X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fext%2Fensembl%2FEnsemblRestClient.java;fp=src%2Fjalview%2Fext%2Fensembl%2FEnsemblRestClient.java;h=b1bc8e5edb4abb24fd3d10fbab66f4ecd37878ec;hb=0efc6021e63ff62bb2e275501586c1900561b303;hp=ad6c70cb28e99269c963fc78d34c1a257b8d3eb2;hpb=771d1758a42abc6e3e330dd6c5117734acf0067f;p=jalview.git diff --git a/src/jalview/ext/ensembl/EnsemblRestClient.java b/src/jalview/ext/ensembl/EnsemblRestClient.java index ad6c70c..b1bc8e5 100644 --- a/src/jalview/ext/ensembl/EnsemblRestClient.java +++ b/src/jalview/ext/ensembl/EnsemblRestClient.java @@ -31,6 +31,7 @@ import java.io.InputStream; import java.io.InputStreamReader; import java.net.HttpURLConnection; import java.net.MalformedURLException; +import java.net.ProtocolException; import java.net.URL; import java.util.HashMap; import java.util.List; @@ -42,8 +43,6 @@ import org.json.simple.JSONArray; import org.json.simple.JSONObject; import org.json.simple.parser.JSONParser; -import com.stevesoft.pat.Regex; - /** * Base class for Ensembl REST service clients * @@ -55,15 +54,21 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher 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 (ref JAL-2105) * @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 = "5.0"; + private static final String LATEST_ENSEMBLGENOMES_REST_VERSION = "6.0"; - private static final String LATEST_ENSEMBL_REST_VERSION = "5.0"; + private static final String LATEST_ENSEMBL_REST_VERSION = "6.1"; private static final String REST_CHANGE_LOG = "https://github.com/Ensembl/ensembl-rest/wiki/Change-log"; @@ -76,18 +81,11 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher private final static long VERSION_RETEST_INTERVAL = 1000L * 3600; // 1 hr - private static final Regex PROTEIN_REGEX = new Regex( - "(ENS)([A-Z]{3}|)P[0-9]{11}$"); - - private static final Regex TRANSCRIPT_REGEX = new Regex( - "(ENS)([A-Z]{3}|)T[0-9]{11}$"); - - 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 = new HashMap<>(); domainData.put(ENSEMBL_REST, new EnsemblInfo(ENSEMBL_REST, LATEST_ENSEMBL_REST_VERSION)); domainData.put(ENSEMBL_GENOMES_REST, new EnsemblInfo( @@ -114,42 +112,6 @@ 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 protein stable identifier - * - * @param query - * @return - */ - public boolean isProteinIdentifier(String query) - { - return query == null ? false : PROTEIN_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() { @@ -204,21 +166,25 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher * @see http://rest.ensembl.org/documentation/info/ping * @return */ - private boolean checkEnsembl() + boolean checkEnsembl() { BufferedReader br = null; 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"); + URL ping = new URL(getDomain() + "/info/ping" + CONTENT_TYPE_JSON); /* * expect {"ping":1} if ok * if ping takes more than 2 seconds to respond, treat as if unavailable */ br = getHttpResponse(ping, null, 2 * 1000); + if (br == null) + { + // error reponse status + return false; + } JSONParser jp = new JSONParser(); JSONObject val = (JSONObject) jp.parse(br); String pingString = val.get("ping").toString(); @@ -281,7 +247,7 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher } /** - * Writes the HTTP request and gets the response as a reader. + * Sends the HTTP request and gets the response as a reader * * @param url * @param ids @@ -295,7 +261,56 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher protected BufferedReader getHttpResponse(URL url, List ids, int readTimeout) throws IOException { - // long now = System.currentTimeMillis(); + int retriesLeft = MAX_RETRIES; + HttpURLConnection connection = null; + int responseCode = 0; + + 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(); + + // System.out.println(getClass().getName() + " took " + // + (System.currentTimeMillis() - now) + "ms to fetch"); + + BufferedReader reader = null; + reader = new BufferedReader(new InputStreamReader(response, "UTF-8")); + return reader; + } + + /** + * @param url + * @param ids + * @param readTimeout + * @return + * @throws IOException + * @throws ProtocolException + */ + protected HttpURLConnection tryConnection(URL url, List ids, + int readTimeout) throws IOException, ProtocolException + { + // System.out.println(System.currentTimeMillis() + " " + url); HttpURLConnection connection = (HttpURLConnection) url.openConnection(); /* @@ -320,77 +335,40 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher { writePostBody(connection, ids); } - - 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 - */ - System.err.println("Response code " + responseCode + " for " + url); - return null; - } - // get content - InputStream response = connection.getInputStream(); - - // 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)); - } } /** @@ -408,20 +386,6 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher 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 @@ -497,9 +461,12 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher URL url = null; try { - url = new URL( - getDomain() + "/info/rest?content-type=application/json"); + url = new URL(getDomain() + "/info/rest" + CONTENT_TYPE_JSON); BufferedReader br = getHttpResponse(url, null); + if (br == null) + { + return; + } JSONObject val = (JSONObject) jp.parse(br); String version = val.get("release").toString(); String majorVersion = version.substring(0, version.indexOf(".")); @@ -558,18 +525,35 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher { JSONParser jp = new JSONParser(); URL url = null; + BufferedReader br = null; + try { - url = new URL( - getDomain() + "/info/data?content-type=application/json"); - BufferedReader br = getHttpResponse(url, null); - JSONObject val = (JSONObject) jp.parse(br); - JSONArray versions = (JSONArray) val.get("releases"); - domainData.get(getDomain()).dataVersion = versions.get(0).toString(); + url = new URL(getDomain() + "/info/data" + CONTENT_TYPE_JSON); + br = getHttpResponse(url, null); + if (br != null) + { + JSONObject val = (JSONObject) jp.parse(br); + JSONArray versions = (JSONArray) val.get("releases"); + domainData.get(getDomain()).dataVersion = versions.get(0) + .toString(); + } } catch (Throwable t) { System.err.println( "Error checking Ensembl data version: " + t.getMessage()); + } finally + { + if (br != null) + { + try + { + br.close(); + } catch (IOException e) + { + // ignore + } + } } }