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 = "5.0";
64 private static final String LATEST_ENSEMBL_REST_VERSION = "5.0";
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 private static final Regex PROTEIN_REGEX = new Regex(
78 "(ENS)([A-Z]{3}|)P[0-9]{11}$");
80 private static final Regex TRANSCRIPT_REGEX = new Regex(
81 "(ENS)([A-Z]{3}|)T[0-9]{11}$");
83 private static final Regex GENE_REGEX = new Regex(
84 "(ENS)([A-Z]{3}|)G[0-9]{11}$");
88 domainData = new HashMap<String, EnsemblInfo>();
89 domainData.put(ENSEMBL_REST, new EnsemblInfo(ENSEMBL_REST,
90 LATEST_ENSEMBL_REST_VERSION));
91 domainData.put(ENSEMBL_GENOMES_REST, new EnsemblInfo(
92 ENSEMBL_GENOMES_REST, LATEST_ENSEMBLGENOMES_REST_VERSION));
95 protected volatile boolean inProgress = false;
98 * Default constructor to use rest.ensembl.org
100 public EnsemblRestClient()
106 * Constructor given the target domain to fetch data from
110 public EnsemblRestClient(String d)
117 * Answers true if the query matches the regular expression pattern for an
118 * Ensembl transcript stable identifier
123 public boolean isTranscriptIdentifier(String query)
125 return query == null ? false : TRANSCRIPT_REGEX.search(query);
129 * Answers true if the query matches the regular expression pattern for an
130 * Ensembl protein stable identifier
135 public boolean isProteinIdentifier(String query)
137 return query == null ? false : PROTEIN_REGEX.search(query);
141 * Answers true if the query matches the regular expression pattern for an
142 * Ensembl gene stable identifier
147 public boolean isGeneIdentifier(String query)
149 return query == null ? false : GENE_REGEX.search(query);
153 public boolean queryInProgress()
159 public StringBuffer getRawRecords()
165 * Returns the URL for the client http request
169 * @throws MalformedURLException
171 protected abstract URL getUrl(List<String> ids)
172 throws MalformedURLException;
175 * Returns true if client uses GET method, false if it uses POST
179 protected abstract boolean useGetRequest();
182 * Return the desired value for the Content-Type request header
187 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
189 protected abstract String getRequestMimeType(boolean multipleIds);
192 * Return the desired value for the Accept request header
195 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
197 protected abstract String getResponseMimeType();
200 * Checks Ensembl's REST 'ping' endpoint, and returns true if response
201 * indicates available, else false
203 * @see http://rest.ensembl.org/documentation/info/ping
206 private boolean checkEnsembl()
208 BufferedReader br = null;
211 // note this format works for both ensembl and ensemblgenomes
212 // info/ping.json works for ensembl only (March 2016)
213 URL ping = new URL(getDomain()
214 + "/info/ping?content-type=application/json");
217 * expect {"ping":1} if ok
218 * if ping takes more than 2 seconds to respond, treat as if unavailable
220 br = getHttpResponse(ping, null, 2 * 1000);
223 // error reponse status
226 JSONParser jp = new JSONParser();
227 JSONObject val = (JSONObject) jp.parse(br);
228 String pingString = val.get("ping").toString();
229 return pingString != null;
230 } catch (Throwable t)
232 System.err.println("Error connecting to " + PING_URL + ": "
241 } catch (IOException e)
251 * returns a reader to a Fasta response from the Ensembl sequence endpoint
255 * @throws IOException
257 protected FileParse getSequenceReader(List<String> ids)
260 URL url = getUrl(ids);
262 BufferedReader reader = getHttpResponse(url, ids);
268 FileParse fp = new FileParse(reader, url.toString(), DataSourceType.URL);
273 * Gets a reader to the HTTP response, using the default read timeout of 5
279 * @throws IOException
281 protected BufferedReader getHttpResponse(URL url, List<String> ids)
284 return getHttpResponse(url, ids, DEFAULT_READ_TIMEOUT);
288 * Writes the HTTP request and gets the response as a reader.
