// ---------- mussa_class.cc -----------
// ----------------------------------------
+#include <boost/filesystem/path.hpp>
#include <boost/filesystem/operations.hpp>
#include <boost/filesystem/fstream.hpp>
namespace fs = boost::filesystem;
+#include <boost/algorithm/string.hpp>
+
#include <iostream>
#include <sstream>
#include "mussa_exceptions.hpp"
-#include "alg/mussa.hpp"
-#include "alg/flp.hpp"
+
+#include "flp.hpp"
+#include "io.hpp"
+#include "mussa.hpp"
+#include "motif_parser.hpp"
using namespace std;
Mussa::Mussa()
+ : color_mapper(new AnnotationColors)
{
clear();
- connect(&the_paths, SIGNAL(progress(const std::string&, int, int)),
- this, SIGNAL(progress(const std::string&, int, int)));
+ connect(&the_paths, SIGNAL(progress(const QString&, int, int)),
+ this, SIGNAL(progress(const QString&, int, int)));
}
Mussa::Mussa(const Mussa& m)
win_append(m.win_append),
thres_append(m.thres_append),
motif_sequences(m.motif_sequences),
- color_mapper(m.color_mapper)
+ color_mapper(m.color_mapper),
+ analysis_path(m.analysis_path),
+ dirty(m.dirty)
+{
+ connect(&the_paths, SIGNAL(progress(const QString&, int, int)),
+ this, SIGNAL(progress(const QString&, int, int)));
+}
+
+MussaRef Mussa::init()
+{
+ boost::shared_ptr<Mussa> m(new Mussa());
+ return m;
+}
+
+boost::filesystem::path Mussa::get_analysis_path() const
+{
+ return analysis_path;
+}
+
+void Mussa::set_analysis_path(boost::filesystem::path pathname)
{
- connect(&the_paths, SIGNAL(progress(const std::string&, int, int)),
- this, SIGNAL(progress(const std::string&, int, int)));
+ analysis_path = pathname;
}
// set all parameters to null state
win_append = false;
thres_append = false;
motif_sequences.clear();
- color_mapper.clear();
+ if(color_mapper) color_mapper->clear();
+ the_seqs.clear();
the_paths.clear();
+ analysis_path = fs::path();
+ set_dirty(false);
+}
+
+void Mussa::set_append_window(bool v)
+{
+ win_append = v;
+}
+
+bool Mussa::get_append_window()
+{
+ return win_append;
+}
+
+void Mussa::set_append_threshold(bool v)
+{
+ thres_append = v;
+}
+
+bool Mussa::get_append_threshold()
+{
+ return thres_append;
+}
+
+void Mussa::set_dirty(bool new_state)
+{
+ if (dirty != new_state) {
+ dirty = new_state;
+ emit isModified(dirty);
+ }
+}
+
+bool Mussa::is_dirty() const
+{
+ return dirty;
}
+bool Mussa::empty() const
+{
+ return the_seqs.empty();
+}
+
+
// these 5 simple methods manually set the parameters for doing an analysis
// used so that the gui can take input from user and setup the analysis
// note - still need a set_append(bool, bool) method...
