genome.cpp

此代码经过大量使用

/***************************************************************************
                          genome.cpp  -  description
                             -------------------
    begin                : Sat Dec 8 2001
    copyright            : (C) 2001 by Rudiger Koch
    email                : [e-mail]
 ***************************************************************************/

/***************************************************************************
 *                                                                         *
 *   This program is free software; you can redistribute it and/or modify  *
 *   it under the terms of the GNU General Public License as published by  *
 *   the Free Software Foundation; either version 2 of the License, or     *
 *   (at your option) any later version.                                   *
 *                                                                         *
 ***************************************************************************/

#include "genome.h"

#include <netdb.h>
#include <arpa/inet.h>
#include <netinet/in.h>
#include <sys/socket.h>

#include <sys/stat.h>
#include <fcntl.h>
#include <sys/types.h>
#include <unistd.h>
#include <stdlib.h>
#include <stdio.h>
#include <iostream>



Genome::Genome(){
    gHandler = NULL;
}
Genome::~Genome(){
}


void Genome::get(string uri){
    if(uri.substr(0,4) == string("file")){
        string path = uri.substr(5);
        getFile(path);
    }  else if(uri.substr(0,7) == string("http://")){
        getHttp(uri);
    }
}


void Genome::getFile(string path){
	cout << "Loading Genefile " << path << endl;

    int fd;
    char buffer[20];
    string filename = path.substr(path.length()-13);
    geneId = atoi(filename.c_str());
    if((fd = open(path.c_str(),  O_RDONLY)) == -1) throw string("cannot open genefile");
    int amount;
    while((amount = read(fd, buffer, sizeof(buffer)))>0) {
        if(amount <=-1) throw string("read error");

        for(int i=0; i<amount;i++){
            genome += buffer[i];
        }
    }
    if (close(fd) < 0) throw string("closing filedescriptor failed");
}

void Genome::getHttp(string _url){
        struct sockaddr_in  server;
        int msgsock;
        char buffer[100];
        string download;

        URL = _url;
        string hostname = URL.substr(7, URL.find('/', 9) - 7);
        string path     = URL.substr(URL.find('/', 9));

        if ((msgsock = socket( AF_INET, SOCK_STREAM, 0)) < 0)
            throw string("socket error");

        struct hostent *serverh = gethostbyname(hostname.c_str());
        if (!serverh) throw string("cannot resolve hostname ") + hostname;

        memset(&server, 0, sizeof(server));
        server.sin_family = AF_INET;
        memcpy((char *)&server.sin_addr.s_addr, serverh->h_addr, serverh->h_length);
        server.sin_port = htons(80);

        if (connect(msgsock, (struct sockaddr*)(void*)&server, sizeof(server)) < 0)
            throw string("cannot connect to server");

        string req = string("GET ") + path + string (" HTTP/1.0\n\n");
        if (write(msgsock, req.c_str(), req.length()) < 0)
            throw string("cannot send request");

        int amount;
        while((amount = read(msgsock, buffer, sizeof(buffer)))>0) {
            if(amount <=-1){ throw string("error reading from server");}

            for(int i=0; i<amount;i++){
                download += buffer[i];
            }
        }

        const char* idHeader = "Gene-Id:";
        unsigned int pos = download.find(idHeader)+strlen(idHeader);
        if(pos == string::npos) throw "No Gene-Id header";
        string id = download.substr(pos, 10);
        geneId = atoi(id.c_str());
        if(!geneId) throw string("Malformed Gene-Id header. ID = ") + id;

        const char* formatHeader = "Chromosome-format:";
        unsigned int pos1 = download.find(formatHeader);
        if(pos1 == string::npos) throw string("No Format header");
        unsigned int pos2 = download.find("\r\n", pos1);
        string format = download.substr(pos1, pos2-pos1);
        if(!geneId) throw string("Malformed Chromosome-Format header");
        Format(format);

        genome = download.substr(download.find("\r\n\r\n")+4);
        if (close(msgsock) < 0) throw string("cannot close socket");
}


void Genome::submit(int score){
        struct sockaddr_in  server;
        int msgsock;
        char buffer[100];

        if(URL.empty()) throw string("No server URL specified - cannot submit a score");

        if ((msgsock = socket( AF_INET, SOCK_STREAM, 0)) < 0)
            throw string("cannot open socket");

        string hostname = URL.substr(7, URL.find('/', 9) - 7);
        string path     = URL.substr(URL.find('/', 9));

        struct hostent *serverh = gethostbyname(hostname.c_str());
        if (!serverh) throw string("cannot resolve hostname ") + hostname;

        memset(&server, 0, sizeof(server));
        server.sin_family = AF_INET;
        memcpy((char *)&server.sin_addr.s_addr, serverh->h_addr, serverh->h_length);
        server.sin_port = htons(80);

        if (connect(msgsock, (struct sockaddr*)(void*)&server, sizeof(server)) < 0)
            throw string("cannot connect to server");

        char args[50];
        sprintf(args, "?%d&%d HTTP/1.0\n\n", geneId, score);
        string req(string("GET ") + path + string(args));
        if (write(msgsock, req.c_str(), req.length()) < 0)
            throw string("write error");

        int amount;
        while((amount = read(msgsock, buffer, sizeof(buffer)))>0) {
            if(amount <=-1) throw string("read error");
        }
        if (close(msgsock) < 0)
            throw string("cannot close socket");
}


void Genome::setHandler(GenomeHandler *gh){
    gHandler = gh;
}


void Genome::parse(){
    if(!gHandler) throw string("no genome handler registered");
    int pos=0;
    for(unsigned int i=0; i<chromosomes.size(); i++){
        gHandler->startChromosome(chromosomes[i].second,  chromosomes[i].first);
        for(int j=0; j<chromosomes[i].first; j++){
            gHandler->gene(genome.substr(pos, chromosomes[i].second));
            pos += chromosomes[i].second;
        }
    }
    gHandler->finish();
}


void Genome::Format(string fmt){
    unsigned int lbracket = 0;
    unsigned int rbracket, comma;

    chromosomes.clear();
    while ((lbracket = fmt.find('{', lbracket)) != string::npos){
        lbracket++;
        comma = fmt.find(',', lbracket);
        if(comma == string::npos) throw string("error in format string\n");
        rbracket = fmt.find('}', lbracket);
        if(rbracket == string::npos) throw string("error in format string\n");
        pair <int, int> chr;
        chr.first = atoi(fmt.substr(lbracket, comma).c_str());
        chr.second = atoi(fmt.substr(comma+1, rbracket).c_str());
        chromosomes.push_back(chr);
    }
}