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;
45 import com.stevesoft.pat.Regex;
48 * Base class for Ensembl REST service clients
52 abstract class EnsemblRestClient extends EnsemblSequenceFetcher
54 private static final int DEFAULT_READ_TIMEOUT = 5 * 60 * 1000; // 5 minutes
56 private static final int CONNECT_TIMEOUT_MS = 10 * 1000; // 10 seconds
59 * update these constants when Jalview has been checked / updated for
60 * changes to Ensembl REST API
61 * @see https://github.com/Ensembl/ensembl-rest/wiki/Change-log
62 * @see http://rest.ensembl.org/info/rest?content-type=application/json
64 private static final String LATEST_ENSEMBLGENOMES_REST_VERSION = "4.8";
66 private static final String LATEST_ENSEMBL_REST_VERSION = "4.8";
68 private static final String REST_CHANGE_LOG = "https://github.com/Ensembl/ensembl-rest/wiki/Change-log";
70 private static Map<String, EnsemblInfo> domainData;
72 // @see https://github.com/Ensembl/ensembl-rest/wiki/Output-formats
73 private static final String PING_URL = "http://rest.ensembl.org/info/ping.json";
75 private final static long AVAILABILITY_RETEST_INTERVAL = 10000L; // 10 seconds
77 private final static long VERSION_RETEST_INTERVAL = 1000L * 3600; // 1 hr
79 private static final Regex TRANSCRIPT_REGEX = new Regex(
80 "(ENS)([A-Z]{3}|)T[0-9]{11}$");
82 private static final Regex GENE_REGEX = new Regex(
83 "(ENS)([A-Z]{3}|)G[0-9]{11}$");
87 domainData = new HashMap<String, EnsemblInfo>();
88 domainData.put(ENSEMBL_REST, new EnsemblInfo(ENSEMBL_REST,
89 LATEST_ENSEMBL_REST_VERSION));
90 domainData.put(ENSEMBL_GENOMES_REST, new EnsemblInfo(
91 ENSEMBL_GENOMES_REST, LATEST_ENSEMBLGENOMES_REST_VERSION));
94 protected volatile boolean inProgress = false;
97 * Default constructor to use rest.ensembl.org
99 public EnsemblRestClient()
105 * Constructor given the target domain to fetch data from
109 public EnsemblRestClient(String d)
115 * Answers true if the query matches the regular expression pattern for an
116 * Ensembl transcript stable identifier
121 public boolean isTranscriptIdentifier(String query)
123 return query == null ? false : TRANSCRIPT_REGEX.search(query);
127 * Answers true if the query matches the regular expression pattern for an
128 * Ensembl gene stable identifier
133 public boolean isGeneIdentifier(String query)
135 return query == null ? false : GENE_REGEX.search(query);
139 public boolean queryInProgress()
145 public StringBuffer getRawRecords()
151 * Returns the URL for the client http request
155 * @throws MalformedURLException
157 protected abstract URL getUrl(List<String> ids)
158 throws MalformedURLException;
161 * Returns true if client uses GET method, false if it uses POST
165 protected abstract boolean useGetRequest();
168 * Return the desired value for the Content-Type request header
173 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
175 protected abstract String getRequestMimeType(boolean multipleIds);
178 * Return the desired value for the Accept request header
181 * @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
183 protected abstract String getResponseMimeType();
186 * Checks Ensembl's REST 'ping' endpoint, and returns true if response
187 * indicates available, else false
189 * @see http://rest.ensembl.org/documentation/info/ping
192 private boolean checkEnsembl()
194 BufferedReader br = null;
197 // note this format works for both ensembl and ensemblgenomes
198 // info/ping.json works for ensembl only (March 2016)
199 URL ping = new URL(getDomain()
200 + "/info/ping?content-type=application/json");
203 * expect {"ping":1} if ok
204 * if ping takes more than 2 seconds to respond, treat as if unavailable
206 br = getHttpResponse(ping, null, 2 * 1000);
207 JSONParser jp = new JSONParser();
208 JSONObject val = (JSONObject) jp.parse(br);
209 String pingString = val.get("ping").toString();
210 return pingString != null;
211 } catch (Throwable t)
213 System.err.println("Error connecting to " + PING_URL + ": "
222 } catch (IOException e)
232 * returns a reader to a Fasta response from the Ensembl sequence endpoint
236 * @throws IOException
238 protected FileParse getSequenceReader(List<String> ids)
241 URL url = getUrl(ids);
243 BufferedReader reader = getHttpResponse(url, ids);
249 FileParse fp = new FileParse(reader, url.toString(), DataSourceType.URL);
254 * Gets a reader to the HTTP response, using the default read timeout of 5
260 * @throws IOException
262 protected BufferedReader getHttpResponse(URL url, List<String> ids)
265 return getHttpResponse(url, ids, DEFAULT_READ_TIMEOUT);
269 * Writes the HTTP request and gets the response as a reader.