Mussa::set_name(string a_name)
{
analysis_name = a_name;
+ set_dirty(true);
}
-string Mussa::get_name()
+string Mussa::get_name() const
{
return analysis_name;
}
+string Mussa::get_title() const
+{
+ fs::path analysis_path = get_analysis_path();
+ if (not analysis_path.empty()) {
+ return analysis_path.native_file_string();
+ } else if (get_name().size() > 0) {
+ return get_name();
+ } else {
+ return std::string("Unnamed");
+ }
+}
+
int
Mussa::size() const
{
Mussa::set_window(int a_window)
{
window = a_window;
+ set_dirty(true);
}
int Mussa::get_window() const
Mussa::set_threshold(int a_threshold)
{
threshold = a_threshold;
- if (a_threshold > soft_thres)
+ set_dirty(true);
+ if (a_threshold > soft_thres) {
soft_thres = a_threshold;
+ }
}
int Mussa::get_threshold() const
Mussa::set_soft_threshold(int new_threshold)
{
if (new_threshold < threshold) {
- soft_thres = threshold;
+ new_threshold = threshold;
} else if (new_threshold > window) {
- soft_thres = window;
- } else {
+ new_threshold = window;
+ }
+ if (soft_thres != new_threshold) {
soft_thres = new_threshold;
+ nway();
}
}
return soft_thres;
}
+
void
Mussa::set_analysis_mode(enum analysis_modes new_ana_mode)
{
ana_mode = new_ana_mode;
+ set_dirty(true);
}
enum Mussa::analysis_modes Mussa::get_analysis_mode() const
std::list<ConservedPath::path_type>& result,
std::list<std::vector<bool> >& reversed)
{
- const vector<Sequence>& raw_seq = the_seqs;
+ const vector_sequence_type& raw_seq = the_seqs;
ConservedPath::path_type aligned_path;
size_t i2, i3;
int x_start, x_end;
x_start = (abs(a_path[i2]-rc_1+win_i));
x_end = (abs(a_path[i2+1]-rc_2+win_i));
+ boost::shared_ptr<Sequence> cur(raw_seq[i2]) ;
+ boost::shared_ptr<Sequence> next(raw_seq[i2+1]);
// RC case handling
// ugh, and xor...only want rc coloring if just one of the nodes is rc
// if both nodes are rc, then they are 'normal' relative to each other
if((rc_list[i2] || rc_list[i2+1] )&&!(rc_list[i2] && rc_list[i2+1]))
{ //the hideous rc matching logic - not complex, but annoying
- if(!(( (raw_seq[i2][x_start]=='A')&&(raw_seq[i2+1][x_end]=='T')) ||
- ((raw_seq[i2][x_start]=='T')&&(raw_seq[i2+1][x_end]=='A')) ||
- ((raw_seq[i2][x_start]=='G')&&(raw_seq[i2+1][x_end]=='C')) ||
- ((raw_seq[i2][x_start]=='C')&&(raw_seq[i2+1][x_end]=='G'))) )
+ if(!(( ((*cur)[x_start]=='A')&&((*next)[x_end]=='T')) ||
+ (((*cur)[x_start]=='T')&&((*next)[x_end]=='A')) ||
+ (((*cur)[x_start]=='G')&&((*next)[x_end]=='C')) ||
+ (((*cur)[x_start]=='C')&&((*next)[x_end]=='G'))) )
{
full_match = false;
} else {
}
else
{ // forward match
- if (!( (raw_seq[i2][x_start] == raw_seq[i2+1][x_end]) &&
- (raw_seq[i2][x_start] != 'N') &&
- (raw_seq[i2+1][x_end] != 'N') ) ) {
+ if (!( ((*cur)[x_start] == (*next)[x_end]) &&
+ ((*cur)[x_start] != 'N') && ((*next)[x_end] != 'N') ) ) {
full_match = false;
} else {
aligned_path.push_back(x_start);
}
-void Mussa::append_sequence(Sequence a_seq)
+void Mussa::append_sequence(const Sequence& a_seq)
+{
+ boost::shared_ptr<Sequence> seq_copy(new Sequence(a_seq));
+ the_seqs.push_back(seq_copy);
+ set_dirty(true);
+}
+
+void Mussa::append_sequence(boost::shared_ptr<Sequence> a_seq)
{
the_seqs.push_back(a_seq);
+ set_dirty(true);
}
-const vector<Sequence>&
+
+const vector<SequenceRef>&
Mussa::sequences() const
{
return the_seqs;
}
void Mussa::load_sequence(fs::path seq_file, fs::path annot_file,
- int fasta_index, int sub_seq_start, int sub_seq_end)
+ int fasta_index, int sub_seq_start, int sub_seq_end,
+ std::string *name)
{
- Sequence aseq;
- aseq.load_fasta(seq_file, fasta_index, sub_seq_start, sub_seq_end);
+ boost::shared_ptr<Sequence> aseq(new Sequence);
+ aseq->load_fasta(seq_file, fasta_index, sub_seq_start, sub_seq_end);
if ( not annot_file.empty() ) {
- aseq.load_annot(annot_file, sub_seq_start, sub_seq_end);
+ aseq->load_annot(annot_file, sub_seq_start, sub_seq_end);
+ }
+ if (name != 0 and name->size() > 0 ) {
+ aseq->set_species(*name);
}
the_seqs.push_back(aseq);
+ set_dirty(true);
}
void
Mussa::load_mupa_file(fs::path para_file_path)
{
- fs::ifstream para_file;
- string file_data_line;
- string param, value;
+ if (not fs::exists(para_file_path))
+ {
+ throw mussa_load_error("Config File: " + para_file_path.string() + " not found");
+ } else if (fs::is_directory(para_file_path)) {
+ throw mussa_load_error("Config File: " + para_file_path.string() + " is a directory.");
+ } else if (fs::is_empty(para_file_path)) {
+ throw mussa_load_error("Config File: " + para_file_path.string() + " is empty");
+ } else {
+ // what directory is the mupa file in?
