cURL / Mailing Lists / curl-library / Single Mail

curl-library

Re: [OT] Simple HTML parser needed

From: Jeff Pohlmeyer <yetanothergeek_at_yahoo.com>
Date: Wed, 2 Feb 2005 05:46:21 -0800 (PST)

> I know plenty of XML parsers, but none of them seems to like
> digesting the typical broken HTML found on many web pages.

LibTidy will swallow just about anything you can throw at it:
  http://tidy.sourceforge.net

> If anyone has an example source code of HTML parsing data
> retrieved with libcurl ...

Here is a crude example using libtidy:

#include <stdio.h>
#include <tidy/tidy.h>
#include <tidy/buffio.h>
#include <curl/curl.h>

/* curl write callback, to fill tidy's input buffer... */
uint write_cb(char *in, uint size, uint nmemb, TidyBuffer *out) {
  uint r;
  r = size * nmemb;
  tidyBufAppend( out, in, r );
  return(r);
}

/* Traverse the document tree */
void dumpNode(TidyDoc doc, TidyNode tnod, int indent ) {
  TidyNode child;
  for ( child = tidyGetChild(tnod); child; child = tidyGetNext(child) )
  {
    ctmbstr name = tidyNodeGetName( child );
    if ( name )
    {
      /* if it has a name, then it's an HTML tag ... */
      TidyAttr attr;
      printf( "%*.*s%s ", indent, indent, "<", name);
      /* walk the attribute list */
      for ( attr=tidyAttrFirst(child); attr; attr=tidyAttrNext(attr) ) {
        printf(tidyAttrName(attr));
        tidyAttrValue(attr)?printf("=\"%s\" ", tidyAttrValue(attr)):printf(" ");
      }
      printf( ">\n");
    } else {
      /* if it doesn't have a name, then it's probably text, cdata, etc... */
      TidyBuffer buf;
      tidyBufInit(&buf);
      tidyNodeGetText(doc, child, &buf);
      printf("%*.*s\n", indent, indent, buf.bp?(char *)buf.bp:"");
      tidyBufFree(&buf);
    }
    dumpNode( doc, child, indent + 4 ); /* recursive */
  }
}

int main(int argc, char **argv ) {
  CURL *curl;
  char curl_errbuf[CURL_ERROR_SIZE];
  TidyDoc tdoc;
  TidyBuffer docbuf = {0};
  TidyBuffer tidy_errbuf = {0};
  int err;
  if ( argc == 2) {
    curl = curl_easy_init();
    curl_easy_setopt(curl, CURLOPT_URL, argv[1]);
    curl_easy_setopt(curl, CURLOPT_ERRORBUFFER, curl_errbuf);
    curl_easy_setopt(curl, CURLOPT_NOPROGRESS, no);
    curl_easy_setopt(curl, CURLOPT_VERBOSE, yes);
    curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, write_cb);

    tdoc = tidyCreate();
    tidyOptSetBool(tdoc, TidyForceOutput, yes); /* try harder */
    tidyOptSetInt(tdoc, TidyWrapLen, 4096);
    tidySetErrorBuffer( tdoc, &tidy_errbuf );
    tidyBufInit(&docbuf);

    curl_easy_setopt(curl, CURLOPT_WRITEDATA, &docbuf);
    err=curl_easy_perform(curl);
    if ( !err ) {
      err = tidyParseBuffer(tdoc, &docbuf); /* parse the input */
      if ( err >= 0 ) {
        err = tidyCleanAndRepair(tdoc); /* fix any problems */
        if ( err >= 0 ) {
          err = tidyRunDiagnostics(tdoc); /* load tidy error buffer */
          if ( err >= 0 ) {
            dumpNode( tdoc, tidyGetRoot(tdoc), 0 ); /* walk the tree */
            fprintf(stderr, "%s\n", tidy_errbuf.bp); /* show errors */
         }
        }
      }
    } else fprintf(stderr, "%s\n", curl_errbuf);

    /* clean-up */
    curl_easy_cleanup(curl);
    tidyBufFree(&docbuf);
    tidyBufFree(&tidy_errbuf);
    tidyRelease(tdoc);
    return(err);

  } else printf( "usage: %s <url>\n", argv[0] );
  return(0);
}

/* gcc -Wall -I/usr/local/include tidycurl.c -lcurl -ltidy -o tidycurl */

.

__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Received on 2005-02-02