An example of parsing LIBSVM input format with Boost library

parse_libsvm_input.cpp

/*
 * No Copyright
 */
#include <boost/algorithm/string.hpp>
#include <boost/foreach.hpp>
#include <boost/shared_ptr.hpp>
#include <boost/lexical_cast.hpp>
#include <iostream>  // NOLINT(readability/streams)
#include <fstream>  // NOLINT(readability/streams)
#include <string>
#include <vector>
#include <utility>

using boost::shared_ptr;
using boost::lexical_cast;
using std::pair;
using std::vector;
using std::string;
using std::cout;
using std::cerr;
using std::endl;

typedef vector<float> blob_type;
typedef pair<string, shared_ptr<blob_type> > datum_type;
typedef vector< shared_ptr<datum_type> > data_type;

shared_ptr<datum_type> get_datum_form_line_or_die(string line, int channels) {
  // buff
  vector<string> cells, indval;
  
  // split lines
  boost::split(cells, line, boost::is_any_of(" \t"));
  string label = cells[0];
  shared_ptr<blob_type> blob(new blob_type(channels));

  // Parse sparse format features
  int i = 0;
  BOOST_FOREACH(string cell, cells) {
    if (i++ == 0) { continue; }
    boost::split(indval, cell, boost::is_any_of(":"));
    if (indval.size() != 2) {
      cerr << "aaaaaa" << cell << endl;
      exit(1);
    }
    unsigned int ind = lexical_cast<unsigned int>(indval[0]);
    float val = lexical_cast<float>(indval[1]);
    (*blob)[ind] = val;
  }
  return shared_ptr<datum_type>(new datum_type(label, blob));
}

void display_data(data_type data) {
  BOOST_FOREACH(shared_ptr<datum_type> d, data) {
    cout << d->first << " -- ";
    BOOST_FOREACH(float v, *(d->second)) {
        cout << v << " ";
    }
    cout << endl;
  }
}

int main() {
  int channels = 123;
  std::ifstream infile("input.txt");
  data_type data;

  // read datum for each line
  string line;
  while (std::getline(infile, line)) {
    // trim spaces
    boost::trim(line);
    // skip empty lines
    if (line.empty()) {
      cout << "empty line" << endl;
      continue;
    }
    data.push_back(get_datum_form_line_or_die(line, channels));
  }
  display_data(data);
  return 0;
}