X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fext%2Fensembl%2FEnsemblRestClient.java;h=01b26d051366b8157b3db2323e2b684e0b78d701;hb=3da878124135ff033f42d19d8733891b09e953cd;hp=f9cfe052a2f7487113cb48fe68a4e5f017ffbc94;hpb=a96386b344a49391588762301ad71f1d39c11ca7;p=jalview.git
diff --git a/src/jalview/ext/ensembl/EnsemblRestClient.java b/src/jalview/ext/ensembl/EnsemblRestClient.java
index f9cfe05..01b26d0 100644
--- a/src/jalview/ext/ensembl/EnsemblRestClient.java
+++ b/src/jalview/ext/ensembl/EnsemblRestClient.java
@@ -1,6 +1,29 @@
+/*
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
+ *
+ * This file is part of Jalview.
+ *
+ * Jalview is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation, either version 3
+ * of the License, or (at your option) any later version.
+ *
+ * Jalview is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+ * PURPOSE. See the GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Jalview. If not, see .
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
+ */
package jalview.ext.ensembl;
-import jalview.io.FileParse;
+import jalview.bin.Cache;
+import jalview.javascript.json.JSON;
+import jalview.util.JSONUtils;
+import jalview.util.Platform;
import jalview.util.StringUtils;
import java.io.BufferedReader;
@@ -8,8 +31,10 @@ import java.io.DataOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
+import java.io.Reader;
import java.net.HttpURLConnection;
import java.net.MalformedURLException;
+import java.net.ProtocolException;
import java.net.URL;
import java.util.HashMap;
import java.util.List;
@@ -17,11 +42,7 @@ import java.util.Map;
import javax.ws.rs.HttpMethod;
-import org.json.simple.JSONArray;
-import org.json.simple.JSONObject;
-import org.json.simple.parser.JSONParser;
-
-import com.stevesoft.pat.Regex;
+import org.json.simple.parser.ParseException;
/**
* Base class for Ensembl REST service clients
@@ -30,37 +51,49 @@ import com.stevesoft.pat.Regex;
*/
abstract class EnsemblRestClient extends EnsemblSequenceFetcher
{
+
+ static {
+ Cache.addJ2SDirectDatabaseCall("http://rest.ensembl");
+ Cache.addJ2SDirectDatabaseCall("https://rest.ensembl");
+ }
+
+ private static final int DEFAULT_READ_TIMEOUT = 5 * 60 * 1000; // 5 minutes
+
+ private static final int CONNECT_TIMEOUT_MS = 10 * 1000; // 10 seconds
+
+ private static final int MAX_RETRIES = 3;
+
+ private static final int HTTP_OK = 200;
+
+ private static final int HTTP_OVERLOAD = 429;
+
/*
* update these constants when Jalview has been checked / updated for
- * changes to Ensembl REST API
+ * changes to Ensembl REST API, and updated JAL-3018
* @see https://github.com/Ensembl/ensembl-rest/wiki/Change-log
+ * @see http://rest.ensembl.org/info/rest?content-type=application/json
*/
- private static final String LATEST_ENSEMBLGENOMES_REST_VERSION = "4.4";
+ private static final String LATEST_ENSEMBLGENOMES_REST_VERSION = "9.0";
- private static final String LATEST_ENSEMBL_REST_VERSION = "4.5";
+ private static final String LATEST_ENSEMBL_REST_VERSION = "9.0";
- private static Map domainData;
+ private static final String REST_CHANGE_LOG = "https://github.com/Ensembl/ensembl-rest/wiki/Change-log";