+ fs::path file_path_base( para_file_path.branch_path()) ;
+
+ fs::ifstream para_file;
+ para_file.open(para_file_path, ios::in);
+
+ load_mupa_stream(para_file, file_path_base);
+ para_file.close();
+ }
+}
+
+void
+Mussa::load_mupa_stream(std::istream& para_file, fs::path& file_path_base)
+{
+ std::string line;
+ std::vector< std::string > tokens;
+ int line_count = 0;
+
+ enum parsing_state_enum { START, INSEQUENCE };
+ parsing_state_enum parsing_state = START;
+
+ // sequence file parameters
+ fs::path seq_file;
fs::path annot_file;
int split_index, fasta_index;
int sub_seq_start, sub_seq_end;
bool seq_params, did_seq;
- string err_msg;
- bool parsing_path;
- string::size_type new_index, dir_index;
+ std::string err_msg;
// initialize values
clear();
- // if file was opened, read the parameter values
- if (fs::exists(para_file_path))
+
+ while (para_file.good())
{
- para_file.open(para_file_path, ios::in);
-
- // what directory is the mupa file in?
- fs::path file_path_base = para_file_path.branch_path();
-
+ ++line_count;
// setup loop by getting file's first line
- getline(para_file,file_data_line);
- split_index = file_data_line.find(" ");
- param = file_data_line.substr(0,split_index);
- value = file_data_line.substr(split_index+1);
-
- while (!para_file.eof())
- {
- did_seq = false;
- if (param == "ANA_NAME")
- analysis_name = value;
- else if (param == "APPEND_WIN")
- win_append = true;
- else if (param == "APPEND_THRES")
- thres_append = true;
- else if (param == "SEQUENCE_NUM")
- ; // ignore sequence_num now
- else if (param == "WINDOW")
- window = atoi(value.c_str());
- else if (param == "THRESHOLD")
- threshold = atoi(value.c_str());
- else if (param == "SEQUENCE")
- {
- fs::path seq_file = file_path_base / value;
- //cout << "seq_file_name " << seq_files.back() << endl;
- fasta_index = 1;
- annot_file = "";
- sub_seq_start = 0;
- sub_seq_end = 0;
- seq_params = true;
-
- while ((!para_file.eof()) && seq_params)
- {
- getline(para_file,file_data_line);
- split_index = file_data_line.find(" ");
- param = file_data_line.substr(0,split_index);
- value = file_data_line.substr(split_index+1);
-
- if (param == "FASTA_INDEX")
- fasta_index = atoi(value.c_str());
- else if (param == "ANNOTATION")
- annot_file = file_path_base / value;
- else if (param == "SEQ_START")
- sub_seq_start = atoi(value.c_str());
- else if (param == "SEQ_END")
- {
- sub_seq_end = atoi(value.c_str());
- }
- //ignore empty lines or that start with '#'
- else if ((param == "") || (param == "#")) {}
- else seq_params = false;
- }
+ multiplatform_getline(para_file, line);
+ // strip leading/trailing whitespace
+ boost::trim(line);
+ // ignore commented out or blank lines
+ if ( line.size() == 0 or line[0] == '#' ) {
+ continue;
+ }
+
+ // split the line on white spance
+ boost::split(tokens, line, boost::is_space());
+ // do we have a name/value pair?
