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
57 Cache.addJ2SDirectDatabaseCall("http://rest.ensembl");
58 Cache.addJ2SDirectDatabaseCall("https://rest.ensembl");
61 private static final int DEFAULT_READ_TIMEOUT = 5 * 60 * 1000; // 5 minutes
63 private static final int CONNECT_TIMEOUT_MS = 10 * 1000; // 10 seconds
65 private static final int MAX_RETRIES = 3;
67 private static final int HTTP_OK = 200;
69 private static final int HTTP_OVERLOAD = 429;
72 * update these constants when Jalview has been checked / updated for
73 * changes to Ensembl REST API, and updated JAL-3018
74 * @see https://github.com/Ensembl/ensembl-rest/wiki/Change-log
75 * @see http://rest.ensembl.org/info/rest?content-type=application/json
77 private static final String LATEST_ENSEMBLGENOMES_REST_VERSION = "9.0";
79 private static final String LATEST_ENSEMBL_REST_VERSION = "9.0";
81 private static final String REST_CHANGE_LOG = "https://github.com/Ensembl/ensembl-rest/wiki/Change-log";
83 private static Map<String, EnsemblData> domainData;
85 private final static long AVAILABILITY_RETEST_INTERVAL = 10000L; // 10 seconds
87 private final static long VERSION_RETEST_INTERVAL = 1000L * 3600; // 1 hr
89 protected static final String CONTENT_TYPE_JSON = "?content-type=application/json";
93 domainData = new HashMap<>();
94 domainData.put(DEFAULT_ENSEMBL_BASEURL, new EnsemblData(
95 DEFAULT_ENSEMBL_BASEURL, LATEST_ENSEMBL_REST_VERSION));
96 domainData.put(DEFAULT_ENSEMBL_GENOMES_BASEURL,
97 new EnsemblData(DEFAULT_ENSEMBL_GENOMES_BASEURL,
98 LATEST_ENSEMBLGENOMES_REST_VERSION));
101 protected volatile boolean inProgress = false;
104 * Default constructor to use rest.ensembl.org
106 public EnsemblRestClient()
111 * initialise domain info lazily
113 if (!domainData.containsKey(ensemblDomain))
115 domainData.put(ensemblDomain,
116 new EnsemblData(ensemblDomain, LATEST_ENSEMBL_REST_VERSION));
118 if (!domainData.containsKey(ensemblGenomesDomain))
120 domainData.put(ensemblGenomesDomain, new EnsemblData(
121 ensemblGenomesDomain, LATEST_ENSEMBLGENOMES_REST_VERSION));
126 * Constructor given the target domain to fetch data from
130 public EnsemblRestClient(String d)
136 public boolean queryInProgress()
142 public StringBuffer getRawRecords()
148 * Returns the URL for the client http request
152 * @throws MalformedURLException
154 protected abstract URL getUrl(List<String> ids)
155 throws MalformedURLException;
158 * Returns true if client uses GET method, false if it uses POST
162 protected abstract boolean useGetRequest();
165 * Returns the desired value for the Content-Type request header. Default is
166 * application/json, override if required to vary this.
169 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
171 protected String getRequestMimeType()
173 return "application/json";
177 * Return the desired value for the Accept request header. Default is
178 * application/json, override if required to vary this.
181 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
183 protected String getResponseMimeType()
185 return "application/json";
189 * Checks Ensembl's REST 'ping' endpoint, and returns true if response
190 * indicates available, else false
192 * @see http://rest.ensembl.org/documentation/info/ping
195 @SuppressWarnings("unchecked")
196 boolean checkEnsembl()
198 BufferedReader br = null;
199 String pingUrl = getDomain() + "/info/ping" + CONTENT_TYPE_JSON;
202 // note this format works for both ensembl and ensemblgenomes
203 // info/ping.json works for ensembl only (March 2016)
206 * expect {"ping":1} if ok
207 * if ping takes more than 2 seconds to respond, treat as if unavailable
209 Map<String, Object> val = (Map<String, Object>) getJSON(
210 new URL(pingUrl), null, 2 * 1000, MODE_MAP, null);
215 String pingString = val.get("ping").toString();
216 return pingString != null;
217 } catch (Throwable t)
220 "Error connecting to " + pingUrl + ": " + t.getMessage());
228 } catch (IOException e)
237 protected final static int MODE_ARRAY = 0;
239 protected final static int MODE_MAP = 1;
241 protected final static int MODE_ITERATOR = 2;
244 // * Returns a reader to a (Json) response from the Ensembl sequence endpoint.