- // @see https://github.com/Ensembl/ensembl-rest/wiki/Output-formats
- private static final String PING_URL = "http://rest.ensembl.org/info/ping.json";
+ private static Map domainData;
private final static long AVAILABILITY_RETEST_INTERVAL = 10000L; // 10 seconds
private final static long VERSION_RETEST_INTERVAL = 1000L * 3600; // 1 hr
- private static final Regex TRANSCRIPT_REGEX = new Regex(
- "(ENS)([A-Z]{3}|)T[0-9]{11}$");
-
- private static final Regex GENE_REGEX = new Regex(
- "(ENS)([A-Z]{3}|)G[0-9]{11}$");
+ protected static final String CONTENT_TYPE_JSON = "?content-type=application/json";
static
{
- domainData = new HashMap();
- domainData.put(ENSEMBL_REST, new EnsemblInfo(ENSEMBL_REST,
- LATEST_ENSEMBL_REST_VERSION));
- domainData.put(ENSEMBL_GENOMES_REST, new EnsemblInfo(
- ENSEMBL_GENOMES_REST, LATEST_ENSEMBLGENOMES_REST_VERSION));
+ domainData = new HashMap<>();
+ domainData.put(DEFAULT_ENSEMBL_BASEURL,
+ new EnsemblData(DEFAULT_ENSEMBL_BASEURL, LATEST_ENSEMBL_REST_VERSION));
+ domainData.put(DEFAULT_ENSEMBL_GENOMES_BASEURL, new EnsemblData(
+ DEFAULT_ENSEMBL_GENOMES_BASEURL, LATEST_ENSEMBLGENOMES_REST_VERSION));
}
protected volatile boolean inProgress = false;
@@ -70,7 +103,21 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
*/
public EnsemblRestClient()
{
- this(ENSEMBL_REST);
+ super();
+
+ /*
+ * initialise domain info lazily
+ */
+ if (!domainData.containsKey(ensemblDomain))
+ {
+ domainData.put(ensemblDomain,
+ new EnsemblData(ensemblDomain, LATEST_ENSEMBL_REST_VERSION));
+ }
+ if (!domainData.containsKey(ensemblGenomesDomain))
+ {
+ domainData.put(ensemblGenomesDomain, new EnsemblData(
+ ensemblGenomesDomain, LATEST_ENSEMBLGENOMES_REST_VERSION));
+ }
}
/**
@@ -83,30 +130,6 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
setDomain(d);
}
- /**
- * Answers true if the query matches the regular expression pattern for an
- * Ensembl transcript stable identifier
- *
- * @param query
- * @return
- */
- public boolean isTranscriptIdentifier(String query)
- {
- return query == null ? false : TRANSCRIPT_REGEX.search(query);
- }
-
- /**
- * Answers true if the query matches the regular expression pattern for an
- * Ensembl gene stable identifier
- *
- * @param query
- * @return
- */
- public boolean isGeneIdentifier(String query)
- {
- return query == null ? false : GENE_REGEX.search(query);
- }
-
@Override
public boolean queryInProgress()
{
@@ -137,22 +160,28 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
protected abstract boolean useGetRequest();
/**
- * Return the desired value for the Content-Type request header
- *
- * @param multipleIds
+ * Returns the desired value for the Content-Type request header. Default is
+ * application/json, override if required to vary this.
*
* @return
* @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
*/
- protected abstract String getRequestMimeType(boolean multipleIds);
+ protected String getRequestMimeType()
+ {
+ return "application/json";
+ }
/**
- * Return the desired value for the Accept request header
+ * Return the desired value for the Accept request header. Default is
+ * application/json, override if required to vary this.