+ if (tokens.size() != 2) {
+ std::stringstream errmsg;
+ errmsg << "Error parsing MUPA file line: "
+ << line_count << std::endl
+ << line;
+ throw mussa_load_error(errmsg.str());
+ }
+
+ boost::to_upper(tokens[0]);
+ // Parameters only useful after a sequence block
+ if (parsing_state == INSEQUENCE) {
+ // in the following if blocks, if we do
+ // successfully match a token we should continue
+ // on to the next token
+ // but if we don't match a token we want to
+ // fall through to the top level parsing
+
+ if (tokens[0] == "FASTA_INDEX") {
+ fasta_index = atoi(tokens[1].c_str());
+ continue;
+ } else if (tokens[0] == "ANNOTATION") {
+ annot_file = file_path_base / tokens[1];
+ continue;
+ } else if (tokens[0] == "SEQ_START") {
+ sub_seq_start = atoi(tokens[1].c_str());
+ continue;
+ } else if (tokens[0] == "SEQ_END") {
+ sub_seq_end = atoi(tokens[1].c_str());
+ continue;
+ } else {
+ // any other token means we're done with this
+ // sequence so we should load it
+ // (and let the "unknown" token fall through into the
+ // top level token parser)
load_sequence(seq_file, annot_file, fasta_index, sub_seq_start,
sub_seq_end);
- did_seq = true;
- }
- //ignore empty lines or that start with '#'
- else if ((param == "") || (param == "#")) {}
- else
- {
- clog << "Illegal/misplaced mussa parameter in file\n";
- clog << param << "\n";
+ parsing_state = START;
}
-
- if (!did_seq)
- {
- getline(para_file,file_data_line);
- split_index = file_data_line.find(" ");
- param = file_data_line.substr(0,split_index);
- value = file_data_line.substr(split_index+1);
- did_seq = false;
+ }
+ // if we didn't consume a token from the previous if block
+ // try
+ // top level token parsing
+ if (tokens[0] == "ANA_NAME") {
+ analysis_name = tokens[1];
+ } else if (tokens[0] == "APPEND_WIN") {
+ win_append = true;
+ } else if (tokens[0] == "APPEND_THRES") {
+ thres_append = true;
+ } else if (tokens[0] == "SEQUENCE_NUM") {
+ ; // ignore sequence_num now
+ } else if (tokens[0] == "WINDOW") {
+ window = atoi(tokens[1].c_str());
+ } else if (tokens[0] == "THRESHOLD") {
+ threshold = atoi(tokens[1].c_str());
+ } else if (tokens[0] == "SEQUENCE") {
+ if (parsing_state == INSEQUENCE) {
+ load_sequence(seq_file, annot_file, fasta_index, sub_seq_start,
+ sub_seq_end);
+ parsing_state = START;
}
+ // reset sequence parameters
+ seq_file = file_path_base / tokens[1];
+ fasta_index = 1;
+ annot_file = "";
+ sub_seq_start = 0;
+ sub_seq_end = 0;
+ seq_params = true;
+ parsing_state = INSEQUENCE;
+ } else {
+ clog << "Illegal/misplaced mussa parameter in file\n";
+ clog << tokens[0] << "\n";
+ std::stringstream errmsg;
+ errmsg << "Invalid mussa paaramater '"
+ << tokens[0]
+ << "' on line: "
+ << line_count << std::endl
+ << line;
+ throw mussa_load_error(errmsg.str());
+ throw mussa_load_error("Error parsing MUPA file");
}
-
- para_file.close();
-
- soft_thres = threshold;
- //cout << "nway mupa: analysis_name = " << analysis_name
- // << " window = " << window
- // << " threshold = " << threshold << endl;
}
- // no file was loaded, signal error
- else
- {
- throw mussa_load_error("Config File: " + para_file_path.string() + " not found");
+
+ // if we hit the end of the file and there's a sequence
+ // pending, go ahead and load it
+ if (parsing_state == INSEQUENCE) {
+ load_sequence(seq_file, annot_file, fasta_index, sub_seq_start,
+ sub_seq_end);
}
+
+ soft_thres = threshold;
+ set_dirty(true);
}
void
Mussa::analyze()
{
- time_t t1, t2, begin, end;
- double seqloadtime, seqcomptime, nwaytime, savetime, totaltime;
-
- begin = time(NULL);
-
- t1 = time(NULL);
-
if (the_seqs.size() < 2) {
