5 * Created by Pat Schloss on 6/17/09.
6 * Copyright 2009 Patrick D. Schloss. All rights reserved.
10 #include "phylotree.h"
12 /**************************************************************************************************/
14 PhyloTree::PhyloTree(){
16 m = MothurOut::getInstance();
19 tree.push_back(TaxNode("Root"));
20 tree[0].heirarchyID = "0";
23 addSeqToTree("unknown", "unknown;");
26 m->errorOut(e, "PhyloTree", "PhyloTree");
30 /**************************************************************************************************/
32 PhyloTree::PhyloTree(ifstream& in, string filename){
34 m = MothurOut::getInstance();
44 char inFileName[1024];
45 strcpy(inFileName, filename.c_str());
47 MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);
48 MPI_File_get_size(inMPI, &size);
50 char* buffer = new char[size];
51 MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status);
53 string tempBuf = buffer;
54 if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size); }
55 istringstream iss (tempBuf,istringstream::in);
59 m->getline(iss); m->gobble(iss);
61 iss >> numNodes; m->gobble(iss);
63 tree.resize(numNodes);
65 for (int i = 0; i < tree.size(); i++) {
66 iss >> tree[i].name >> tree[i].level >> tree[i].parent; m->gobble(iss);
71 iss >> numGenus; m->gobble(iss);
75 for (int i = 0; i < numGenus; i++) {
76 iss >> gnode >> gsize; m->gobble(iss);
78 uniqueTaxonomies[gnode] = gnode;
79 totals.push_back(gsize);
82 MPI_File_close(&inMPI);
86 string line = m->getline(in); m->gobble(in);
88 in >> numNodes; m->gobble(in);
90 tree.resize(numNodes);
92 for (int i = 0; i < tree.size(); i++) {
93 in >> tree[i].name >> tree[i].level >> tree[i].parent; m->gobble(in);
98 in >> numGenus; m->gobble(in);
102 for (int i = 0; i < numGenus; i++) {
103 in >> gnode >> gsize; m->gobble(in);
105 uniqueTaxonomies[gnode] = gnode;
106 totals.push_back(gsize);
114 catch(exception& e) {
115 m->errorOut(e, "PhyloTree", "PhyloTree");
119 /**************************************************************************************************/
121 PhyloTree::PhyloTree(string tfile){
123 m = MothurOut::getInstance();
126 tree.push_back(TaxNode("Root"));
127 tree[0].heirarchyID = "0";
133 int pid, num, processors;
134 vector<unsigned long long> positions;
138 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
139 MPI_Comm_size(MPI_COMM_WORLD, &processors);
141 char inFileName[1024];
142 strcpy(inFileName, tfile.c_str());
144 MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
147 positions = m->setFilePosEachLine(tfile, num);
149 //send file positions to all processes
150 for(int i = 1; i < processors; i++) {
151 MPI_Send(&num, 1, MPI_INT, i, 2001, MPI_COMM_WORLD);
152 MPI_Send(&positions[0], (num+1), MPI_LONG, i, 2001, MPI_COMM_WORLD);
155 MPI_Recv(&num, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
156 positions.resize(num+1);
157 MPI_Recv(&positions[0], (num+1), MPI_LONG, 0, 2001, MPI_COMM_WORLD, &status);
161 for(int i=0;i<num;i++){
163 int length = positions[i+1] - positions[i];
164 char* buf4 = new char[length];
166 MPI_File_read_at(inMPI, positions[i], buf4, length, MPI_CHAR, &status);
168 string tempBuf = buf4;
169 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
172 istringstream iss (tempBuf,istringstream::in);
174 addSeqToTree(name, tax);
177 MPI_File_close(&inMPI);
178 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
182 m->openInputFile(tfile, in);
184 //read in users taxonomy file and add sequences to tree
186 in >> name >> tax; m->gobble(in);
188 addSeqToTree(name, tax);
193 assignHeirarchyIDs(0);
196 string unknownTax = "unknown;";
197 //added last taxon until you get desired level
198 for (int i = 1; i < maxLevel; i++) {
199 unknownTax += "unclassfied;";
202 addSeqToTree("unknown", unknownTax);
204 //create file for summary if needed
207 catch(exception& e) {
208 m->errorOut(e, "PhyloTree", "PhyloTree");