*
* @return
* @see https://github.com/Ensembl/ensembl-rest/wiki/HTTP-Headers
*/
- protected abstract String getResponseMimeType();
+ protected String getResponseMimeType()
+ {
+ return "application/json";
+ }
/**
* Checks Ensembl's REST 'ping' endpoint, and returns true if response
@@ -161,162 +190,218 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
* @see http://rest.ensembl.org/documentation/info/ping
* @return
*/
- private boolean checkEnsembl()
+ @SuppressWarnings("unchecked")
+boolean checkEnsembl()
{
- HttpURLConnection conn = null;
+ BufferedReader br = null;
+ String pingUrl = getDomain() + "/info/ping" + CONTENT_TYPE_JSON;
try
{
// note this format works for both ensembl and ensemblgenomes
// info/ping.json works for ensembl only (March 2016)
- URL ping = new URL(getDomain()
- + "/info/ping?content-type=application/json");
-
+
+
+
+
/*
* expect {"ping":1} if ok
+ * if ping takes more than 2 seconds to respond, treat as if unavailable
*/
- BufferedReader br = getHttpResponse(ping, null);
- JSONParser jp = new JSONParser();
- JSONObject val = (JSONObject) jp.parse(br);
+ Map val = (Map) getJSON(new URL(pingUrl), null, 2 * 1000, MODE_MAP, null);
+ if (val == null)
+ return false;
String pingString = val.get("ping").toString();
return pingString != null;
} catch (Throwable t)
{
- System.err.println("Error connecting to " + PING_URL + ": "
- + t.getMessage());
+ System.err.println(
+ "Error connecting to " + pingUrl + ": " + t.getMessage());
} finally
{
- if (conn != null)
+ if (br != null)
{
- conn.disconnect();
+ try
+ {
+ br.close();
+ } catch (IOException e)
+ {
+ // ignore
+ }
}
}
return false;
}
+
+ protected final static int MODE_ARRAY = 0;
+ protected final static int MODE_MAP = 1;
+ protected final static int MODE_ITERATOR = 2;
+
+// /**
+// * Returns a reader to a (Json) response from the Ensembl sequence endpoint.
+// * If the request failed the return value may be null.
+// *
+// * @param ids
+// * @return
+// * @throws IOException
+// * @throws ParseException
+// */
+// protected Object getSequenceJSON(List ids, int mode)
+// throws IOException, ParseException
+// {
+// URL url = getUrl(ids);
+// return getJSON(url, ids, -1, mode);
+// }
+//
+// /**
+// * Gets a reader to the HTTP response, using the default read timeout of 5
+// * minutes
+// *
+// * @param url
+// * @param ids
+// * @return
+// * @throws IOException
+// */
+// protected BufferedReader getHttpResponse(URL url, List ids)
+// throws IOException
+// {
+// return getHttpResponse(url, ids, DEFAULT_READ_TIMEOUT);
+// }
+
/**
- * returns a reader to a Fasta response from the Ensembl sequence endpoint
+ * Sends the HTTP request and gets the response as a reader. Returns null if
+ * the HTTP response code was not 200.
*
+ * @param url
* @param ids
+ * written as Json POST body if more than one
+ * @param readTimeout
+ * in milliseconds
* @return
* @throws IOException
*/
- protected FileParse getSequenceReader(List ids)
- throws IOException
+ private BufferedReader getHttpResponse(URL url, List ids,
+ int readTimeout) throws IOException
{
- URL url = getUrl(ids);
-
- BufferedReader reader = getHttpResponse(url, ids);
- FileParse fp = new FileParse(reader, url.toString(), "HTTP_POST");
- return fp;
+ if (readTimeout < 0)
+ readTimeout = DEFAULT_READ_TIMEOUT;
+ int retriesLeft = MAX_RETRIES;
+ HttpURLConnection connection = null;
+ int responseCode = 0;
+
+ if (Platform.isJS()) {
+ JSON.setAjax(url);
+ }
+
+ while (retriesLeft > 0)
+ {
+ connection = tryConnection(url, ids, readTimeout);
+ responseCode = connection.getResponseCode();
+ if (responseCode == HTTP_OVERLOAD) // 429
+ {
+ retriesLeft--;
+ checkRetryAfter(connection);
+ }
+ else
+ {
+ retriesLeft = 0;
+ }
+ }
+ if (responseCode != HTTP_OK) // 200
+ {
+ /*
+ * note: a GET request for an invalid id returns an error code e.g. 415
+ * but POST request returns 200 and an empty Fasta response
+ */
+ System.err.println("Response code " + responseCode);// + " for " + url);
+ return null;
+ }
+
+
+ InputStream response = connection.getInputStream();
+
+ if (Platform.isJS()) {
+ return JSON.getJSONReader(response);
+ }
+
+ // System.out.println(getClass().getName() + " took "
+ // + (System.currentTimeMillis() - now) + "ms to fetch");
+
+ return new BufferedReader(new InputStreamReader(response, "UTF-8"));
}
/**
- * Writes the HTTP request and gets the response as a reader.
