newsenum.c

harvest是一个下载html网页得机器人

static char rcsid[] = "$Id: newsenum.c,v 2.1 1997/03/21 19:21:39 sxw Exp $";
/*
 *  newsenum.c - Enumerates news URLs
 *
 *  DEBUG: none
 *  AUTHOR: Harvest derived
 *
 *  Harvest Indexer http://harvest.sourceforge.net/
 *  -----------------------------------------------
 *
 *  The Harvest Indexer is a continued development of code developed by
 *  the Harvest Project. Development is carried out by numerous individuals
 *  in the Internet community, and is not officially connected with the
 *  original Harvest Project or its funding sources.
 *
 *  Please mail [e-mail] if you are interested in participating
 *  in the development effort.
 *
 *  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.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

/*  Copyright (c) 1994, 1995.  All rights reserved.
 *
 *    The Harvest software was developed by the Internet Research Task
 *    Force Research Group on Resource Discovery (IRTF-RD):
 *
 *          Mic Bowman of Transarc Corporation.
 *          Peter Danzig of the University of Southern California.
 *          Darren R. Hardy of the University of Colorado at Boulder.
 *          Udi Manber of the University of Arizona.
 *          Michael F. Schwartz of the University of Colorado at Boulder.
 *          Duane Wessels of the University of Colorado at Boulder.
 *
 *    This copyright notice applies to software in the Harvest
 *    ``src/'' directory only.  Users should consult the individual
 *    copyright notices in the ``components/'' subdirectories for
 *    copyright information about other software bundled with the
 *    Harvest source code distribution.
 *
 *  TERMS OF USE
 *
 *    The Harvest software may be used and re-distributed without
 *    charge, provided that the software origin and research team are
 *    cited in any use of the system.  Most commonly this is
 *    accomplished by including a link to the Harvest Home Page
 *    (http://harvest.cs.colorado.edu/) from the query page of any
 *    Broker you deploy, as well as in the query result pages.  These
 *    links are generated automatically by the standard Broker
 *    software distribution.
 *
 *    The Harvest software is provided ``as is'', without express or
 *    implied warranty, and with no support nor obligation to assist
 *    in its use, correction, modification or enhancement.  We assume
 *    no liability with respect to the infringement of copyrights,
 *    trade secrets, or any patents, and are not responsible for
 *    consequential damages.  Proper use of the Harvest software is
 *    entirely the responsibility of the user.
 *
 *  DERIVATIVE WORKS
 *
 *    Users may make derivative works from the Harvest software, subject
 *    to the following constraints:
 *
 *      - You must include the above copyright notice and these
 *        accompanying paragraphs in all forms of derivative works,
 *        and any documentation and other materials related to such
 *        distribution and use acknowledge that the software was
 *        developed at the above institutions.
 *
 *      - You must notify IRTF-RD regarding your distribution of
 *        the derivative work.
 *
 *      - You must clearly notify users that your are distributing
 *        a modified version and not the original Harvest software.
 *
 *      - Any derivative product is also subject to these copyright
 *        and use restrictions.
 *
 *    Note that the Harvest software is NOT in the public domain.  We
 *    retain copyright, as specified above.
 *
 *  HISTORY OF FREE SOFTWARE STATUS
 *
 *    Originally we required sites to license the software in cases
 *    where they were going to build commercial products/services
 *    around Harvest.  In June 1995 we changed this policy.  We now
 *    allow people to use the core Harvest software (the code found in
 *    the Harvest ``src/'' directory) for free.  We made this change
 *    in the interest of encouraging the widest possible deployment of
 *    the technology.  The Harvest software is really a reference
 *    implementation of a set of protocols and formats, some of which
 *    we intend to standardize.  We encourage commercial
 *    re-implementations of code complying to this set of standards.
 */
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <unistd.h>
#include <ctype.h>
#include <gdbm.h>
#include "util.h"
#include "url.h"

/* Local variables */
static char *url_file = NULL;
static GDBM_FILE urldbf = NULL;
static int url_max = 250, nurls = 0;

/* Local functions */
static void process_url();
static void usage();
static void cleanup();
static int url_in_db();
static void mark_retrieved();
static Buffer *news_enum();

/*
 *  mark_retrieved() - Mark that the given URL was successfully retrieved,
 *  so that the URL is not retrieved again.  This prevents cycles in the
 *  enumeration.
 */
static void mark_retrieved(up)
     URL *up;
{
    datum k, d;

    k.dptr = xstrdup(up->url);
    k.dsize = strlen(k.dptr) + 1;
    d.dptr = xstrdup(up->md5);
    d.dsize = strlen(d.dptr) + 1;
    if (gdbm_store(urldbf, k, d, GDBM_REPLACE)) {
	fatal("gdbm_store: %s: urldb: %s\n", k.dptr,
	    gdbm_strerror(gdbm_errno));
    }
    xfree(k.dptr);
    xfree(d.dptr);
#ifdef DEBUG
    fprintf(stderr, "ADDING: %s\n", up->url);
    fprintf(stderr, "ADDING: %s\n", up->md5);
#endif

    if (up->type == URL_NEWS) {
	fprintf(stdout, "%s\t%s\n", up->url, up->md5);
	fflush(stdout);
	nurls++;
    } else {
	fprintf(stderr, "!!! URL=%s  TYPE=%d\n", up->url, up->type);
    }
}

/*
 *  url_in_db() - check to see if the URL is in the database
 */
static int url_in_db(url)
     char *url;
{
    datum k;
    int r;

    k.dptr = xstrdup(url);
    k.dsize = strlen(k.dptr) + 1;
    r = gdbm_exists(urldbf, k);
    xfree(k.dptr);
    return (r);
}