292 * written as Json POST body if more than one
296 * @throws IOException
297 * if response code was not 200, or other I/O error
299 protected BufferedReader getHttpResponse(URL url, List<String> ids,
300 int readTimeout) throws IOException
302 // long now = System.currentTimeMillis();
303 HttpURLConnection connection = (HttpURLConnection) url.openConnection();
306 * POST method allows multiple queries in one request; it is supported for
307 * sequence queries, but not for overlap
309 boolean multipleIds = ids != null && ids.size() > 1;
310 connection.setRequestMethod(multipleIds ? HttpMethod.POST
312 connection.setRequestProperty("Content-Type",
313 getRequestMimeType(multipleIds));
314 connection.setRequestProperty("Accept", getResponseMimeType());
316 connection.setUseCaches(false);
317 connection.setDoInput(true);
318 connection.setDoOutput(multipleIds);
320 connection.setConnectTimeout(CONNECT_TIMEOUT_MS);
321 connection.setReadTimeout(readTimeout);
325 writePostBody(connection, ids);
328 int responseCode = connection.getResponseCode();
330 if (responseCode != 200)
333 * note: a GET request for an invalid id returns an error code e.g. 415
334 * but POST request returns 200 and an empty Fasta response
336 System.err.println("Response code " + responseCode + " for " + url);
340 InputStream response = connection.getInputStream();
342 // System.out.println(getClass().getName() + " took "
343 // + (System.currentTimeMillis() - now) + "ms to fetch");
345 checkRateLimits(connection);
347 BufferedReader reader = null;
348 reader = new BufferedReader(new InputStreamReader(response, "UTF-8"));
353 * Inspect response headers for any sign of server overload and respect any
354 * 'retry-after' directive
356 * @see https://github.com/Ensembl/ensembl-rest/wiki/Rate-Limits
359 void checkRateLimits(HttpURLConnection connection)
361 // number of requests allowed per time interval:
362 // String limit = connection.getHeaderField("X-RateLimit-Limit");
363 // length of quota time interval in seconds:
364 // String period = connection.getHeaderField("X-RateLimit-Period");
365 // seconds remaining until usage quota is reset:
366 // String reset = connection.getHeaderField("X-RateLimit-Reset");
367 // number of requests remaining from quota for current period:
368 // String remaining = connection.getHeaderField("X-RateLimit-Remaining");
369 // number of seconds to wait before retrying (if remaining == 0)
370 String retryDelay = connection.getHeaderField("Retry-After");
375 EnsemblInfo info = domainData.get(getDomain());
376 if (retryDelay != null)
378 System.err.println("Ensembl REST service rate limit exceeded, wait "
379 + retryDelay + " seconds before retrying");
382 info.retryAfter = System.currentTimeMillis()
383 + (1000 * Integer.valueOf(retryDelay));
384 } catch (NumberFormatException e)
386 System.err.println("Unexpected value for Retry-After: "
394 // System.out.println(String.format(
395 // "%s Ensembl requests remaining of %s (reset in %ss)",
396 // remaining, limit, reset));
401 * Rechecks if Ensembl is responding, unless the last check was successful and
402 * the retest interval has not yet elapsed. Returns true if Ensembl is up,
403 * else false. Also retrieves and saves the current version of Ensembl data
404 * and REST services at intervals.
408 protected boolean isEnsemblAvailable()
410 EnsemblInfo info = domainData.get(getDomain());
412 long now = System.currentTimeMillis();
415 * check if we are waiting for 'Retry-After' to expire
417 if (info.retryAfter > now)
419 System.err.println("Still " + (1 + (info.retryAfter - now) / 1000)
420 + " secs to wait before retrying Ensembl");
429 * recheck if Ensembl is up if it was down, or the recheck period has elapsed
431 boolean retestAvailability = (now - info.lastAvailableCheckTime) > AVAILABILITY_RETEST_INTERVAL;
432 if (!info.restAvailable || retestAvailability)
434 info.restAvailable = checkEnsembl();
435 info.lastAvailableCheckTime = now;
439 * refetch Ensembl versions if the recheck period has elapsed
441 boolean refetchVersion = (now - info.lastVersionCheckTime) > VERSION_RETEST_INTERVAL;
444 checkEnsemblRestVersion();
445 checkEnsemblDataVersion();
446 info.lastVersionCheckTime = now;
449 return info.restAvailable;
453 * Constructs, writes and flushes the POST body of the request, containing the
454 * query ids in JSON format
458 * @throws IOException
460 protected void writePostBody(HttpURLConnection connection,
461 List<String> ids) throws IOException
464 StringBuilder postBody = new StringBuilder(64);
465 postBody.append("{\"ids\":[");
467 for (String id : ids)
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(connection.getOutputStream());
483 wr.write(thepostbody);
489 * Fetches and checks Ensembl's REST version number
493 private void checkEnsemblRestVersion()
495 EnsemblInfo info = domainData.get(getDomain());
497 JSONParser jp = new JSONParser();
501 url = new URL(getDomain()
502 + "/info/rest?content-type=application/json");
503 BufferedReader br = getHttpResponse(url, null);
508 JSONObject val = (JSONObject) jp.parse(br);
509 String version = val.get("release").toString();
510 String majorVersion = version.substring(0, version.indexOf("."));
511 String expected = info.expectedRestVersion;
512 String expectedMajorVersion = expected.substring(0,
513 expected.indexOf("."));
514 info.restMajorVersionMismatch = false;
518 * if actual REST major version is ahead of what we expect,
519 * record this in case we want to warn the user
521 if (Float.valueOf(majorVersion) > Float
522 .valueOf(expectedMajorVersion))
524 info.restMajorVersionMismatch = true;
526 } catch (NumberFormatException e)
528 System.err.println("Error in REST version: " + e.toString());
532 * check if REST version is later than what Jalview has tested against,
533 * if so warn; we don't worry if it is earlier (this indicates Jalview has
534 * been tested in advance against the next pending REST version)
536 boolean laterVersion = StringUtils.compareVersions(version, expected) == 1;
541 .format("EnsemblRestClient expected %s REST version %s but found %s, see %s",
542 getDbSource(), expected, version,
545 info.restVersion = version;
546 } catch (Throwable t)
548 System.err.println("Error checking Ensembl REST version: "
553 public boolean isRestMajorVersionMismatch()
555 return domainData.get(getDomain()).restMajorVersionMismatch;
559 * Fetches and checks Ensembl's data version number
563 private void checkEnsemblDataVersion()
565 JSONParser jp = new JSONParser();
569 url = new URL(getDomain()
570 + "/info/data?content-type=application/json");
571 BufferedReader br = getHttpResponse(url, null);
574 JSONObject val = (JSONObject) jp.parse(br);
575 JSONArray versions = (JSONArray) val.get("releases");
576 domainData.get(getDomain()).dataVersion = versions.get(0)
579 } catch (Throwable t)
581 System.err.println("Error checking Ensembl data version: "
586 public String getEnsemblDataVersion()
588 return domainData.get(getDomain()).dataVersion;
592 public String getDbVersion()
594 return getEnsemblDataVersion();