273 * written as Json POST body if more than one
277 * @throws IOException
278 * if response code was not 200, or other I/O error
280 protected BufferedReader getHttpResponse(URL url, List<String> ids,
281 int readTimeout) throws IOException
283 // long now = System.currentTimeMillis();
284 HttpURLConnection connection = (HttpURLConnection) url.openConnection();
287 * POST method allows multiple queries in one request; it is supported for
288 * sequence queries, but not for overlap
290 boolean multipleIds = ids != null && ids.size() > 1;
291 connection.setRequestMethod(multipleIds ? HttpMethod.POST
293 connection.setRequestProperty("Content-Type",
294 getRequestMimeType(multipleIds));
295 connection.setRequestProperty("Accept", getResponseMimeType());
297 connection.setUseCaches(false);
298 connection.setDoInput(true);
299 connection.setDoOutput(multipleIds);
301 connection.setConnectTimeout(CONNECT_TIMEOUT_MS);
302 connection.setReadTimeout(readTimeout);
306 writePostBody(connection, ids);
309 int responseCode = connection.getResponseCode();
311 if (responseCode != 200)
314 * note: a GET request for an invalid id returns an error code e.g. 415
315 * but POST request returns 200 and an empty Fasta response
317 System.err.println("Response code " + responseCode + " for " + url);
321 InputStream response = connection.getInputStream();
323 // System.out.println(getClass().getName() + " took "
324 // + (System.currentTimeMillis() - now) + "ms to fetch");
326 checkRateLimits(connection);
328 BufferedReader reader = null;
329 reader = new BufferedReader(new InputStreamReader(response, "UTF-8"));
334 * Inspect response headers for any sign of server overload and respect any
335 * 'retry-after' directive
337 * @see https://github.com/Ensembl/ensembl-rest/wiki/Rate-Limits
340 void checkRateLimits(HttpURLConnection connection)
342 // number of requests allowed per time interval:
343 String limit = connection.getHeaderField("X-RateLimit-Limit");
344 // length of quota time interval in seconds:
345 // String period = connection.getHeaderField("X-RateLimit-Period");
346 // seconds remaining until usage quota is reset:
347 String reset = connection.getHeaderField("X-RateLimit-Reset");
348 // number of requests remaining from quota for current period:
349 String remaining = connection.getHeaderField("X-RateLimit-Remaining");
350 // number of seconds to wait before retrying (if remaining == 0)
351 String retryDelay = connection.getHeaderField("Retry-After");
356 EnsemblInfo info = domainData.get(getDomain());
357 if (retryDelay != null)
359 System.err.println("Ensembl REST service rate limit exceeded, wait "
360 + retryDelay + " seconds before retrying");
363 info.retryAfter = System.currentTimeMillis()
364 + (1000 * Integer.valueOf(retryDelay));
365 } catch (NumberFormatException e)
367 System.err.println("Unexpected value for Retry-After: "
375 // System.out.println(String.format(
376 // "%s Ensembl requests remaining of %s (reset in %ss)",
377 // remaining, limit, reset));
382 * Rechecks if Ensembl is responding, unless the last check was successful and
383 * the retest interval has not yet elapsed. Returns true if Ensembl is up,
384 * else false. Also retrieves and saves the current version of Ensembl data
385 * and REST services at intervals.