245 // * If the request failed the return value may be null.
249 // * @throws IOException
250 // * @throws ParseException
252 // protected Object getSequenceJSON(List<String> ids, int mode)
253 // throws IOException, ParseException
255 // URL url = getUrl(ids);
256 // return getJSON(url, ids, -1, mode);
260 // * Gets a reader to the HTTP response, using the default read timeout of 5
266 // * @throws IOException
268 // protected BufferedReader getHttpResponse(URL url, List<String> ids)
269 // throws IOException
271 // return getHttpResponse(url, ids, DEFAULT_READ_TIMEOUT);
275 * Sends the HTTP request and gets the response as a reader. Returns null if
276 * the HTTP response code was not 200.
280 * written as Json POST body if more than one
284 * @throws IOException
286 private BufferedReader getHttpResponse(URL url, List<String> ids,
287 int readTimeout) throws IOException
291 readTimeout = DEFAULT_READ_TIMEOUT;
293 int retriesLeft = MAX_RETRIES;
294 HttpURLConnection connection = null;
295 int responseCode = 0;
302 while (retriesLeft > 0)
304 connection = tryConnection(url, ids, readTimeout);
305 responseCode = connection.getResponseCode();
306 if (responseCode == HTTP_OVERLOAD) // 429
309 checkRetryAfter(connection);
316 if (responseCode != HTTP_OK) // 200
319 * note: a GET request for an invalid id returns an error code e.g. 415
320 * but POST request returns 200 and an empty Fasta response
322 System.err.println("Response code " + responseCode);// + " for " + url);
326 InputStream response = connection.getInputStream();
330 return JSON.getJSONReader(response);
333 // System.out.println(getClass().getName() + " took "
334 // + (System.currentTimeMillis() - now) + "ms to fetch");
336 return new BufferedReader(new InputStreamReader(response, "UTF-8"));
344 * @throws IOException
345 * @throws ProtocolException
347 protected HttpURLConnection tryConnection(URL url, List<String> ids,
348 int readTimeout) throws IOException, ProtocolException
350 // System.out.println(System.currentTimeMillis() + " " + url);
352 HttpURLConnection connection = (HttpURLConnection) url.openConnection();
355 * POST method allows multiple queries in one request; it is supported for
356 * sequence queries, but not for overlap
358 boolean multipleIds = ids != null && ids.size() > 1;
359 connection.setRequestMethod(
360 multipleIds ? HttpMethod.POST : HttpMethod.GET);
361 connection.setRequestProperty("Content-Type", getRequestMimeType());
362 connection.setRequestProperty("Accept", getResponseMimeType());
364 connection.setDoInput(true);
365 connection.setDoOutput(multipleIds);
367 connection.setUseCaches(false);
368 connection.setConnectTimeout(CONNECT_TIMEOUT_MS);
369 connection.setReadTimeout(readTimeout);
373 writePostBody(connection, ids);
379 * Inspects response headers for a 'retry-after' directive, and waits for the
380 * directed period (if less than 10 seconds)
382 * @see https://github.com/Ensembl/ensembl-rest/wiki/Rate-Limits
385 void checkRetryAfter(HttpURLConnection connection)
387 String retryDelay = connection.getHeaderField("Retry-After");
392 if (retryDelay != null)
396 int retrySecs = Integer.valueOf(retryDelay);
397 if (retrySecs > 0 && retrySecs < 10)
400 "Ensembl REST service rate limit exceeded, waiting "
401 + retryDelay + " seconds before retrying");
402 Thread.sleep(1000 * retrySecs);
404 } catch (NumberFormatException | InterruptedException e)
406 System.err.println("Error handling Retry-After: " + e.getMessage());
412 * Rechecks if Ensembl is responding, unless the last check was successful and
413 * the retest interval has not yet elapsed. Returns true if Ensembl is up,
414 * else false. Also retrieves and saves the current version of Ensembl data
415 * and REST services at intervals.