213 /**************************************************************************************************/
215 string PhyloTree::getNextTaxon(string& heirarchy, string seqname){
217 string currentLevel = "";
219 int pos = heirarchy.find_first_of(';');
221 if (pos == -1) { //you can't find another ;
222 currentLevel = heirarchy;
224 m->mothurOut(seqname + " is missing a ;, please check for other errors."); m->mothurOutEndLine();
226 currentLevel=heirarchy.substr(0,pos);
227 if (pos != (heirarchy.length()-1)) { heirarchy=heirarchy.substr(pos+1); }
228 else { heirarchy = ""; }
234 catch(exception& e) {
235 m->errorOut(e, "PhyloTree", "getNextTaxon");
240 /**************************************************************************************************/
242 int PhyloTree::addSeqToTree(string seqName, string seqTaxonomy){
246 map<string, int>::iterator childPointer;
251 tree[0].accessions.push_back(seqName);
252 m->removeConfidences(seqTaxonomy);
254 string taxon;// = getNextTaxon(seqTaxonomy);
256 while(seqTaxonomy != ""){
260 if (m->control_pressed) { return 0; }
262 //somehow the parent is getting one too many accnos
263 //use print to reassign the taxa id
264 taxon = getNextTaxon(seqTaxonomy, seqName);
266 if (taxon == "") { m->mothurOut(seqName + " has an error in the taxonomy. This may be due to a ;;"); m->mothurOutEndLine(); if (currentNode != 0) { uniqueTaxonomies[currentNode] = currentNode; } break; }
268 childPointer = tree[currentNode].children.find(taxon);
270 if(childPointer != tree[currentNode].children.end()){ //if the node already exists, move on
271 currentNode = childPointer->second;
272 tree[currentNode].accessions.push_back(seqName);
273 name2Taxonomy[seqName] = currentNode;
275 else{ //otherwise, create it
276 tree.push_back(TaxNode(taxon));
278 tree[currentNode].children[taxon] = numNodes-1;
279 tree[numNodes-1].parent = currentNode;
281 currentNode = tree[currentNode].children[taxon];
282 tree[currentNode].accessions.push_back(seqName);
283 name2Taxonomy[seqName] = currentNode;
286 if (seqTaxonomy == "") { uniqueTaxonomies[currentNode] = currentNode; }
291 catch(exception& e) {
292 m->errorOut(e, "PhyloTree", "addSeqToTree");
296 /**************************************************************************************************/
297 vector<int> PhyloTree::getGenusNodes() {
300 //generate genusIndexes
301 map<int, int>::iterator it2;
302 for (it2=uniqueTaxonomies.begin(); it2!=uniqueTaxonomies.end(); it2++) { genusIndex.push_back(it2->first); }
306 catch(exception& e) {
307 m->errorOut(e, "PhyloTree", "getGenusNodes");
311 /**************************************************************************************************/
312 vector<int> PhyloTree::getGenusTotals() {
317 //reset counts because we are on a new word
318 for (int j = 0; j < genusIndex.size(); j++) {
319 totals.push_back(tree[genusIndex[j]].accessions.size());
327 catch(exception& e) {
328 m->errorOut(e, "PhyloTree", "getGenusNodes");
332 /**************************************************************************************************/
334 void PhyloTree::assignHeirarchyIDs(int index){
336 map<string,int>::iterator it;
339 for(it=tree[index].children.begin();it!=tree[index].children.end();it++){
340 tree[it->second].heirarchyID = tree[index].heirarchyID + '.' + toString(counter);
342 tree[it->second].level = tree[index].level + 1;
344 //save maxLevel for binning the unclassified seqs
345 if (tree[it->second].level > maxLevel) { maxLevel = tree[it->second].level; }
347 assignHeirarchyIDs(it->second);
350 catch(exception& e) {
351 m->errorOut(e, "PhyloTree", "assignHeirarchyIDs");
355 /**************************************************************************************************/
356 void PhyloTree::setUp(string tfile){
358 string taxFileNameTest = tfile.substr(0,tfile.find_last_of(".")+1) + "tree.sum";
