Drupal function to extract text using Tika deployed from within Solr

apachesolr_extract_text.module

/**
 * Extracts text using Tika deployed from within Solr. Assumes the following
 * about your Solr config:
 * - ExtractingRequestHandler lives at (solr URL)/extract, not /update/extract
 * - ExtractingRequestHandler is set to extract only (ext.extract.only=true)
 * - ExtractingRequestHandler only returns text within the body tags of the
 *   XHTML response (ext.xpath=/xhtml:html/xhtml:body/descendant:node())
 * 
 * @param $path
 *   string containing path of file to have text extracted
 * 
 * @return string
 *   XHTML string containing text extracted from document
 */

function apachesolr_extract_text($path) {
  $headers = array('Content-type' => 'application/octet-stream');
  $solr_url = 'http://'. variable_get('apachesolr_host', 'localhost') .':'
    . variable_get('apachesolr_port', '8983')
    . variable_get('apachesolr_path', '/solr') . '/extract';
  $rsp = drupal_http_request($solr_url, $headers, 'POST',
    file_get_contents($path));
  if ($rsp->code != 200) {
    $msg = "HTTP %code error posting file %path to Solr server at %solr.";
    $vars = array('%code' => $rsp->code, '%path' => $path, '%solr' => $solr_url);
    watchdog('Apache Solr', $msg, $vars, WATCHDOG_WARNING);
    $return = '';
  } else {
    $xmldata = simplexml_load_string($rsp->data);
    $extract = $xmldata->str;
    $return = str_replace('<?xml version="1.0" encoding="UTF-8"?>', '', $extract);
  }
  return $return;
}