419 protected boolean isEnsemblAvailable()
421 EnsemblData info = domainData.get(getDomain());
423 long now = System.currentTimeMillis();
426 * recheck if Ensembl is up if it was down, or the recheck period has elapsed
428 boolean retestAvailability = (now
429 - info.lastAvailableCheckTime) > AVAILABILITY_RETEST_INTERVAL;
430 if (!info.restAvailable || retestAvailability)
432 info.restAvailable = checkEnsembl();
433 info.lastAvailableCheckTime = now;
437 * refetch Ensembl versions if the recheck period has elapsed
439 boolean refetchVersion = (now
440 - info.lastVersionCheckTime) > VERSION_RETEST_INTERVAL;
443 checkEnsemblRestVersion();
444 checkEnsemblDataVersion();
445 info.lastVersionCheckTime = now;
448 return info.restAvailable;
452 * Constructs, writes and flushes the POST body of the request, containing the
453 * query ids in JSON format
457 * @throws IOException
459 protected void writePostBody(HttpURLConnection connection,
460 List<String> ids) throws IOException
463 StringBuilder postBody = new StringBuilder(64);
464 postBody.append("{\"ids\":[");
466 for (int i = 0, n = ids.size(); i < n; i++)
468 String id = ids.get(i);
471 postBody.append(",");
474 postBody.append("\"");
475 postBody.append(id.trim());
476 postBody.append("\"");
478 postBody.append("]}");
479 byte[] thepostbody = postBody.toString().getBytes();
480 connection.setRequestProperty("Content-Length",
481 Integer.toString(thepostbody.length));
482 DataOutputStream wr = new DataOutputStream(
483 connection.getOutputStream());
484 wr.write(thepostbody);
490 * Primary access point to parsed JSON data, including the call to retrieve
494 * request url; if null, getUrl(ids) will be used
496 * optional; may be null
498 * -1 for default delay
500 * map, array, or array iterator
502 * an optional key for an outer map
503 * @return a Map, List, Iterator, or null
504 * @throws IOException
505 * @throws ParseException
507 * @author Bob Hanson 2019
509 @SuppressWarnings("unchecked")
510 protected Object getJSON(URL url, List<String> ids, int msDelay, int mode,
511 String mapKey) throws IOException, ParseException
521 Platform.timeCheck(null, Platform.TIME_MARK);
523 br = (url == null ? null : getHttpResponse(url, ids, msDelay));
525 Object ret = (br == null ? null : JSONUtils.parse(br));
527 Platform.timeCheck("EnsemblRestClient.getJSON " + url,
530 if (ret != null && mapKey != null)
532 ret = ((Map<String, Object>) ret).get(mapKey);
544 ret = ((List<Object>) ret).iterator();
556 } catch (IOException e)
565 * Fetches and checks Ensembl's REST version number
569 @SuppressWarnings("unchecked")
570 private void checkEnsemblRestVersion()
572 EnsemblData info = domainData.get(getDomain());
576 Map<String, Object> val = (Map<String, Object>) getJSON(
577 new URL(getDomain() + "/info/rest" + CONTENT_TYPE_JSON), null,
583 String version = val.get("release").toString();
584 String majorVersion = version.substring(0, version.indexOf("."));
585 String expected = info.expectedRestVersion;
586 String expectedMajorVersion = expected.substring(0,
587 expected.indexOf("."));
588 info.restMajorVersionMismatch = false;
592 * if actual REST major version is ahead of what we expect,
593 * record this in case we want to warn the user
595 if (Float.valueOf(majorVersion) > Float
596 .valueOf(expectedMajorVersion))
598 info.restMajorVersionMismatch = true;
600 } catch (NumberFormatException e)
602 System.err.println("Error in REST version: " + e.toString());
606 * check if REST version is later than what Jalview has tested against,
607 * if so warn; we don't worry if it is earlier (this indicates Jalview has
608 * been tested in advance against the next pending REST version)
610 boolean laterVersion = StringUtils.compareVersions(version,
614 System.err.println(String.format(
615 "EnsemblRestClient expected %s REST version %s but found %s, see %s",
616 getDbSource(), expected, version, REST_CHANGE_LOG));
618 info.restVersion = version;
619 } catch (Throwable t)
622 "Error checking Ensembl REST version: " + t.getMessage());
626 public boolean isRestMajorVersionMismatch()
628 return domainData.get(getDomain()).restMajorVersionMismatch;
632 * Fetches and checks Ensembl's data version number
636 @SuppressWarnings("unchecked")
637 private void checkEnsemblDataVersion()
639 Map<String, Object> val;
642 val = (Map<String, Object>) getJSON(
643 new URL(getDomain() + "/info/data" + CONTENT_TYPE_JSON), null,
649 List<Object> versions = (List<Object>) val.get("releases");
650 domainData.get(getDomain()).dataVersion = versions.get(0).toString();
651 } catch (Throwable e)
652 {// could be IOException | ParseException e) {
654 "Error checking Ensembl data version: " + e.getMessage());
658 public String getEnsemblDataVersion()
660 return domainData.get(getDomain()).dataVersion;
664 public String getDbVersion()
666 return getEnsemblDataVersion();