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;
42 fasta_indices.clear();
44 sub_seq_starts.clear();
48 // these 5 simple methods manually set the parameters for doing an analysis
49 // used so that the gui can take input from user and setup the analysis
50 // note - still need a set_append(bool, bool) method...
52 Mussa::set_name(string a_name)
54 analysis_name = a_name;
57 string Mussa::get_name()
65 if (the_seqs.size() > 0)
66 return the_seqs.size();
67 else if (seq_files.size() > 0)
68 return seq_files.size();
74 Mussa::set_window(int a_window)
79 int Mussa::get_window() const
85 Mussa::set_threshold(int a_threshold)
87 threshold = a_threshold;
88 //soft_thres = a_threshold;
91 int Mussa::get_threshold() const
97 Mussa::set_soft_thres(int sft_thres)
99 soft_thres = sft_thres;
103 Mussa::set_analysis_mode(enum analysis_modes new_ana_mode)
105 ana_mode = new_ana_mode;
108 enum Mussa::analysis_modes Mussa::get_analysis_mode() const
113 string Mussa::get_analysis_mode_name() const
118 return string("Transitive");
121 return string("Radial");
124 return string("Entropy");
127 return string("[deprecated] Recursive");
130 throw runtime_error("invalid analysis mode type");
135 const NwayPaths& Mussa::paths() const
140 // takes a string and sets it as the next seq
142 Mussa::add_a_seq(string a_seq)
147 the_seqs.push_back(aSeq);
151 // sets info for just 1 seq at a time
153 Mussa::set_seq_info(string seq_file, string annot_file, int fa_i, int a_start, int the_end)
155 seq_files.push_back(seq_file);
156 fasta_indices.push_back(fa_i);
157 annot_files.push_back(annot_file);
158 sub_seq_starts.push_back(a_start);
159 sub_seq_ends.push_back(the_end);
162 const vector<Sequence>&
163 Mussa::sequences() const
169 Mussa::load_mupa_file(string para_file_path)
172 string file_data_line;
173 string param, value, annot_file;
174 int split_index, fasta_index;
175 int sub_seq_start, sub_seq_end;
176 bool seq_params, did_seq;
179 string::size_type new_index, dir_index;
184 para_file.open(para_file_path.c_str(), ios::in);
186 // if file was opened, read the parameter values
189 // need to find the path to the .mupa file
194 new_index = (para_file_path.substr(dir_index)).find("/");
195 if (new_index != string::npos)
196 dir_index += new_index + 1;
198 parsing_path = false;
201 file_path_base = para_file_path.substr(0,dir_index);
203 // setup loop by getting file's first line
204 getline(para_file,file_data_line);
205 split_index = file_data_line.find(" ");
206 param = file_data_line.substr(0,split_index);
207 value = file_data_line.substr(split_index+1);
209 while (!para_file.eof())
212 if (param == "ANA_NAME")
213 analysis_name = value;
214 else if (param == "APPEND_WIN")
216 else if (param == "APPEND_THRES")
218 else if (param == "SEQUENCE_NUM")
219 ; // ignore sequence_num now
220 else if (param == "WINDOW")
221 window = atoi(value.c_str());
222 else if (param == "THRESHOLD")
223 threshold = atoi(value.c_str());
224 else if (param == "SEQUENCE")
226 seq_files.push_back(file_path_base + value);
227 //cout << "seq_file_name " << seq_files.back() << endl;
234 while ((!para_file.eof()) && seq_params)
236 getline(para_file,file_data_line);
237 split_index = file_data_line.find(" ");
238 param = file_data_line.substr(0,split_index);
239 value = file_data_line.substr(split_index+1);
241 if (param == "FASTA_INDEX")
242 fasta_index = atoi(value.c_str());
243 else if (param == "ANNOTATION")
244 annot_file = file_path_base + value;
245 else if (param == "SEQ_START")
246 sub_seq_start = atoi(value.c_str());
247 else if (param == "SEQ_END")
249 sub_seq_end = atoi(value.c_str());
251 //ignore empty lines or that start with '#'
252 else if ((param == "") || (param == "#")) {}
253 else seq_params = false;
256 fasta_indices.push_back(fasta_index);
257 annot_files.push_back(annot_file);
258 sub_seq_starts.push_back(sub_seq_start);
259 sub_seq_ends.push_back(sub_seq_end);
262 //ignore empty lines or that start with '#'
263 else if ((param == "") || (param == "#")) {}
266 clog << "Illegal/misplaced mussa parameter in file\n";
267 clog << param << "\n";
272 getline(para_file,file_data_line);
273 split_index = file_data_line.find(" ");
274 param = file_data_line.substr(0,split_index);
275 value = file_data_line.substr(split_index+1);
282 soft_thres = threshold;
283 //cout << "nway mupa: analysis_name = " << analysis_name