throw mussa_analysis_error("you need to have at least 2 sequences to "
"do an analysis.");
}
- //cout << "nway ana: seq_num = " << the_seqs.size() << endl;
-
- t2 = time(NULL);
- seqloadtime = difftime(t2, t1);
-
- t1 = time(NULL);
+
seqcomp();
- t2 = time(NULL);
- seqcomptime = difftime(t2, t1);
-
- t1 = time(NULL);
the_paths.setup(window, threshold);
nway();
- t2 = time(NULL);
- nwaytime = difftime(t2, t1);
-
- t1 = time(NULL);
- save();
- t2 = time(NULL);
- savetime = difftime(t2, t1);
-
- end = time(NULL);
- totaltime = difftime(end, begin);
-
- //cout << "seqload\tseqcomp\tnway\tsave\ttotal\n";
- //cout << seqloadtime << "\t";
- //cout << seqcomptime << "\t";
- //cout << nwaytime << "\t";
- //cout << savetime << "\t";
- //cout << totaltime << "\n";
}
void
all_comps[i].push_back(dummy_comp);
}
for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++) {
- seq_lens.push_back(the_seqs[i].size());
+ seq_lens.push_back(the_seqs[i]->size());
}
int seqcomps_done = 0;
int seqcomps_todo = (the_seqs.size() * (the_seqs.size()-1)) / 2;
{
//cout << "seqcomping: " << i << " v. " << i2 << endl;
all_comps[i][i2].setup(window, threshold);
- all_comps[i][i2].seqcomp(the_seqs[i], the_seqs[i2], false);
- all_comps[i][i2].seqcomp(the_seqs[i], the_seqs[i2].rev_comp(),true);
+ all_comps[i][i2].seqcomp(*the_seqs[i], *the_seqs[i2], false);
+ all_comps[i][i2].seqcomp(*the_seqs[i], the_seqs[i2]->rev_comp(),true);
++seqcomps_done;
emit progress("seqcomp", seqcomps_done, seqcomps_todo);
}
void
Mussa::nway()
{
- vector<string> some_Seqs;
the_paths.set_soft_threshold(soft_thres);
}
else if (ana_mode == EntropyNway)
{
+ vector<Sequence> some_Seqs;
//unlike other methods, entropy needs to look at the sequence at this stage
some_Seqs.clear();
for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
{
- some_Seqs.push_back(the_seqs[i]);
+ some_Seqs.push_back(*the_seqs[i]);
}
the_paths.setup_ent(ent_thres, some_Seqs); // ent analysis extra setup
}
void
-Mussa::save()
+Mussa::save(fs::path save_path)
{
- string save_name;
- fs::path flp_filepath;
fs::fstream save_file;
ostringstream append_info;
int dir_create_status;
- if (not analysis_name.empty()) {
- // not sure why, but gotta close file each time since can't pass
- // file streams
- save_name = analysis_name;
+ if (save_path.empty()) {
+ if (not analysis_path.empty()) {
+ save_path = analysis_path;
+ } else if (not analysis_name.empty()) {
+ std::string save_name = analysis_name;
+ // gotta do bit with adding win & thres if to be appended
+ if (win_append) {
+ append_info.str("");
+ append_info << "_w" << window;
+ save_name += append_info.str();
+ }
- // gotta do bit with adding win & thres if to be appended
- if (win_append)
- {
- append_info.str("");
- append_info << "_w" << window;
- save_name += append_info.str();
+ if (thres_append) {
+ append_info.str("");
+ append_info << "_t" << threshold;
+ save_name += append_info.str();
+ }
+ save_path = save_name;
+ } else {
+ throw mussa_save_error("Need filename or analysis name to save");
}
+ }
- if (thres_append)
- {
- append_info.str("");
- append_info << "_t" << threshold;
- save_name += append_info.str();
- }
- fs::path save_path( save_name);
+ if (not fs::exists(save_path)) {
+ fs::create_directory(save_path);
+ }
+
+ std::string basename = save_path.leaf();
+ fs::path museq(basename + ".museq", fs::native);
+
+ // save sequence and annots to a special mussa file
+ save_file.open(save_path / museq, ios::out);
+ save_file << "<Mussa_Sequence>" << endl;
- if (not fs::exists(save_path)) {
- fs::create_directory(save_path);
- }
- // save sequence and annots to a special mussa file
- save_file.open(save_path / (save_name+".museq"), ios::out);
