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.bin.Cache;
24 import jalview.javascript.json.JSON;
25 import jalview.util.JSONUtils;
26 import jalview.util.Platform;
27 import jalview.util.StringUtils;
29 import java.io.BufferedReader;
30 import java.io.DataOutputStream;
31 import java.io.IOException;
32 import java.io.InputStream;
33 import java.io.InputStreamReader;
34 import java.io.Reader;
35 import java.net.HttpURLConnection;
36 import java.net.MalformedURLException;
37 import java.net.ProtocolException;
39 import java.util.HashMap;
40 import java.util.List;
43 import javax.ws.rs.HttpMethod;
45 import org.json.simple.parser.ParseException;
48 * Base class for Ensembl REST service clients
52 abstract class EnsemblRestClient extends EnsemblSequenceFetcher
56 Cache.addJ2SDirectDatabaseCall("http://rest.ensembl");
57 Cache.addJ2SDirectDatabaseCall("https://rest.ensembl");
60 private static final int DEFAULT_READ_TIMEOUT = 5 * 60 * 1000; // 5 minutes
62 private static final int CONNECT_TIMEOUT_MS = 10 * 1000; // 10 seconds
64 private static final int MAX_RETRIES = 3;
66 private static final int HTTP_OK = 200;
68 private static final int HTTP_OVERLOAD = 429;
71 * update these constants when Jalview has been checked / updated for
72 * changes to Ensembl REST API, and updated JAL-3018
73 * @see https://github.com/Ensembl/ensembl-rest/wiki/Change-log
74 * @see http://rest.ensembl.org/info/rest?content-type=application/json
76 private static final String LATEST_ENSEMBLGENOMES_REST_VERSION = "9.0";
78 private static final String LATEST_ENSEMBL_REST_VERSION = "9.0";
80 private static final String REST_CHANGE_LOG = "https://github.com/Ensembl/ensembl-rest/wiki/Change-log";
82 private static Map<String, EnsemblData> domainData;
84 private final static long AVAILABILITY_RETEST_INTERVAL = 10000L; // 10 seconds
86 private final static long VERSION_RETEST_INTERVAL = 1000L * 3600; // 1 hr
88 protected static final String CONTENT_TYPE_JSON = "?content-type=application/json";
92 domainData = new HashMap<>();
93 domainData.put(DEFAULT_ENSEMBL_BASEURL,
94 new EnsemblData(DEFAULT_ENSEMBL_BASEURL, LATEST_ENSEMBL_REST_VERSION));
95 domainData.put(DEFAULT_ENSEMBL_GENOMES_BASEURL, new EnsemblData(
96 DEFAULT_ENSEMBL_GENOMES_BASEURL, LATEST_ENSEMBLGENOMES_REST_VERSION));
99 protected volatile boolean inProgress = false;
102 * Default constructor to use rest.ensembl.org
104 public EnsemblRestClient()
109 * initialise domain info lazily
111 if (!domainData.containsKey(ensemblDomain))
113 domainData.put(ensemblDomain,
114 new EnsemblData(ensemblDomain, LATEST_ENSEMBL_REST_VERSION));
116 if (!domainData.containsKey(ensemblGenomesDomain))
118 domainData.put(ensemblGenomesDomain, new EnsemblData(
119 ensemblGenomesDomain, LATEST_ENSEMBLGENOMES_REST_VERSION));
124 * Constructor given the target domain to fetch data from
128 public EnsemblRestClient(String d)
134 public boolean queryInProgress()
140 public StringBuffer getRawRecords()
146 * Returns the URL for the client http request
150 * @throws MalformedURLException
152 protected abstract URL getUrl(List<String> ids)
153 throws MalformedURLException;
156 * Returns true if client uses GET method, false if it uses POST
160 protected abstract boolean useGetRequest();
163 * Returns the desired value for the Content-Type request header. Default is
164 * application/json, override if required to vary this.
167 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
169 protected String getRequestMimeType()
171 return "application/json";
175 * Return the desired value for the Accept request header. Default is
176 * application/json, override if required to vary this.
179 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
181 protected String getResponseMimeType()
183 return "application/json";
187 * Checks Ensembl's REST 'ping' endpoint, and returns true if response
188 * indicates available, else false
190 * @see http://rest.ensembl.org/documentation/info/ping
193 @SuppressWarnings("unchecked")
194 boolean checkEnsembl()
196 BufferedReader br = null;
197 String pingUrl = getDomain() + "/info/ping" + CONTENT_TYPE_JSON;
200 // note this format works for both ensembl and ensemblgenomes
201 // info/ping.json works for ensembl only (March 2016)
207 * expect {"ping":1} if ok
208 * if ping takes more than 2 seconds to respond, treat as if unavailable
210 Map<String, Object> val = (Map<String, Object>) getJSON(new URL(pingUrl), null, 2 * 1000, MODE_MAP, null);
213 String pingString = val.get("ping").toString();
214 return pingString != null;
215 } catch (Throwable t)
218 "Error connecting to " + pingUrl + ": " + t.getMessage());
226 } catch (IOException e)
236 protected final static int MODE_ARRAY = 0;
237 protected final static int MODE_MAP = 1;
238 protected final static int MODE_ITERATOR = 2;
241 // * Returns a reader to a (Json) response from the Ensembl sequence endpoint.