362 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
364 if (pid == 0) { binUnclassified(taxFileNameTest); }
367 binUnclassified(taxFileNameTest);
370 catch(exception& e) {
371 m->errorOut(e, "PhyloTree", "setUp");
375 /**************************************************************************************************/
376 void PhyloTree::binUnclassified(string file){
380 m->openOutputFile(file, out);
382 map<string, int>::iterator itBin;
383 map<string, int>::iterator childPointer;
385 vector<TaxNode> copy = tree;
388 fillOutTree(0, copy);
390 //get leaf nodes that may need extension
391 for (int i = 0; i < copy.size(); i++) {
393 if (copy[i].children.size() == 0) {
398 int copyNodes = copy.size();
400 //go through the seqs and if a sequence finest taxon is not the same level as the most finely defined taxon then classify it as unclassified where necessary
401 map<int, int>::iterator itLeaf;
402 for (itLeaf = leafNodes.begin(); itLeaf != leafNodes.end(); itLeaf++) {
404 if (m->control_pressed) { out.close(); break; }
406 int level = copy[itLeaf->second].level;
407 int currentNode = itLeaf->second;
409 //this sequence is unclassified at some levels
410 while(level < maxLevel){
414 string taxon = "unclassified";
416 //does the parent have a child names 'unclassified'?
417 childPointer = copy[currentNode].children.find(taxon);
419 if(childPointer != copy[currentNode].children.end()){ //if the node already exists, move on
420 currentNode = childPointer->second; //currentNode becomes 'unclassified'
422 else{ //otherwise, create it
423 copy.push_back(TaxNode(taxon));
425 copy[currentNode].children[taxon] = copyNodes-1;
426 copy[copyNodes-1].parent = currentNode;
427 copy[copyNodes-1].level = copy[currentNode].level + 1;
429 currentNode = copy[currentNode].children[taxon];
434 if (!m->control_pressed) {
440 catch(exception& e) {
441 m->errorOut(e, "PhyloTree", "binUnclassified");
445 /**************************************************************************************************/
446 void PhyloTree::fillOutTree(int index, vector<TaxNode>& copy) {
449 map<string,int>::iterator it;
451 it = copy[index].children.find("unclassified");
452 if (it == copy[index].children.end()) { //no unclassified at this level
453 string taxon = "unclassified";
454 copy.push_back(TaxNode(taxon));
455 copy[index].children[taxon] = copy.size()-1;
456 copy[copy.size()-1].parent = index;
457 copy[copy.size()-1].level = copy[index].level + 1;
460 if (tree[index].level < maxLevel) {
461 for(it=tree[index].children.begin();it!=tree[index].children.end();it++){ //check your children
462 fillOutTree(it->second, copy);
467 catch(exception& e) {
468 m->errorOut(e, "PhyloTree", "fillOutTree");
472 /**************************************************************************************************/
473 string PhyloTree::getFullTaxonomy(string seqName) {
477 int currentNode = name2Taxonomy[seqName];
479 while (tree[currentNode].parent != -1) {
480 tax = tree[currentNode].name + ";" + tax;
481 currentNode = tree[currentNode].parent;
486 catch(exception& e) {
487 m->errorOut(e, "PhyloTree", "getFullTaxonomy");
491 /**************************************************************************************************/
493 void PhyloTree::print(ofstream& out, vector<TaxNode>& copy){
496 //output mothur version
497 out << "#" << m->getVersion() << endl;
499 out << copy.size() << endl;
501 out << maxLevel << endl;
503 for (int i = 0; i < copy.size(); i++) {
505 out << copy[i].level << '\t'<< copy[i].name << '\t' << copy[i].children.size() << '\t';
507 map<string,int>::iterator it;
508 for(it=copy[i].children.begin();it!=copy[i].children.end();it++){
509 out << it->first << '\t' << it->second << '\t';
516 catch(exception& e) {
517 m->errorOut(e, "PhyloTree", "print");
521 /**************************************************************************************************/