389 protected boolean isEnsemblAvailable()
391 EnsemblInfo info = domainData.get(getDomain());
393 long now = System.currentTimeMillis();
396 * check if we are waiting for 'Retry-After' to expire
398 if (info.retryAfter > now)
400 System.err.println("Still " + (1 + (info.retryAfter - now) / 1000)
401 + " secs to wait before retrying Ensembl");
410 * recheck if Ensembl is up if it was down, or the recheck period has elapsed
412 boolean retestAvailability = (now - info.lastAvailableCheckTime) > AVAILABILITY_RETEST_INTERVAL;
413 if (!info.restAvailable || retestAvailability)
415 info.restAvailable = checkEnsembl();
416 info.lastAvailableCheckTime = now;
420 * refetch Ensembl versions if the recheck period has elapsed
422 boolean refetchVersion = (now - info.lastVersionCheckTime) > VERSION_RETEST_INTERVAL;
425 checkEnsemblRestVersion();
426 checkEnsemblDataVersion();
427 info.lastVersionCheckTime = now;
430 return info.restAvailable;
434 * Constructs, writes and flushes the POST body of the request, containing the
435 * query ids in JSON format
439 * @throws IOException
441 protected void writePostBody(HttpURLConnection connection,
442 List<String> ids) throws IOException
445 StringBuilder postBody = new StringBuilder(64);
446 postBody.append("{\"ids\":[");
448 for (String id : ids)
452 postBody.append(",");
455 postBody.append("\"");
456 postBody.append(id.trim());
457 postBody.append("\"");
459 postBody.append("]}");
460 byte[] thepostbody = postBody.toString().getBytes();
461 connection.setRequestProperty("Content-Length",
462 Integer.toString(thepostbody.length));
463 DataOutputStream wr = new DataOutputStream(connection.getOutputStream());
464 wr.write(thepostbody);
470 * Fetches and checks Ensembl's REST version number
474 private void checkEnsemblRestVersion()
476 EnsemblInfo info = domainData.get(getDomain());
478 JSONParser jp = new JSONParser();
482 url = new URL(getDomain()
483 + "/info/rest?content-type=application/json");
484 BufferedReader br = getHttpResponse(url, null);
485 JSONObject val = (JSONObject) jp.parse(br);
486 String version = val.get("release").toString();
487 String majorVersion = version.substring(0, version.indexOf("."));
488 String expected = info.expectedRestVersion;
489 String expectedMajorVersion = expected.substring(0,
490 expected.indexOf("."));
491 info.restMajorVersionMismatch = false;
495 * if actual REST major version is ahead of what we expect,
496 * record this in case we want to warn the user
498 if (Float.valueOf(majorVersion) > Float
499 .valueOf(expectedMajorVersion))
501 info.restMajorVersionMismatch = true;
503 } catch (NumberFormatException e)
505 System.err.println("Error in REST version: " + e.toString());
509 * check if REST version is later than what Jalview has tested against,
510 * if so warn; we don't worry if it is earlier (this indicates Jalview has
511 * been tested in advance against the next pending REST version)
513 boolean laterVersion = StringUtils.compareVersions(version, expected) == 1;
516 System.err.println(String.format(
517 "Expected %s REST version %s but found %s, see %s",
518 getDbSource(), expected, version, REST_CHANGE_LOG));
520 info.restVersion = version;
521 } catch (Throwable t)
523 System.err.println("Error checking Ensembl REST version: "
528 public boolean isRestMajorVersionMismatch()
530 return domainData.get(getDomain()).restMajorVersionMismatch;
534 * Fetches and checks Ensembl's data version number
538 private void checkEnsemblDataVersion()
540 JSONParser jp = new JSONParser();
544 url = new URL(getDomain()
545 + "/info/data?content-type=application/json");
546 BufferedReader br = getHttpResponse(url, null);
547 JSONObject val = (JSONObject) jp.parse(br);
548 JSONArray versions = (JSONArray) val.get("releases");
549 domainData.get(getDomain()).dataVersion = versions.get(0).toString();
550 } catch (Throwable t)
552 System.err.println("Error checking Ensembl data version: "
557 public String getEnsemblDataVersion()
559 return domainData.get(getDomain()).dataVersion;
563 public String getDbVersion()
565 return getEnsemblDataVersion();