2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
7 * Jalview is free software: you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation, either version 3
10 * of the License, or (at your option) any later version.
12 * Jalview is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty
14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 * PURPOSE. See the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19 * The Jalview Authors are detailed in the 'AUTHORS' file.
21 package jalview.ext.ensembl;
23 import jalview.io.DataSourceType;
24 import jalview.io.FileParse;
25 import jalview.util.StringUtils;
27 import java.io.BufferedReader;
28 import java.io.DataOutputStream;
29 import java.io.IOException;
30 import java.io.InputStream;
31 import java.io.InputStreamReader;
32 import java.net.HttpURLConnection;
33 import java.net.MalformedURLException;
35 import java.util.HashMap;
36 import java.util.List;
39 import javax.ws.rs.HttpMethod;
41 import org.json.simple.JSONArray;
42 import org.json.simple.JSONObject;
43 import org.json.simple.parser.JSONParser;
46 * Base class for Ensembl REST service clients
50 abstract class EnsemblRestClient extends EnsemblSequenceFetcher
52 private static final int DEFAULT_READ_TIMEOUT = 5 * 60 * 1000; // 5 minutes
54 private static final int CONNECT_TIMEOUT_MS = 10 * 1000; // 10 seconds
57 * update these constants when Jalview has been checked / updated for
58 * changes to Ensembl REST API (ref JAL-2105)
59 * @see https://github.com/Ensembl/ensembl-rest/wiki/Change-log
60 * @see http://rest.ensembl.org/info/rest?content-type=application/json
62 private static final String LATEST_ENSEMBLGENOMES_REST_VERSION = "6.0";
64 private static final String LATEST_ENSEMBL_REST_VERSION = "6.1";
66 private static final String REST_CHANGE_LOG = "https://github.com/Ensembl/ensembl-rest/wiki/Change-log";
68 private static Map<String, EnsemblInfo> domainData;
70 // @see https://github.com/Ensembl/ensembl-rest/wiki/Output-formats
71 private static final String PING_URL = "http://rest.ensembl.org/info/ping.json";
73 private final static long AVAILABILITY_RETEST_INTERVAL = 10000L; // 10 seconds
75 private final static long VERSION_RETEST_INTERVAL = 1000L * 3600; // 1 hr
77 protected static final String CONTENT_TYPE_JSON = "?content-type=application/json";
81 domainData = new HashMap<String, EnsemblInfo>();
82 domainData.put(ENSEMBL_REST,
83 new EnsemblInfo(ENSEMBL_REST, LATEST_ENSEMBL_REST_VERSION));
84 domainData.put(ENSEMBL_GENOMES_REST, new EnsemblInfo(
85 ENSEMBL_GENOMES_REST, LATEST_ENSEMBLGENOMES_REST_VERSION));
88 protected volatile boolean inProgress = false;
91 * Default constructor to use rest.ensembl.org
93 public EnsemblRestClient()
99 * Constructor given the target domain to fetch data from
103 public EnsemblRestClient(String d)
109 public boolean queryInProgress()
115 public StringBuffer getRawRecords()
121 * Returns the URL for the client http request
125 * @throws MalformedURLException
127 protected abstract URL getUrl(List<String> ids)
128 throws MalformedURLException;
131 * Returns true if client uses GET method, false if it uses POST
135 protected abstract boolean useGetRequest();
138 * Return the desired value for the Content-Type request header
143 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
145 protected abstract String getRequestMimeType(boolean multipleIds);
148 * Return the desired value for the Accept request header
151 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
153 protected abstract String getResponseMimeType();
156 * Checks Ensembl's REST 'ping' endpoint, and returns true if response
157 * indicates available, else false
159 * @see http://rest.ensembl.org/documentation/info/ping
162 private boolean checkEnsembl()
164 BufferedReader br = null;
167 // note this format works for both ensembl and ensemblgenomes
168 // info/ping.json works for ensembl only (March 2016)
169 URL ping = new URL(getDomain() + "/info/ping" + CONTENT_TYPE_JSON);
172 * expect {"ping":1} if ok
173 * if ping takes more than 2 seconds to respond, treat as if unavailable
175 br = getHttpResponse(ping, null, 2 * 1000);
178 // error reponse status
181 JSONParser jp = new JSONParser();
182 JSONObject val = (JSONObject) jp.parse(br);
183 String pingString = val.get("ping").toString();
184 return pingString != null;
185 } catch (Throwable t)
188 "Error connecting to " + PING_URL + ": " + t.getMessage());
196 } catch (IOException e)
206 * returns a reader to a Fasta response from the Ensembl sequence endpoint
210 * @throws IOException
212 protected FileParse getSequenceReader(List<String> ids) throws IOException
214 URL url = getUrl(ids);
216 BufferedReader reader = getHttpResponse(url, ids);
222 FileParse fp = new FileParse(reader, url.toString(),
228 * Gets a reader to the HTTP response, using the default read timeout of 5
234 * @throws IOException
236 protected BufferedReader getHttpResponse(URL url, List<String> ids)
239 return getHttpResponse(url, ids, DEFAULT_READ_TIMEOUT);
243 * Writes the HTTP request and gets the response as a reader.