242 // * If the request failed the return value may be null.
246 // * @throws IOException
247 // * @throws ParseException
249 // protected Object getSequenceJSON(List<String> ids, int mode)
250 // throws IOException, ParseException
252 // URL url = getUrl(ids);
253 // return getJSON(url, ids, -1, mode);
257 // * Gets a reader to the HTTP response, using the default read timeout of 5
263 // * @throws IOException
265 // protected BufferedReader getHttpResponse(URL url, List<String> ids)
266 // throws IOException
268 // return getHttpResponse(url, ids, DEFAULT_READ_TIMEOUT);
272 * Sends the HTTP request and gets the response as a reader. Returns null if
273 * the HTTP response code was not 200.
277 * written as Json POST body if more than one
281 * @throws IOException
283 private BufferedReader getHttpResponse(URL url, List<String> ids,
284 int readTimeout) throws IOException
287 readTimeout = DEFAULT_READ_TIMEOUT;
288 int retriesLeft = MAX_RETRIES;
289 HttpURLConnection connection = null;
290 int responseCode = 0;
292 if (Platform.isJS()) {
296 while (retriesLeft > 0)
298 connection = tryConnection(url, ids, readTimeout);
299 responseCode = connection.getResponseCode();
300 if (responseCode == HTTP_OVERLOAD) // 429
303 checkRetryAfter(connection);
310 if (responseCode != HTTP_OK) // 200
313 * note: a GET request for an invalid id returns an error code e.g. 415
314 * but POST request returns 200 and an empty Fasta response
316 System.err.println("Response code " + responseCode);// + " for " + url);
321 InputStream response = connection.getInputStream();
323 if (Platform.isJS()) {
324 return JSON.getJSONReader(response);
327 // System.out.println(getClass().getName() + " took "
328 // + (System.currentTimeMillis() - now) + "ms to fetch");
330 return new BufferedReader(new InputStreamReader(response, "UTF-8"));
338 * @throws IOException
339 * @throws ProtocolException
341 protected HttpURLConnection tryConnection(URL url, List<String> ids,
342 int readTimeout) throws IOException, ProtocolException
344 // System.out.println(System.currentTimeMillis() + " " + url);
346 HttpURLConnection connection = (HttpURLConnection) url.openConnection();
349 * POST method allows multiple queries in one request; it is supported for
350 * sequence queries, but not for overlap
352 boolean multipleIds = ids != null && ids.size() > 1;
353 connection.setRequestMethod(
354 multipleIds ? HttpMethod.POST : HttpMethod.GET);
355 connection.setRequestProperty("Content-Type", getRequestMimeType());
356 connection.setRequestProperty("Accept", getResponseMimeType());
358 connection.setDoInput(true);
359 connection.setDoOutput(multipleIds);
361 connection.setUseCaches(false);
362 connection.setConnectTimeout(CONNECT_TIMEOUT_MS);
363 connection.setReadTimeout(readTimeout);
367 writePostBody(connection, ids);
373 * Inspects response headers for a 'retry-after' directive, and waits for the
374 * directed period (if less than 10 seconds)
376 * @see https://github.com/Ensembl/ensembl-rest/wiki/Rate-Limits
379 void checkRetryAfter(HttpURLConnection connection)
381 String retryDelay = connection.getHeaderField("Retry-After");
386 if (retryDelay != null)
390 int retrySecs = Integer.valueOf(retryDelay);
391 if (retrySecs > 0 && retrySecs < 10)
394 .println("Ensembl REST service rate limit exceeded, waiting "
395 + retryDelay + " seconds before retrying");
396 Thread.sleep(1000 * retrySecs);
398 } catch (NumberFormatException | InterruptedException e)
400 System.err.println("Error handling Retry-After: " + e.getMessage());
406 * Rechecks if Ensembl is responding, unless the last check was successful and
407 * the retest interval has not yet elapsed. Returns true if Ensembl is up,
408 * else false. Also retrieves and saves the current version of Ensembl data
409 * and REST services at intervals.
