/*
 * $Id: main.cpp,v 1.8 2005/02/16 10:45:39 tsuruoka Exp $
 */

#include <stdio.h>
#include <fstream>
#include <map>
#include <list>
#include <iostream>
#include <sstream>
#include "maxent.h"
#include "common.h"

#include <sys/types.h>
#include <sys/socket.h>
#include <sys/fcntl.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <sys/un.h>
#include <unistd.h>
#include <iostream>
#include <string.h>
#include <errno.h>
#include <stdlib.h>
#include <signal.h>
#include <ctype.h>


using namespace std;

string bidir_postag(const string & s, const vector<ME_Model> & vme, const vector<ME_Model> & cvme, bool dont_tokenize);
void bidir_chunking(vector<Sentence> & vs, const vector<ME_Model> & vme);
void init_morphdic();


static bool dont_tokenize = false;
static vector<ME_Model> vme(16);
static vector<ME_Model> vme_chunking(16);

static int client_sockfd;
static volatile bool connection_closed = false;

void iohandler(int dummy)
{
  const int BUFLEN = 10000;
  
  char ch;
  char buf[BUFLEN];
  buf[0] = 0;
  if (read(client_sockfd, buf, BUFLEN) == -1) return;
  for (int i = 0; i < BUFLEN-1; i++) {
    if (buf[i] == '\r') {
      //      buf[i] = '\n';
      buf[i] = 0;
      break;
    }
    if (buf[i] == '\n') {
      buf[i] = 0;
      break;
    }
  }
  string line(buf);
  if (line.size() == 0) {
    close(client_sockfd);
    cerr << "empty line. connection closed." << endl;
    connection_closed = true;
    return;
  }

  string postagged = bidir_postag(line, vme, vme_chunking, dont_tokenize);
  postagged += "\n";
  write(client_sockfd, postagged.c_str(), postagged.size());

  cerr << ".";

  /*  
  int len = 0;
  for (int i = 0; i < BUFLEN; i++) {
    if (buf[i] == '\n') break;
    len++;
    buf[i] = toupper(buf[i]);
  }
  //  cout << "len = " << len << endl;
  // note) CR+LF?
  write(client_sockfd, buf, len+1);
  */
}

int wait_for_connection(int port)
{
  int server_sockfd;
  socklen_t server_len, client_len;
  struct sockaddr_in server_address;
  struct sockaddr_in client_address;

  if ((server_sockfd = socket(PF_INET, SOCK_STREAM, 0)) < 0) {
    cerr << "error: socket" << endl;
    exit(1);
  }

  bzero((char *)&server_address, sizeof(server_address));
  server_address.sin_family = AF_INET;
  server_address.sin_addr.s_addr = htonl(INADDR_ANY);
  //  server_address.sin_addr.s_addr = inet_addr("127.0.0.1");
  server_address.sin_port = htons(port);

  int on = 1;
  setsockopt(server_sockfd, SOL_SOCKET, SO_REUSEADDR, (char*) &on, sizeof(on));
  
  server_len = sizeof(server_address);
  if (bind(server_sockfd, (struct sockaddr *)&server_address, server_len) < 0){
    cerr << "error. bind(): " << strerror(errno) << endl;
    exit(1);
  }

  if (listen(server_sockfd, 5) < 0) {
    cerr << "error. listen(): " << strerror(errno) << endl;
    exit(1);
  }

  cerr << "GENIA Tagger server is waiting at port " << port << "." << endl;

  client_len = sizeof(client_address);
  client_sockfd = accept(server_sockfd, (struct sockaddr *)&client_address,
                         &client_len);
  if (client_sockfd == -1) {
    cerr << "error. accept(): " << strerror(errno) << endl;
    exit(1);
  }
  cerr << "connected." << endl;
  close(server_sockfd);

  struct sigaction act;
  act.sa_handler = iohandler;
  sigemptyset(&act.sa_mask);
  act.sa_flags = 0;
  sigaction(SIGIO, &act, NULL);
    
  if (fcntl(client_sockfd, F_SETOWN, getpid()) == -1) {
    cerr << "error. fcntl(): " << strerror(errno) << endl;
    exit(1);
  }
  if (fcntl(client_sockfd, F_SETFL, O_NONBLOCK|O_ASYNC) == -1) {
    cerr << "error. fcntl(): " << strerror(errno) << endl;
    exit(1);
  }

  while(1) {
    if (connection_closed) {
      connection_closed = false;
      return 0;
    }
  }

  close(client_sockfd);
  exit(0);

}


void help()
{
  cout << "Usage: geniatagger [OPTION]... [FILE]..." << endl;
  cout << "Analyze English sentences and print the base forms, part-of-speech tags, and" << endl;
  cout << "chunk tags." << endl;
  cout << endl;
  cout << "Options:" << endl;
  cout << "  --server [PORT]  server mode" << endl;
  cout << "  -nt              don't perform tokenization" << endl;
  cout << "  --help           display this help and exit" << endl;
  cout << endl;
  cout << "Report bugs to <tsuruoka@is.s.u-tokyo.ac.jp>." << endl;
}

void version()
{
  cout << "GENIA Tagger 2.0.1" << endl << endl;
}

int main(int argc, char** argv)
{
  istream *is(&std::cin);

  string ifilename, ofilename;
  int port = -1;
  for (int i = 1; i < argc; i++) {
    string v = argv[i];
    if (v == "-nt") { dont_tokenize = true; continue; }
    if (v == "--server") { port = atoi(argv[i+1]); i++; continue; }
    if (v == "--help") { help(); exit(0); }
    ifilename = argv[i];
  }
  ifstream ifile;
  if (ifilename != "" && ifilename != "-") {
    ifile.open(ifilename.c_str());
    if (!ifile) { cerr << "error: cannot open " << ifilename << endl; exit(1); }
    is = &ifile;
  }
                                                                      
  init_morphdic();

  cerr << "loading pos_models";
  for (int i = 0; i < 16; i++) {
    char buf[1000];
    sprintf(buf, "./models_medline/model.bidir.%d", i);
    vme[i].load_from_file(buf);
    cerr << ".";
  }
  cerr << "done." << endl;

  cerr << "loading chunk_models";
  for (int i = 0; i < 8; i +=2 ) {
    char buf[1000];
    sprintf(buf, "./models_chunking/model.bidir.%d", i);
    vme_chunking[i].load_from_file(buf);
    cerr << ".";
  }
  cerr << "done." << endl;

  if (port > 0) {
    while (1) {
      wait_for_connection(port);
    }
    exit(0);
  }
  
  string line;
  while (getline(*is, line)) {
    string postagged = bidir_postag(line, vme, vme_chunking, dont_tokenize);
    cout << postagged << endl;
  }
  
}

/*
 * $Log: main.cpp,v $
 * Revision 1.8  2005/02/16 10:45:39  tsuruoka
 * acl05 submit
 *
 * Revision 1.7  2005/01/10 13:44:36  tsuruoka
 * a
 *
 * Revision 1.6  2004/12/30 10:34:56  tsuruoka
 * add bidir_decode_search
 *
 * Revision 1.5  2004/12/25 10:47:12  tsuruoka
 * add load_tag_dictionary()
 *
 * Revision 1.4  2004/12/25 09:22:57  tsuruoka
 * add make_tag_dictionary()
 *
 * Revision 1.3  2004/12/21 13:54:46  tsuruoka
 * add bidir.cpp
 *
 * Revision 1.2  2004/12/20 12:06:24  tsuruoka
 * change the data
 *
 * Revision 1.1  2004/07/16 13:40:42  tsuruoka
 * init
 *
 */