247 * written as Json POST body if more than one
251 * @throws IOException
252 * if response code was not 200, or other I/O error
254 protected BufferedReader getHttpResponse(URL url, List<String> ids,
255 int readTimeout) throws IOException
257 // long now = System.currentTimeMillis();
258 HttpURLConnection connection = (HttpURLConnection) url.openConnection();
261 * POST method allows multiple queries in one request; it is supported for
262 * sequence queries, but not for overlap
264 boolean multipleIds = ids != null && ids.size() > 1;
265 connection.setRequestMethod(
266 multipleIds ? HttpMethod.POST : HttpMethod.GET);
267 connection.setRequestProperty("Content-Type",
268 getRequestMimeType(multipleIds));
269 connection.setRequestProperty("Accept", getResponseMimeType());
271 connection.setUseCaches(false);
272 connection.setDoInput(true);
273 connection.setDoOutput(multipleIds);
275 connection.setConnectTimeout(CONNECT_TIMEOUT_MS);
276 connection.setReadTimeout(readTimeout);
280 writePostBody(connection, ids);
283 int responseCode = connection.getResponseCode();
285 if (responseCode != 200)
288 * note: a GET request for an invalid id returns an error code e.g. 415
289 * but POST request returns 200 and an empty Fasta response
291 System.err.println("Response code " + responseCode + " for " + url);
295 InputStream response = connection.getInputStream();
297 // System.out.println(getClass().getName() + " took "
298 // + (System.currentTimeMillis() - now) + "ms to fetch");
300 checkRateLimits(connection);
302 BufferedReader reader = null;
303 reader = new BufferedReader(new InputStreamReader(response, "UTF-8"));
308 * Inspect response headers for any sign of server overload and respect any
309 * 'retry-after' directive
311 * @see https://github.com/Ensembl/ensembl-rest/wiki/Rate-Limits
314 void checkRateLimits(HttpURLConnection connection)
316 // number of requests allowed per time interval:
317 // String limit = connection.getHeaderField("X-RateLimit-Limit");
318 // length of quota time interval in seconds:
319 // String period = connection.getHeaderField("X-RateLimit-Period");
320 // seconds remaining until usage quota is reset:
321 // String reset = connection.getHeaderField("X-RateLimit-Reset");
322 // number of requests remaining from quota for current period:
323 // String remaining = connection.getHeaderField("X-RateLimit-Remaining");
324 // number of seconds to wait before retrying (if remaining == 0)
325 String retryDelay = connection.getHeaderField("Retry-After");
330 EnsemblInfo info = domainData.get(getDomain());
331 if (retryDelay != null)
333 System.err.println("Ensembl REST service rate limit exceeded, wait "
334 + retryDelay + " seconds before retrying");
337 info.retryAfter = System.currentTimeMillis()
338 + (1000 * Integer.valueOf(retryDelay));
339 } catch (NumberFormatException e)
342 .println("Unexpected value for Retry-After: " + retryDelay);
349 // System.out.println(String.format(
350 // "%s Ensembl requests remaining of %s (reset in %ss)",
351 // remaining, limit, reset));
356 * Rechecks if Ensembl is responding, unless the last check was successful and
357 * the retest interval has not yet elapsed. Returns true if Ensembl is up,
358 * else false. Also retrieves and saves the current version of Ensembl data
359 * and REST services at intervals.