- save_file << "<Mussa_Sequence>" << endl;
+ for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
+ {
+ the_seqs[i]->save(save_file);
+ }
- for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
- {
- the_seqs[i].save(save_file);
- }
+ save_file << "</Mussa_Sequence>" << endl;
+ save_file.close();
- save_file << "</Mussa_Sequence>" << endl;
- save_file.close();
+ // if we have any motifs, save them.
+ if (motif_sequences.size()) {
+ fs::path mtl(basename + ".mtl", fs::native);
+ save_motifs(save_path / mtl);
+ }
- // save nway paths to its mussa save file
- the_paths.save(save_path / (save_name + ".muway"));
+ // save nway paths to its mussa save file
+ fs::path muway(basename + ".muway", fs::native);
+ the_paths.save(save_path / muway);
- for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
- for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
- {
- append_info.str("");
- append_info << "_sp_" << i << "v" << i2;
- all_comps[i][i2].save(save_path/(save_name+append_info.str()+".flp"));
- }
+ for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++) {
+ for(vector<Sequence>::size_type i2 = i+1; i2 < the_seqs.size(); i2++)
+ {
+ append_info.str("");
+ append_info << "_sp_" << i << "v" << i2;
+ fs::path flp(basename+append_info.str()+".flp", fs::native);
+ all_comps[i][i2].save(save_path / flp);
+ }
}
+
+ set_dirty(false);
+ analysis_path = save_path;
}
void
fs::path a_file_path;
fs::path ana_path(ana_file);
bool parsing_path;
- Sequence tmp_seq;
string err_msg;
ostringstream append_info;
vector<FLPs> empty_FLP_vector;
FLPs dummy_comp;
- //cout << "ana_file name " << ana_file.string() << endl;
+
+ //--------------------------------------------------------
+ // Load Muway
+ //--------------------------------------------------------
+ analysis_path = ana_file;
analysis_name = ana_path.leaf();
- //cout << " ana_name " << analysis_name << endl;
- file_path_base = ana_path.branch_path() / analysis_name;
- a_file_path = file_path_base / (analysis_name + ".muway");
- //cout << " loading museq: " << a_file_path.string() << endl;
+ fs::path muway(analysis_name+".muway", fs::native);
+ a_file_path = analysis_path / muway;
the_paths.load(a_file_path);
// perhaps this could be more elegent, but at least this'll let
// us know what our threshold and window sizes were when we load a muway
window = the_paths.get_window();
threshold = the_paths.get_threshold();
- soft_thres = threshold;
+ soft_thres = the_paths.get_soft_threshold();
+
- int seq_num = the_paths.sequence_count();
+ //--------------------------------------------------------
+ // Load Sequence
+ //--------------------------------------------------------
+ //int seq_num = the_paths.sequence_count();
- a_file_path = file_path_base / (analysis_name + ".museq");
+ fs::path museq(analysis_name + ".museq", fs::native);
+ a_file_path = analysis_path / museq;
// this is a bit of a hack due to C++ not acting like it should with files
- for (i = 1; i <= seq_num; i++)
+ /*for (i = 1; i <= seq_num; i++)
{
- tmp_seq.clear();
- //cout << "mussa_class: loading museq frag... " << a_file_path.string() << endl;
- tmp_seq.load_museq(a_file_path, i);
+ boost::shared_ptr<Sequence> tmp_seq(new Sequence);
+ tmp_seq->load_museq(a_file_path, i);
the_seqs.push_back(tmp_seq);
+ }*/
+
+ i = 1;
+ //int seq_num = 0;
+ boost::filesystem::fstream load_museq_fs;
+ load_museq_fs.open(a_file_path, std::ios::in);
+ boost::shared_ptr<Sequence> tmp_seq;
+ while (1)
+ {
+ tmp_seq = Sequence::load_museq(load_museq_fs);
+
+ if (tmp_seq)
+ {
+ the_seqs.push_back(tmp_seq);
+ }
+ else
+ {
+ break;
+ }
+
+
+ //safe guard in case of an infinate loop.
