#include <boost/filesystem/fstream.hpp>
namespace fs = boost::filesystem;
+#include <boost/algorithm/string.hpp>
+
#include <iostream>
#include <sstream>
#include "mussa_exceptions.hpp"
-#include "alg/flp.hpp"
-#include "alg/mussa.hpp"
-#include "alg/motif_parser.hpp"
+
+#include "flp.hpp"
+#include "io.hpp"
+#include "mussa.hpp"
+#include "motif_parser.hpp"
using namespace std;
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) {
return soft_thres;
}
+
void
Mussa::set_analysis_mode(enum analysis_modes new_ana_mode)
{
set_dirty(true);
}
+void Mussa::load_mupa_file(std::string para_file_path) {
+ load_mupa_file(boost::filesystem::path(para_file_path));
+}
+
void
Mussa::load_mupa_file(fs::path para_file_path)
{
- fs::ifstream para_file;
- string file_data_line;
- string param, value;
- 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;
-
- // initialize values
- clear();
-
- // if file was opened, read the parameter values
if (not fs::exists(para_file_path))
{
throw mussa_load_error("Config File: " + para_file_path.string() + " not found");
} 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();
+ }
+}
- // what directory is the mupa file in?
- fs::path file_path_base = para_file_path.branch_path();
+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;
+ std::string err_msg;
+ // initialize values
+ clear();
+
+
+ while (para_file.good())
+ {
+ ++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)
- {
- 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 && 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) {
+ cout << "seq_file_name call2" << seq_file << endl;
+ 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;
+ // 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);
}
- // no file was loaded, signal error
+
+ soft_thres = threshold;
set_dirty(true);
}
vector<FLPs> empty_FLP_vector;
FLPs dummy_comp;
+
+ //--------------------------------------------------------
+ // Load Muway
+ //--------------------------------------------------------
analysis_path = ana_file;
analysis_name = ana_path.leaf();
fs::path muway(analysis_name+".muway", fs::native);
// 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();
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++)
{
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++)
projects / mussa.git / blobdiff