363 protected boolean isEnsemblAvailable()
365 EnsemblInfo info = domainData.get(getDomain());
367 long now = System.currentTimeMillis();
370 * check if we are waiting for 'Retry-After' to expire
372 if (info.retryAfter > now)
374 System.err.println("Still " + (1 + (info.retryAfter - now) / 1000)
375 + " secs to wait before retrying Ensembl");
384 * recheck if Ensembl is up if it was down, or the recheck period has elapsed
386 boolean retestAvailability = (now
387 - info.lastAvailableCheckTime) > AVAILABILITY_RETEST_INTERVAL;
388 if (!info.restAvailable || retestAvailability)
390 info.restAvailable = checkEnsembl();
391 info.lastAvailableCheckTime = now;
395 * refetch Ensembl versions if the recheck period has elapsed
397 boolean refetchVersion = (now
398 - info.lastVersionCheckTime) > VERSION_RETEST_INTERVAL;
401 checkEnsemblRestVersion();
402 checkEnsemblDataVersion();
403 info.lastVersionCheckTime = now;
406 return info.restAvailable;
410 * Constructs, writes and flushes the POST body of the request, containing the
411 * query ids in JSON format
415 * @throws IOException
417 protected void writePostBody(HttpURLConnection connection,
418 List<String> ids) throws IOException
421 StringBuilder postBody = new StringBuilder(64);
422 postBody.append("{\"ids\":[");
424 for (String id : ids)
428 postBody.append(",");
431 postBody.append("\"");
432 postBody.append(id.trim());
433 postBody.append("\"");
435 postBody.append("]}");
436 byte[] thepostbody = postBody.toString().getBytes();
437 connection.setRequestProperty("Content-Length",
438 Integer.toString(thepostbody.length));
439 DataOutputStream wr = new DataOutputStream(
440 connection.getOutputStream());
441 wr.write(thepostbody);
447 * Fetches and checks Ensembl's REST version number
451 private void checkEnsemblRestVersion()
453 EnsemblInfo info = domainData.get(getDomain());
455 JSONParser jp = new JSONParser();
459 url = new URL(getDomain() + "/info/rest" + CONTENT_TYPE_JSON);
460 BufferedReader br = getHttpResponse(url, null);
465 JSONObject val = (JSONObject) jp.parse(br);
466 String version = val.get("release").toString();
467 String majorVersion = version.substring(0, version.indexOf("."));
468 String expected = info.expectedRestVersion;
469 String expectedMajorVersion = expected.substring(0,
470 expected.indexOf("."));
471 info.restMajorVersionMismatch = false;
475 * if actual REST major version is ahead of what we expect,
476 * record this in case we want to warn the user
478 if (Float.valueOf(majorVersion) > Float
479 .valueOf(expectedMajorVersion))
481 info.restMajorVersionMismatch = true;
483 } catch (NumberFormatException e)
485 System.err.println("Error in REST version: " + e.toString());
489 * check if REST version is later than what Jalview has tested against,
490 * if so warn; we don't worry if it is earlier (this indicates Jalview has
491 * been tested in advance against the next pending REST version)
493 boolean laterVersion = StringUtils.compareVersions(version,
497 System.err.println(String.format(
498 "EnsemblRestClient expected %s REST version %s but found %s, see %s",
499 getDbSource(), expected, version, REST_CHANGE_LOG));
501 info.restVersion = version;
502 } catch (Throwable t)
505 "Error checking Ensembl REST version: " + t.getMessage());
509 public boolean isRestMajorVersionMismatch()
511 return domainData.get(getDomain()).restMajorVersionMismatch;
515 * Fetches and checks Ensembl's data version number
519 private void checkEnsemblDataVersion()
521 JSONParser jp = new JSONParser();
525 url = new URL(getDomain() + "/info/data" + CONTENT_TYPE_JSON);
526 BufferedReader br = getHttpResponse(url, null);
529 JSONObject val = (JSONObject) jp.parse(br);
530 JSONArray versions = (JSONArray) val.get("releases");
531 domainData.get(getDomain()).dataVersion = versions.get(0)
534 } catch (Throwable t)
537 "Error checking Ensembl data version: " + t.getMessage());
541 public String getEnsemblDataVersion()
543 return domainData.get(getDomain()).dataVersion;
547 public String getDbVersion()
549 return getEnsemblDataVersion();