522 void PhyloTree::printTreeNodes(string treefilename) {
527 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
534 m->openOutputFile(treefilename, outTree);
536 //output mothur version
537 outTree << "#" << m->getVersion() << endl;
540 outTree << tree.size() << endl;
541 for (int i = 0; i < tree.size(); i++) {
542 outTree << tree[i].name << '\t' << tree[i].level << '\t' << tree[i].parent << endl;
546 outTree << endl << uniqueTaxonomies.size() << endl;
547 map<int, int>::iterator it2;
548 for (it2=uniqueTaxonomies.begin(); it2!=uniqueTaxonomies.end(); it2++) { outTree << it2->first << '\t' << tree[it2->first].accessions.size() << endl; }
559 catch(exception& e) {
560 m->errorOut(e, "PhyloTree", "printTreeNodes");
564 /**************************************************************************************************/
565 TaxNode PhyloTree::get(int i ){
567 if (i < tree.size()) { return tree[i]; }
568 else { cout << i << '\t' << tree.size() << endl ; m->mothurOut("Mismatch with taxonomy and template files. Cannot continue."); m->mothurOutEndLine(); exit(1); }
570 catch(exception& e) {
571 m->errorOut(e, "PhyloTree", "get");
575 /**************************************************************************************************/
576 TaxNode PhyloTree::get(string seqName){
578 map<string, int>::iterator itFind = name2Taxonomy.find(seqName);
580 if (itFind != name2Taxonomy.end()) { return tree[name2Taxonomy[seqName]]; }
581 else { m->mothurOut("Cannot find " + seqName + ". Mismatch with taxonomy and template files. Cannot continue."); m->mothurOutEndLine(); exit(1);}
583 catch(exception& e) {
584 m->errorOut(e, "PhyloTree", "get");
588 /**************************************************************************************************/
589 string PhyloTree::getName(int i ){
591 if (i < tree.size()) { return tree[i].name; }
592 else { m->mothurOut("Mismatch with taxonomy and template files. Cannot continue."); m->mothurOutEndLine(); exit(1); }
594 catch(exception& e) {
595 m->errorOut(e, "PhyloTree", "get");
599 /**************************************************************************************************/
600 int PhyloTree::getIndex(string seqName){
602 map<string, int>::iterator itFind = name2Taxonomy.find(seqName);
604 if (itFind != name2Taxonomy.end()) { return name2Taxonomy[seqName]; }
605 else { m->mothurOut("Cannot find " + seqName + ". Mismatch with taxonomy and template files. Cannot continue."); m->mothurOutEndLine(); exit(1);}
607 catch(exception& e) {
608 m->errorOut(e, "PhyloTree", "get");
612 /**************************************************************************************************/
613 bool PhyloTree::ErrorCheck(vector<string> templateFileNames){
617 templateFileNames.push_back("unknown");
619 map<string, int>::iterator itFind;
620 map<string, int> taxonomyFileNames = name2Taxonomy;
622 for (int i = 0; i < templateFileNames.size(); i++) {
623 itFind = taxonomyFileNames.find(templateFileNames[i]);
625 if (itFind != taxonomyFileNames.end()) { //found it so erase it
626 taxonomyFileNames.erase(itFind);
628 m->mothurOut(templateFileNames[i] + " is in your template file and is not in your taxonomy file. Please correct."); m->mothurOutEndLine();
632 //templateFileNames.erase(templateFileNames.begin()+i);
635 templateFileNames.clear();
637 if (taxonomyFileNames.size() > 0) { //there are names in tax file that are not in template
640 for (itFind = taxonomyFileNames.begin(); itFind != taxonomyFileNames.end(); itFind++) {
641 m->mothurOut(itFind->first + " is in your taxonomy file and is not in your template file. Please correct."); m->mothurOutEndLine();
647 catch(exception& e) {
648 m->errorOut(e, "PhyloTree", "ErrorCheck");
652 /**************************************************************************************************/