- *
* @param url
* @param ids
- * written as Json POST body if more than one
+ * @param readTimeout
* @return
* @throws IOException
- * if response code was not 200, or other I/O error
+ * @throws ProtocolException
*/
- protected BufferedReader getHttpResponse(URL url, List ids)
- throws IOException
+ protected HttpURLConnection tryConnection(URL url, List ids,
+ int readTimeout) throws IOException, ProtocolException
{
- // long now = System.currentTimeMillis();
+ // System.out.println(System.currentTimeMillis() + " " + url);
+
HttpURLConnection connection = (HttpURLConnection) url.openConnection();
-
+
/*
* POST method allows multiple queries in one request; it is supported for
* sequence queries, but not for overlap
*/
boolean multipleIds = ids != null && ids.size() > 1;
- connection.setRequestMethod(multipleIds ? HttpMethod.POST
- : HttpMethod.GET);
- connection.setRequestProperty("Content-Type",
- getRequestMimeType(multipleIds));
+ connection.setRequestMethod(
+ multipleIds ? HttpMethod.POST : HttpMethod.GET);
+ connection.setRequestProperty("Content-Type", getRequestMimeType());
connection.setRequestProperty("Accept", getResponseMimeType());
- connection.setUseCaches(false);
connection.setDoInput(true);
connection.setDoOutput(multipleIds);
+ connection.setUseCaches(false);
+ connection.setConnectTimeout(CONNECT_TIMEOUT_MS);
+ connection.setReadTimeout(readTimeout);
+
if (multipleIds)
{
writePostBody(connection, ids);
}
-
- InputStream response = connection.getInputStream();
- int responseCode = connection.getResponseCode();
-
- if (responseCode != 200)
- {
- /*
- * note: a GET request for an invalid id returns an error code e.g. 415
- * but POST request returns 200 and an empty Fasta response
- */
- throw new IOException(
- "Response code was not 200. Detected response was "
- + responseCode);
- }
- // System.out.println(getClass().getName() + " took "
- // + (System.currentTimeMillis() - now) + "ms to fetch");
-
- checkRateLimits(connection);
-
- BufferedReader reader = null;
- reader = new BufferedReader(new InputStreamReader(response, "UTF-8"));
- return reader;
+ return connection;
}
/**
- * Inspect response headers for any sign of server overload and respect any
- * 'retry-after' directive
+ * Inspects response headers for a 'retry-after' directive, and waits for the
+ * directed period (if less than 10 seconds)
*
* @see https://github.com/Ensembl/ensembl-rest/wiki/Rate-Limits
* @param connection
*/
- void checkRateLimits(HttpURLConnection connection)
+ void checkRetryAfter(HttpURLConnection connection)
{
- // number of requests allowed per time interval:
- String limit = connection.getHeaderField("X-RateLimit-Limit");
- // length of quota time interval in seconds:
- // String period = connection.getHeaderField("X-RateLimit-Period");
- // seconds remaining until usage quota is reset:
- String reset = connection.getHeaderField("X-RateLimit-Reset");
- // number of requests remaining from quota for current period:
- String remaining = connection.getHeaderField("X-RateLimit-Remaining");
- // number of seconds to wait before retrying (if remaining == 0)
String retryDelay = connection.getHeaderField("Retry-After");
// to test:
// retryDelay = "5";
- EnsemblInfo info = domainData.get(getDomain());
if (retryDelay != null)
{
- System.err.println("Ensembl REST service rate limit exceeded, wait "
- + retryDelay + " seconds before retrying");
try
{
- info.retryAfter = System.currentTimeMillis()
- + (1000 * Integer.valueOf(retryDelay));
- } catch (NumberFormatException e)
+ int retrySecs = Integer.valueOf(retryDelay);
+ if (retrySecs > 0 && retrySecs < 10)
+ {
+ System.err
+ .println("Ensembl REST service rate limit exceeded, waiting "
+ + retryDelay + " seconds before retrying");
