/**
* @param accession
* @return
* @throws Exception
*/
private Document getUniprotXML(String accession) throws Exception {
int index = accession.lastIndexOf(".");
String key = accession;
if (index != -1) {
key = accession.substring(0, index);
}
StringBuilder sb = new StringBuilder();
File f = null;
if (uniprotDirectoryCache != null && uniprotDirectoryCache.length() > 0) {
f = new File(uniprotDirectoryCache + File.separatorChar + key + ".xml");
if (f.exists()) {
FileReader fr = new FileReader(f);
int size = (int) f.length();
char[] data = new char[size];
fr.read(data);
fr.close();
sb.append(data);
index = sb.indexOf("xmlns="); // strip out name space stuff to make it easier on xpath
if (index != -1) {
int lastIndex = sb.indexOf(">", index);
sb.replace(index, lastIndex, "");
}
}
}
// http://www.uniprot.org/uniprot/?query=SORBIDRAFT_03g027040&format=xml
if (sb.length() == 0) {
String uniprotURL = getUniprotbaseURL() + "/uniprot/?query=" + key + "&format=xml";
// String uniprotURL = getUniprotbaseURL() + "/uniprot/" + key + ".xml";
logger.info("Loading: {}", uniprotURL);
URL uniprot = new URL(uniprotURL);
URLConnection uniprotConnection = uniprot.openConnection();
uniprotConnection.setRequestProperty("User-Agent", "BioJava");
BufferedReader in =
new BufferedReader(new InputStreamReader(uniprotConnection.getInputStream()));
String inputLine;
while ((inputLine = in.readLine()) != null) {
sb.append(inputLine);
}
in.close();
index = sb.indexOf("xmlns="); // strip out name space stuff to make it easier on xpath
if (index != -1) {
int lastIndex = sb.indexOf(">", index);
sb.replace(index, lastIndex, "");
}
if (f != null) {
FileWriter fw = new FileWriter(f);
fw.write(sb.toString());
fw.close();
}
}
logger.info("Load complete");
try {
// logger.debug(sb.toString());
Document document =
XMLHelper.inputStreamToDocument(new ByteArrayInputStream(sb.toString().getBytes()));
return document;
} catch (Exception e) {
logger.error("Exception on xml parse of: {}", sb.toString());
}
return null;
}