+ //FIXME: If mussa can handle a comparison of 10000 sequences
+ // in the future, then this code should be fixed.
+ if (i == 10000)
+ {
+ throw mussa_load_error(" Run away sequence load!");
+ }
+ i++;
+ }
+ load_museq_fs.close();
+
+ //--------------------------------------------------------
+ // Load Motifs
+ //--------------------------------------------------------
+ fs::path mtl(analysis_name + ".mtl", fs::native);
+ fs::path motif_file = analysis_path / mtl;
+ if (fs::exists(motif_file)) {
+ load_motifs(motif_file);
}
+ vector<Sequence>::size_type seq_num = the_seqs.size();
empty_FLP_vector.clear();
for(i = 0; i < seq_num; i++)
{
all_comps[i].push_back(dummy_comp);
}
+
for(i = 0; i < seq_num; i++)
{
for(i2 = i+1; i2 < seq_num; i2++)
{
append_info.str("");
append_info << analysis_name << "_sp_" << i << "v" << i2 << ".flp";
- //cout << append_info.str() << endl;
- a_file_path = file_path_base / append_info.str();
- //cout << "path " << a_file_path.string() << endl;
+ //clog << append_info.str() << endl;
+ fs::path flp(append_info.str(), fs::native);
+ a_file_path = analysis_path / flp;
all_comps[i][i2].load(a_file_path);
- //cout << "real size = " << all_comps[i][i2].size() << endl;
}
}
}
-
void
Mussa::save_old()
{
save_file.open(analysis_name, ios::out);
for(vector<Sequence>::size_type i = 0; i < the_seqs.size(); i++)
- save_file << the_seqs[i] << endl;
+ save_file << *(the_seqs[i]) << endl;
save_file << window << endl;
save_file.close();
for(i = 0; i < seq_num; i++)
{
getline(save_file, file_data_line);
- a_seq = file_data_line;
+ boost::shared_ptr<Sequence> a_seq(new Sequence(file_data_line));
the_seqs.push_back(a_seq);
}
//the_paths.save("tmp.save");
}
-void Mussa::add_motifs(const vector<string>& motifs,
+void Mussa::add_motif(const Sequence& motif, const Color& color)
+{
+ motif_sequences.insert(motif);
+ color_mapper->appendInstanceColor("motif", motif.get_sequence(), color);
+ set_dirty(true);
+}
+
+void Mussa::set_motifs(const vector<Sequence>& motifs,
const vector<Color>& colors)
{
if (motifs.size() != colors.size()) {
throw mussa_error("motif and color vectors must be the same size");
}
+ motif_sequences.clear();
for(size_t i = 0; i != motifs.size(); ++i)
{
- motif_sequences.insert(motifs[i]);
- color_mapper.appendInstanceColor("motif", motifs[i], colors[i]);
+ add_motif(motifs[i], colors[i]);
}
update_sequences_motifs();
}
-// I mostly split the ifstream out so I can use a stringstream to test it.
