Convert edge lists described by names to those described by 0-indexed vertex IDs

shrink_ids.cc

/*
 * Basic usage:
 *   $ CXXFLAGS=-O3 make shrink_ids
 *   $ ./shrink_ids < EDGES > SHRINKED_EDGES
 *
 * Example:
 *   $ echo "hoge piyo\nfuga hoge" | ./shrink_ids
 *   0	1
 *   2	0
 *
 * Options:
 *   -f TARGET_FIELDS  --- Specify target fields (just like 'cut' command)
 *   -v OUTPUT_VERTEX_NAME_FILE --- Output names of vertices
 *
 * More examples:
 *   $ echo "01:00 piyo fuga\n02:15 fuga hoge" | ./shrink_ids -f 2,3 -v tmp.txt
 *   01:00	0	1
 *   02:15	1	2
 *   $ cat tmp.txt
 *   piyo
 *   fuga
 *   hoge
 */

#include <cstdio>
#include <cstring>
#include <cstdlib>
#include <fstream>
#include <iostream>
#include <sstream>
#include <algorithm>
#include <map>
#include <vector>
using namespace std;

int main(int argc, char **argv) {
  std::ios_base::sync_with_stdio();

  //
  // Parse commandline frags
  //
  vector<int> option_target_fields(2);
  option_target_fields[0] = 1;
  option_target_fields[1] = 2;
  ofstream ofs_vertices;

  for (int i = 1; i < argc; ++i) {
    if (strncmp(argv[i], "-f", 2) == 0) {
      string s(argv[i] + 2);
      if (s.length() == 0) {
        if (++i >= argc) {
          fprintf(stderr, "ERROR: Bad args (after '-f')\n");
          exit(EXIT_FAILURE);
        }
        s = argv[i];
      }
      for (size_t j = 0; j < s.length(); ++j) {
        if (s[j] == ',') s[j] = ' ';
      }
      istringstream ss(s);
      option_target_fields.clear();
      for (int f; ss >> f; ) {
        option_target_fields.push_back(f);
      }
      if (!ss.eof() || ss.bad()) {
        fprintf(stderr, "Error: Bad args (after '-f')\n");
        exit(EXIT_FAILURE);
      }
    } else if (strncmp(argv[i], "-v", 2) == 0) {
      if (++i >= argc) {
        fprintf(stderr, "ERROR: Bad args (after '-v')\n");
        exit(EXIT_FAILURE);
      }
      ofs_vertices.open(argv[i]);
      if (!ofs_vertices) {
        perror("Error: Failed to open file");
        exit(EXIT_FAILURE);
      }
    } else {
      fprintf(stderr, "Error: Bad args\n");
      exit(EXIT_FAILURE);
    }
  }
  sort(option_target_fields.begin(), option_target_fields.end());
  option_target_fields.erase(unique(option_target_fields.begin(), option_target_fields.end()),
                             option_target_fields.end());

  //
  // Shrink
  //
  map<string, int> ma;
  size_t num_e = 0;
  for (string line; getline(cin, line); num_e++) {
    istringstream ss(line);
    string s;
    for (int f = 1; ss >> s; ++f) {
      if (f > 1) cout << "\t";
      if (binary_search(option_target_fields.begin(), option_target_fields.end(), f)) {
        pair<map<string, int>::iterator, bool> p = ma.insert(make_pair(s, ma.size()));
        cout << p.first->second;
        if (p.second && ofs_vertices.is_open()) ofs_vertices << s << endl;
      } else {
        cout << s;
      }
    }
    cout << endl;
  }

  cerr << "Number of Vertices: " << ma.size() << endl;
  cerr << "Number of Edges:    " << num_e << endl;
  exit(EXIT_SUCCESS);
}