function get_sequence($accession, &$item)
{
$id = find_genbank($accession);
if ($id == 0) {
// We don't have this sequence (but see below)
$url = 'http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=nucleotide&id=' . $accession . '&rettype=gb&retmode=xml';
//echo $url;
$xml = get($url);
//echo $xml;
// Did we get an error?
// Nothing returned
if ($xml == '') {
return 0;
}
//echo "\n\n" . __LINE__ . "\n\n";
// NCBI error (sequence doesn't exist, or might not be released
$dom = new DOMDocument();
$dom->loadXML($xml);
$xpath = new DOMXPath($dom);
$xpath_query = "//Error";
$nodeCollection = $xpath->query($xpath_query);
$ok = true;
foreach ($nodeCollection as $node) {
if ($node->firstChild->nodeValue != '') {
$ok = false;
}
}
if (!$ok) {
return 0;
}
//echo "\n\n" . __LINE__ . "\n\n";
$xml = str_replace('<!DOCTYPE GBSet PUBLIC "-//NCBI//NCBI GBSeq/EN" "http://www.ncbi.nlm.nih.gov/dtd/NCBI_GBSeq.dtd">', '', $xml);
$xp = new XsltProcessor();
$xsl = new DomDocument();
$xsl->load('xsl/gb2JSON.xsl');
$xp->importStylesheet($xsl);
$xml_doc = new DOMDocument();
$xml_doc->loadXML($xml);
$json = $xp->transformToXML($xml_doc);
//echo $json;
$data = json_decode($json);
//print_r($data);
// Handle case where was have this sequnece from EMBL harvesting
$id = set_gi($data->accession, $data->gi);
if ($id == 0) {
// new sequence
gb_postprocess($data);
//print_r($data);
$id = store_genbank($data);
$item = $data;
} else {
// we have this already from EMBL
$json = retrieve_genbank_json($id);
$item = json_decode($json);
}
} else {
//echo 'have it' . "\n";
$json = retrieve_genbank_json($id);
$item = json_decode($json);
}
// CouchDB modification
if (isset($item->source->latitude)) {
// Geometry for sequence
$item->source->geometry = new stdclass();
$item->source->geometry->type = "MultiPoint";
$item->source->geometry->coordinates = array();
$item->source->geometry->coordinates[] = array((double) $item->source->longitude, (double) $item->source->latitude);
} else {
if (isset($item->source->specimen)) {
// Use geomtry from specimen
if (isset($item->source->specimen->latitude)) {
$item->source->geometry = new stdclass();
$item->source->geometry->type = "MultiPoint";
$item->source->geometry->coordinates = array();
$item->source->geometry->coordinates[] = array((double) $item->source->specimen->longitude, (double) $item->source->specimen->latitude);
}
}
}
return $id;
}
// fetch page
$url = 'http://localhost/~rpage/biostor/api.php?page=' . $PageID . '&format=html';
$json = get($url);
//echo $json;
$obj = json_decode($json);
if (isset($obj->html)) {
$page = html_to_page($obj->html);
$hits = array();
if (0) {
$hits = array_merge($hits, find_specimens($page->text));
}
if (1) {
$hits = array_merge($hits, find_points($page->text));
}
if (0) {
$hits = array_merge($hits, find_genbank($page->text));
}
$annotations = annotations_from_hits($uri, $page, $hits);
echo "annotations\n";
print_r($annotations);
foreach ($annotations as $annotation) {
echo json_encode($annotation->data);
$h->add_annotation($annotation->data);
echo "\n";
}
}
$page_count++;
}
?>