284 // << " window = " << window
285 // << " threshold = " << threshold << endl;
287 // no file was loaded, signal error
290 throw mussa_load_error("Config File: " + para_file_path + " not found");
296 Mussa::analyze(int w, int t, enum Mussa::analysis_modes the_ana_mode, double new_ent_thres)
298 time_t t1, t2, begin, end;
299 double seqloadtime, seqcomptime, nwaytime, savetime, totaltime;
303 ana_mode = the_ana_mode;
304 ent_thres = new_ent_thres;
314 load_sequence_data();
316 if (the_seqs.size() < 2) {
317 throw mussa_analysis_error("you need to have at least 2 sequences to "
320 //cout << "nway ana: seq_num = " << the_seqs.size() << endl;
323 seqloadtime = difftime(t2, t1);
328 seqcomptime = difftime(t2, t1);
332 the_paths.setup(window, threshold);
335 nwaytime = difftime(t2, t1);
340 savetime = difftime(t2, t1);
343 totaltime = difftime(end, begin);
346 //cout << "seqload\tseqcomp\tnway\tsave\ttotal\n";
347 //cout << seqloadtime << "\t";
348 //cout << seqcomptime << "\t";
349 //cout << nwaytime << "\t";
350 //cout << savetime << "\t";
351 //cout << totaltime << "\n";
356 Mussa::load_sequence_data()
358 list<string>::iterator seq_files_i, annot_files_i;
359 list<int>::iterator fasta_indices_i, seq_starts_i, seq_ends_i;
364 seq_files_i = seq_files.begin();
365 fasta_indices_i = fasta_indices.begin();
366 annot_files_i = annot_files.begin();
367 seq_starts_i = sub_seq_starts.begin();
368 seq_ends_i = sub_seq_ends.begin();
370 while ( (seq_files_i != seq_files.end()) && (err_msg == "") )
371 /* it should be guarenteed that each of the following exist
372 should I bother checking, and how to deal with if not true...
374 (fasta_indices_i != fasta_indices.end()) &&
375 (annot_files_i != annot_files.end()) &&
376 (seq_starts_i != sub_seq_starts.end()) &&
377 (seq_ends_i != sub_seq_ends.end()) )
380 aSeq.load_fasta(*seq_files_i, *fasta_indices_i,*seq_starts_i, *seq_ends_i);
381 if (annot_files_i->size() > 0)
382 aSeq.load_annot(*annot_files_i, *seq_starts_i, *seq_ends_i);
384 the_seqs.push_back(aSeq);
385 //cout << aSeq.get_header() << endl;
386 //cout << aSeq.get_seq() << endl;
388 ++seq_files_i; // advance all the iterators
400 vector<int> seq_lens;
401 vector<FLPs> empty_FLP_vector;
403 string save_file_string;
405 empty_FLP_vector.clear();
406 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
408 all_comps.push_back(empty_FLP_vector);
409 for(vector<Sequence>::size_type i2 = 0; i2 < the_seqs.size(); i2++)
410 all_comps[i].push_back(dummy_comp);
412 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
413 seq_lens.push_back(the_seqs[i].size());
415 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
416 for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
418 //cout << "seqcomping: " << i << " v. " << i2 << endl;
419 all_comps[i][i2].setup(window, threshold);
420 all_comps[i][i2].seqcomp(the_seqs[i].get_seq(), the_seqs[i2].get_seq(), false);
421 all_comps[i][i2].seqcomp(the_seqs[i].get_seq(),the_seqs[i2].rev_comp(),true);
428 vector<string> some_Seqs;
430 the_paths.set_soft_thres(soft_thres);
432 if (ana_mode == TransitiveNway) {
433 the_paths.trans_path_search(all_comps);
435 else if (ana_mode == RadialNway) {
436 the_paths.radiate_path_search(all_comps);
438 else if (ana_mode == EntropyNway)
440 //unlike other methods, entropy needs to look at the sequence at this stage
442 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
444 some_Seqs.push_back(the_seqs[i].get_seq());
447 the_paths.setup_ent(ent_thres, some_Seqs); // ent analysis extra setup
448 the_paths.entropy_path_search(all_comps);
451 // old recursive transitive analysis function
452 else if (ana_mode == RecursiveNway)
453 the_paths.find_paths_r(all_comps);
455 the_paths.simple_refine();
461 string save_name, save_path, create_dir_cmd, flp_filepath;
463 ostringstream append_info;
464 int dir_create_status;
467 // not sure why, but gotta close file each time since can't pass file streams
469 save_name = analysis_name;
471 // gotta do bit with adding win & thres if to be appended
475 append_info << "_w" << window;
476 save_name += append_info.str();
482 append_info << "_t" << threshold;
483 save_name += append_info.str();
486 //#include <stdlib.h>
487 // ******* use appropriate for os ------- 1 of 4
488 // the additions for osX make it more sane where it saves the analysis
489 // will come up with a cleaner sol'n later...
