= " << node << " ";
620 loaded_path.push_back(atoi (node.c_str()));
621 file_data_line = file_data_line.substr(space_split_i+1);
624 // FIXME: do we have any information about what the threshold should be?
625 the_paths.add_path(0, loaded_path);
629 //the_paths.save("tmp.save");