413 protected boolean isEnsemblAvailable()
415 EnsemblData info = domainData.get(getDomain());
417 long now = System.currentTimeMillis();
420 * recheck if Ensembl is up if it was down, or the recheck period has elapsed
422 boolean retestAvailability = (now
423 - info.lastAvailableCheckTime) > AVAILABILITY_RETEST_INTERVAL;
424 if (!info.restAvailable || retestAvailability)
426 info.restAvailable = checkEnsembl();
427 info.lastAvailableCheckTime = now;
431 * refetch Ensembl versions if the recheck period has elapsed
433 boolean refetchVersion = (now
434 - info.lastVersionCheckTime) > VERSION_RETEST_INTERVAL;
437 checkEnsemblRestVersion();
438 checkEnsemblDataVersion();
439 info.lastVersionCheckTime = now;
442 return info.restAvailable;
446 * Constructs, writes and flushes the POST body of the request, containing the
447 * query ids in JSON format
451 * @throws IOException
453 protected void writePostBody(HttpURLConnection connection,
454 List<String> ids) throws IOException
457 StringBuilder postBody = new StringBuilder(64);
458 postBody.append("{\"ids\":[");
460 for (int i = 0, n = ids.size(); i < n; i++)
462 String id = ids.get(i);
465 postBody.append(",");
468 postBody.append("\"");
469 postBody.append(id.trim());
470 postBody.append("\"");
472 postBody.append("]}");
473 byte[] thepostbody = postBody.toString().getBytes();
474 connection.setRequestProperty("Content-Length",
475 Integer.toString(thepostbody.length));
476 DataOutputStream wr = new DataOutputStream(
477 connection.getOutputStream());
478 wr.write(thepostbody);
484 * Primary access point to parsed JSON data, including the call to retrieve and
487 * @param url request url; if null, getUrl(ids) will be used
488 * @param ids optional; may be null
489 * @param msDelay -1 for default delay
490 * @param mode map, array, or array iterator
491 * @param mapKey an optional key for an outer map
492 * @return a Map, List, Iterator, or null
493 * @throws IOException
494 * @throws ParseException
496 * @author Bob Hanson 2019
498 @SuppressWarnings("unchecked")
499 protected Object getJSON(URL url, List<String> ids, int msDelay, int mode, String mapKey) throws IOException, ParseException {
507 Platform.timeCheck(null, Platform.TIME_MARK);
509 br = (url == null ? null : getHttpResponse(url, ids, msDelay));
511 Object ret = (br == null ? null : JSONUtils.parse(br));
513 Platform.timeCheck("EnsembleRestClient.getJSON " + url, Platform.TIME_MARK);
516 if (ret != null && mapKey != null)
517 ret = ((Map<String, Object>) ret).get(mapKey);
527 ret = ((List<Object>) ret).iterator();
539 } catch (IOException e)
550 * Fetches and checks Ensembl's REST version number
554 @SuppressWarnings("unchecked")
555 private void checkEnsemblRestVersion()
557 EnsemblData info = domainData.get(getDomain());
561 Map<String, Object> val = (Map<String, Object>) getJSON(new URL(getDomain() + "/info/rest" + CONTENT_TYPE_JSON), null, -1, MODE_MAP, null);
564 String version = val.get("release").toString();
565 String majorVersion = version.substring(0, version.indexOf("."));
566 String expected = info.expectedRestVersion;
567 String expectedMajorVersion = expected.substring(0,
568 expected.indexOf("."));
569 info.restMajorVersionMismatch = false;
573 * if actual REST major version is ahead of what we expect,
574 * record this in case we want to warn the user
576 if (Float.valueOf(majorVersion) > Float
577 .valueOf(expectedMajorVersion))
579 info.restMajorVersionMismatch = true;
581 } catch (NumberFormatException e)
583 System.err.println("Error in REST version: " + e.toString());
587 * check if REST version is later than what Jalview has tested against,
588 * if so warn; we don't worry if it is earlier (this indicates Jalview has
589 * been tested in advance against the next pending REST version)
591 boolean laterVersion = StringUtils.compareVersions(version,
595 System.err.println(String.format(
596 "EnsemblRestClient expected %s REST version %s but found %s, see %s",
597 getDbSource(), expected, version, REST_CHANGE_LOG));
599 info.restVersion = version;
600 } catch (Throwable t)
603 "Error checking Ensembl REST version: " + t.getMessage());
607 public boolean isRestMajorVersionMismatch()
609 return domainData.get(getDomain()).restMajorVersionMismatch;
613 * Fetches and checks Ensembl's data version number
617 @SuppressWarnings("unchecked")
618 private void checkEnsemblDataVersion()
620 Map<String, Object> val;
623 val = (Map<String, Object>) getJSON(
624 new URL(getDomain() + "/info/data" + CONTENT_TYPE_JSON), null, -1, MODE_MAP, null);
627 List<Object> versions = (List<Object>) val.get("releases");
628 domainData.get(getDomain()).dataVersion = versions.get(0).toString();
629 } catch (Throwable e) {//could be IOException | ParseException e) {
630 System.err.println("Error checking Ensembl data version: " + e.getMessage());
634 public String getEnsemblDataVersion()
636 return domainData.get(getDomain()).dataVersion;
640 public String getDbVersion()
642 return getEnsemblDataVersion();