+void Mussa::load_motifs(fs::path filename)
+{
+ fs::ifstream f;
+ f.open(filename, ifstream::in);
+ load_motifs(f);
+}
+
void Mussa::load_motifs(std::istream &in)
{
- string seq;
- float red;
- float green;
- float blue;
-
- while(in.good())
- {
- in >> seq >> red >> green >> blue;
- // if we couldn't read this line 'cause we're like at the end of the file
- // try to exit the loop
- if (!in.good())
- break;
- try {
- seq = Sequence::motif_normalize(seq);
- } catch(motif_normalize_error e) {
- clog << "unable to parse " << seq << " skipping" << endl;
- clog << e.what() << endl;
- continue;
- }
- if (red < 0.0 or red > 1.0) {
- clog << "invalid red value " << red << ". must be in range [0..1]"
- << endl;
- continue;
- }
- if (green < 0.0 or green > 1.0) {
- clog << "invalid green value " << green << ". must be in range [0..1]"
- << endl;
- continue;
- }
- if (blue < 0.0 or blue > 1.0) {
- clog << "invalid blue value " << blue << ". must be in range [0..1]"
- << endl;
- continue;
- }
- if (motif_sequences.find(seq) == motif_sequences.end()) {
- // sequence wasn't found
- motif_sequences.insert(seq);
- Color c(red, green, blue);
- color_mapper.appendInstanceColor("motif", seq, c);
- } else {
- clog << "sequence " << seq << " was already defined skipping"
- << endl;
- continue;
- }
+ std::string data;
+ const char *alphabet = Alphabet::dna_cstr;
+ motif_parser::ParsedMotifs parsed_motifs(motif_sequences, color_mapper);
+
+ // slurp our data into a string
+ std::streamsize bytes_read = 1;
+ while (in.good() and bytes_read) {
+ const std::streamsize bufsiz=512;
+ char buf[bufsiz];
+ bytes_read = in.readsome(buf, bufsiz);
+ data.append(buf, buf+bytes_read);
}
+ parsed_motifs.parse(data);
update_sequences_motifs();
}
-void Mussa::load_motifs(fs::path filename)
+void Mussa::save_motifs(fs::path filename)
{
- fs::ifstream f;
- f.open(filename, ifstream::in);
- load_motifs(f);
+ fs::ofstream out_stream;
+ out_stream.open(filename, ofstream::out);
+ save_motifs(out_stream);
+}
+
+void Mussa::save_motifs(std::ostream& out)
+{
+ for(motif_set::iterator motif_i = motif_sequences.begin();
+ motif_i != motif_sequences.end();
+ ++motif_i)
+ {
+ out << motif_i->get_sequence() << " ";
+ if (motif_i->get_name().size() > 0) {
+ out << "\"" << motif_i->get_name() << "\" ";
+ }
+ out << color_mapper->lookup("motif", motif_i->get_sequence());
+ out << std::endl;
+ }
}
void Mussa::update_sequences_motifs()
{
// once we've loaded all the motifs from the file,
// lets attach them to the sequences
- for(vector<Sequence>::iterator seq_i = the_seqs.begin();
+ for(vector<SequenceRef >::iterator seq_i = the_seqs.begin();
seq_i != the_seqs.end();
++seq_i)
{
// clear out old motifs
- seq_i->clear_motifs();
+ (*seq_i)->clear_motifs();
// for all the motifs in our set, attach them to the current sequence
- for(set<string>::iterator motif_i = motif_sequences.begin();
+ for(set<Sequence>::iterator motif_i = motif_sequences.begin();
motif_i != motif_sequences.end();
++motif_i)
{
- seq_i->add_motif(*motif_i);
+ (*seq_i)->add_motif(*motif_i);
}
}
}
-const set<string>& Mussa::motifs() const
+const set<Sequence>& Mussa::motifs() const
{
return motif_sequences;
}
-AnnotationColors& Mussa::colorMapper()
+boost::shared_ptr<AnnotationColors> Mussa::colorMapper()
{
return color_mapper;
}