+ Thread.sleep(1000 * retrySecs);
+ }
+ } catch (NumberFormatException | InterruptedException e)
{
- System.err.println("Unexpected value for Retry-After: "
- + retryDelay);
+ System.err.println("Error handling Retry-After: " + e.getMessage());
}
}
- else
- {
- info.retryAfter = 0;
- // debug:
- // System.out.println(String.format(
- // "%s Ensembl requests remaining of %s (reset in %ss)",
- // remaining, limit, reset));
- }
}
-
+
/**
* Rechecks if Ensembl is responding, unless the last check was successful and
* the retest interval has not yet elapsed. Returns true if Ensembl is up,
@@ -327,28 +412,15 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
*/
protected boolean isEnsemblAvailable()
{
- EnsemblInfo info = domainData.get(getDomain());
+ EnsemblData info = domainData.get(getDomain());
long now = System.currentTimeMillis();
/*
- * check if we are waiting for 'Retry-After' to expire
- */
- if (info.retryAfter > now)
- {
- System.err.println("Still " + (1 + (info.retryAfter - now) / 1000)
- + " secs to wait before retrying Ensembl");
- return false;
- }
- else
- {
- info.retryAfter = 0;
- }
-
- /*
* recheck if Ensembl is up if it was down, or the recheck period has elapsed
*/
- boolean retestAvailability = (now - info.lastAvailableCheckTime) > AVAILABILITY_RETEST_INTERVAL;
+ boolean retestAvailability = (now
+ - info.lastAvailableCheckTime) > AVAILABILITY_RETEST_INTERVAL;
if (!info.restAvailable || retestAvailability)
{
info.restAvailable = checkEnsembl();
@@ -358,7 +430,8 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
/*
* refetch Ensembl versions if the recheck period has elapsed
*/
- boolean refetchVersion = (now - info.lastVersionCheckTime) > VERSION_RETEST_INTERVAL;
+ boolean refetchVersion = (now
+ - info.lastVersionCheckTime) > VERSION_RETEST_INTERVAL;
if (refetchVersion)
{
checkEnsemblRestVersion();
@@ -384,8 +457,9 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
StringBuilder postBody = new StringBuilder(64);
postBody.append("{\"ids\":[");
first = true;
- for (String id : ids)
+ for (int i = 0, n = ids.size(); i < n; i++)
{
+ String id = ids.get(i);
if (!first)
{
postBody.append(",");
@@ -399,29 +473,94 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher
byte[] thepostbody = postBody.toString().getBytes();
connection.setRequestProperty("Content-Length",
Integer.toString(thepostbody.length));
- DataOutputStream wr = new DataOutputStream(connection.getOutputStream());
+ DataOutputStream wr = new DataOutputStream(
+ connection.getOutputStream());
wr.write(thepostbody);
wr.flush();
wr.close();
}
+ /**
+ * Primary access point to parsed JSON data, including the call to retrieve and
+ * parsing.
+ *
+ * @param url request url; if null, getUrl(ids) will be used
+ * @param ids optional; may be null
+ * @param msDelay -1 for default delay
+ * @param mode map, array, or array iterator
+ * @param mapKey an optional key for an outer map
+ * @return a Map, List, Iterator, or null
+ * @throws IOException
+ * @throws ParseException
+ *
+ * @author Bob Hanson 2019
+ */
+ @SuppressWarnings("unchecked")
+ protected Object getJSON(URL url, List ids, int msDelay, int mode, String mapKey) throws IOException, ParseException {
+
+ if (url == null)
+ url = getUrl(ids);
+
+ Reader br = null;
+ try {
+
+ Platform.timeCheck(null, Platform.TIME_MARK);
+
+ br = (url == null ? null : getHttpResponse(url, ids, msDelay));
+
+ Object ret = (br == null ? null : JSONUtils.parse(br));
+
+ Platform.timeCheck("EnsembleRestClient.getJSON " + url, Platform.TIME_MARK);
+
+
+ if (ret != null && mapKey != null)
+ ret = ((Map) ret).get(mapKey);
+ if (ret == null)
+ {
+ return null;
+ }
+ switch (mode) {
+ case MODE_ARRAY:
+ case MODE_MAP:
+ break;
+ case MODE_ITERATOR:
+ ret = ((List