490 create_dir_cmd = "mkdir " + save_name; //linux
491 //create_dir_cmd = "mkdir " + file_path_base + save_name; //osX
493 dir_create_status = system( (const char*) create_dir_cmd.c_str());
494 //cout << "action: " << dir_create_status << endl;
496 // save sequence and annots to a special mussa file
498 // ******** use appropriate for OS ---------- 2 of 4
499 save_path = save_name + "/" + save_name + ".museq"; //linux
500 //save_path = file_path_base + save_name + "/" + save_name + ".museq"; //osX
502 save_file.open(save_path.c_str(), ios::out);
503 save_file << "<Mussa_Sequence>" << endl;
506 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
508 the_seqs[i].save(save_file);
511 //save_file.open(save_path.c_str(), ios::app);
512 save_file << "</Mussa_Sequence>" << endl;
515 // save nway paths to its mussa save file
517 // ******** use appropriate for OS -------- 3 of 4
518 save_path = save_name + "/" + save_name + ".muway"; //linux
519 //save_path = file_path_base + save_name + "/" + save_name + ".muway"; //os X
520 the_paths.save(save_path);
522 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
523 for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
526 append_info << "_sp_" << i << "v" << i2;
527 // ******** use appropriate for OS --------- 4 of 4
529 save_path = save_name + "/" + save_name + append_info.str() + ".flp";
531 //save_path = file_path_base + save_name + "/" + save_name + append_info.str() + ".flp";
532 all_comps[i][i2].save(save_path);
537 Mussa::save_muway(string save_path)
539 the_paths.save(save_path);
543 Mussa::load(string ana_file)
546 string::size_type start_index, end_index;
547 string file_path_base, a_file_path, ana_path;
551 ostringstream append_info;
552 vector<FLPs> empty_FLP_vector;
555 //cout << "ana_file name " << ana_file << endl;
558 end_index = ana_path.size()-1;
559 if (ana_path[end_index] == '/') {
562 start_index = ana_path.rfind('/', end_index);
563 if (start_index == string::npos) {
567 // skip the / we found
570 analysis_name = ana_path.substr(start_index, end_index-start_index+1);
571 //cout << " ana_name " << analysis_name << endl;
572 file_path_base = ana_path.substr(0, start_index) + analysis_name
573 + "/" + analysis_name;
574 a_file_path = file_path_base + ".muway";
575 //cout << " loading museq: " << a_file_path << endl;
576 the_paths.load(a_file_path);
578 int seq_num = the_paths.sequence_count();
580 a_file_path = file_path_base + ".museq";
582 // this is a bit of a hack due to C++ not acting like it should with files
583 for (i = 1; i <= seq_num; i++)
586 //cout << "mussa_class: loading museq frag... " << a_file_path << endl;
587 tmp_seq.load_museq(a_file_path, i);
588 the_seqs.push_back(tmp_seq);
591 empty_FLP_vector.clear();
592 for(i = 0; i < seq_num; i++)
594 all_comps.push_back(empty_FLP_vector);
595 for(i2 = 0; i2 < seq_num; i2++)
596 all_comps[i].push_back(dummy_comp);
599 for(i = 0; i < seq_num; i++)
601 for(i2 = i+1; i2 < seq_num; i2++)
604 append_info << "_sp_" << i << "v" << i2;
605 //cout << append_info.str() << endl;
606 a_file_path = file_path_base + append_info.str() + ".flp";
607 all_comps[i][i2].load(a_file_path);
608 //cout << "real size = " << all_comps[i][i2].size() << endl;
619 save_file.open(analysis_name.c_str(), ios::out);
621 for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
622 save_file << the_seqs[i].get_seq() << endl;
624 save_file << window << endl;
626 //note more complex eventually since analysis_name may need to have
627 //window size, threshold and other stuff to modify it...
628 the_paths.save_old(analysis_name);
633 Mussa::load_old(char * load_file_path, int s_num)
636 string file_data_line;
637 int i, space_split_i, comma_split_i;
638 vector<int> loaded_path;
639 string node_pair, node;
643 the_paths.setup(0, 0);
644 save_file.open(load_file_path, ios::in);
646 // currently loads old mussa format
649 for(i = 0; i < seq_num; i++)
651 getline(save_file, file_data_line);
652 a_seq.set_seq(file_data_line);
653 the_seqs.push_back(a_seq);
657 getline(save_file, file_data_line);
658 window = atoi(file_data_line.c_str());
661 while (!save_file.eof())
664 getline(save_file, file_data_line);
665 if (file_data_line != "")
666 for(i = 0; i < seq_num; i++)
668 space_split_i = file_data_line.find(" ");
669 node_pair = file_data_line.substr(0,space_split_i);
670 //cout << "np= " << node_pair;
671 comma_split_i = node_pair.find(",");
672 node = node_pair.substr(comma_split_i+1);
673 //cout << "n= " << node << " ";
674 loaded_path.push_back(atoi (node.c_str()));
675 file_data_line = file_data_line.substr(space_split_i+1);
678 // FIXME: do we have any information about what the threshold should be?
679 the_paths.add_path(0, loaded_path);
683 //the_paths.save("tmp.save");