/*
 *  news_enum() - Returns all of the URLs that News file at url.  The
 *  buffer that is returned has the URLs separated by \n's.  Returns
 *  NULL on error.
 */
static Buffer *news_enum(up)
     URL *up;
{
    FILE *fp = NULL;
    char *news_path = NULL;
    char *p = NULL;
    char *q = NULL;
    int i;
    static Buffer *b = NULL;
    static char buf[BUFSIZ];
    static char newurl[BUFSIZ];

    /* Verify that we haven't seen this before */
    if (url_in_db(up->url)) {
#ifdef DEBUG
	fprintf(stderr, "EXISTS (url): %s\n", up->url);
#endif
	return (NULL);
    }
#ifdef DEBUG
    fprintf(stderr, "RETRIEVING: %s\n", up->url);
#endif
    if (url_retrieve(up)) {
#ifdef DEBUG
	fprintf(stderr, "RETRIEVING FAILED: %s\n", up->url);
#endif
	return (NULL);
    }
    /* Remember that we've been here before */
    if (up->md5)
	mark_retrieved(up);

    /*
     *  For each pointer, convert it to a URL, and add it to
     *  the list of URLs to return.
     */
    b = create_buffer(BUFSIZ);
    if ((fp = fopen(up->filename, "r")) == NULL) {
	log_errno(up->filename);
	return (NULL);
    }
    while (fgets(buf, BUFSIZ, fp)) {
#ifdef DEBUG
	fprintf(stderr, "LINE: %s", buf);
#endif

	/* Here, we are parsing the output of an NNTP server XOVER command  */
	/* Each line is one message; fields separated by tabs:              */
	/* msgnum   subj    from    date    msgid   ??id n1   n2    Xref:   */

	/* NOTE: the msgid field is surrounded by brackets: <[e-mail]> */
	/* But a news URL looks like this:  news:[e-mail]              */

	for (q = buf, i = 0; i < 4; i++)
	    if (q)
		q = strchr(q + 1, '\t');
	if (!q)
	    continue;
	q += 2;			/* skip tab and < */
	p = strchr(q, '>');
	if (!p)
	    continue;
	*p = 0;
	news_path = xstrdup(q);

#ifdef DEBUG
	Log("News Path: %s\n", news_path);
#endif

	sprintf(newurl, "news:%s\n", news_path);
	add_buffer(b, newurl, strlen(newurl));
	if (news_path)
	    xfree(news_path);
    }
    fclose(fp);
    return (b);
}

/*
 *  process_url() - Retrieves the given URL (News only), computes an MD5,
 *  and extracts the list of further News pointers within the document.
 *  If any of the HREF pointers are News and the same host as the given
 *  URL up, then it calls process_url() recursively.
 *
 */
static void process_url(up)
     URL *up;
{
    Buffer *b;
    char buf[BUFSIZ], *p;
    URL *tup;

#ifdef DEBUG
    fprintf(stderr, "DOING: %s\n", up->url);
#endif
    if ((b = news_enum(up)) == NULL)
	return;
    add_buffer(b, "\0", 1);

    /*
     *  Now, for each URL in the buffer, call process_url() if
     *  the URL is an News url and it is on the same host
     */
    p = b->data;
    while (sscanf(p, "%s\n", buf) == 1) {
	p = strchr(p, '\n') + 1;
#ifdef DEBUG
	fprintf(stderr, "PROCESSING: %s\n", buf);
#endif
	if ((tup = url_open(buf)) == NULL)
	    continue;
	/* If it isn't a news url, or the servers are differnet */
	if ((tup->type != URL_NEWS)) {
	    url_close(tup);
	    continue;
	}
	if (url_retrieve(tup)) {
#ifdef DEBUG
	    fprintf(stderr, "RETRIEVING FAILED: %s\n", tup->url);
#endif
	    url_close(tup);
	    continue;
	}
	mark_retrieved(tup);
	url_close(tup);
	if (++nurls > url_max) {
	    Log("Truncating RootNode URL at %d LeafNode URLs.\n",
		url_max);
	    break;
	}
    }
    free_buffer(b);
}

static void usage()
{
    fprintf(stderr, "Usage: newsenum news-URL\n");
    exit(1);
}

int main(argc, argv)
     int argc;
     char **argv;
{
    char *s;
    URL *up;
    FILE *logfp = NULL;

    if (getenv("HARVEST_GATHERER_LOGFILE") != (char *) NULL)
	logfp = fopen(getenv("HARVEST_GATHERER_LOGFILE"), "a+");
    if (logfp == (FILE *) NULL)
	logfp = stderr;
    init_log3("newsenum", logfp, stderr);
    if (argc != 2)
	usage();
    debug_init();

    if ((s = getenv("HARVEST_URL_MAX")) != NULL)
	url_max = atoi(s);

#ifdef DEBUG
    Log("Running News enumeration: %s\n", argv[1]);
#endif

    url_file = xstrdup(tempnam(NULL, "Nurl"));
    urldbf = gdbm_open(url_file, 0, GDBM_NEWDB, 0644, NULL);
    if (urldbf == NULL) {
	errorlog("gdbm_open: %s: %s\n", url_file,
	    gdbm_strerror(gdbm_errno));
	log_errno(url_file);
	exit(1);
    }
    init_url();
    if ((up = url_open(argv[1])) == NULL || up->type != URL_NEWS) {
	(void) unlink(url_file);
	usage();
    }
    printf("%s\n", up->url);	/* Print tree root */
    process_url(up);
    url_close(up);
    finish_url();
    cleanup();
    exit(0);
}

static void cleanup()
{
    gdbm_close(urldbf);
    (void) unlink(url_file);
}