X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fext%2Fensembl%2FEnsemblRestClient.java;h=97172fa50aa7350b55b8e613ef13b8e36fbcca53;hb=181cd6607ecd631aa5972582ff1d99c5bea75b23;hp=7752e9c9c18bfd2bf0218383c1d9d7739584a1f1;hpb=f23f5cefe9c66bfa8b877c707d25862cabba3ddf;p=jalview.git diff --git a/src/jalview/ext/ensembl/EnsemblRestClient.java b/src/jalview/ext/ensembl/EnsemblRestClient.java index 7752e9c..97172fa 100644 --- a/src/jalview/ext/ensembl/EnsemblRestClient.java +++ b/src/jalview/ext/ensembl/EnsemblRestClient.java @@ -20,6 +20,7 @@ */ package jalview.ext.ensembl; +import jalview.bin.Jalview; import jalview.io.DataSourceType; import jalview.io.FileParse; import jalview.util.StringUtils; @@ -31,6 +32,7 @@ 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; @@ -53,34 +55,39 @@ 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 = "7.0"; - private static final String LATEST_ENSEMBL_REST_VERSION = "5.0"; + private static final String LATEST_ENSEMBL_REST_VERSION = "7.0"; private static final String REST_CHANGE_LOG = "https://github.com/Ensembl/ensembl-rest/wiki/Change-log"; - private static Map domainData; - - // @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 + 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; @@ -90,7 +97,21 @@ 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)); + } } /** @@ -133,22 +154,28 @@ 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 @@ -157,15 +184,15 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher * @see http://rest.ensembl.org/documentation/info/ping * @return */ - private boolean checkEnsembl() + boolean checkEnsembl() { 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"); + URL ping = new URL(pingUrl); /* * expect {"ping":1} if ok @@ -184,7 +211,7 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher } catch (Throwable t) { System.err.println( - "Error connecting to " + PING_URL + ": " + t.getMessage()); + "Error connecting to " + pingUrl + ": " + t.getMessage()); } finally { if (br != null) @@ -202,25 +229,20 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher } /** - * returns a reader to a Fasta response from the Ensembl sequence endpoint + * 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 */ - protected FileParse getSequenceReader(List ids) throws IOException + protected BufferedReader getSequenceReader(List ids) + throws IOException { URL url = getUrl(ids); BufferedReader reader = getHttpResponse(url, ids); - if (reader == null) - { - // request failed - return null; - } - FileParse fp = new FileParse(reader, url.toString(), - DataSourceType.URL); - return fp; + return reader; } /** @@ -239,7 +261,8 @@ 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. Returns null if + * the HTTP response code was not 200. * * @param url * @param ids @@ -248,40 +271,29 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher * in milliseconds * @return * @throws IOException - * if response code was not 200, or other I/O error */ protected BufferedReader getHttpResponse(URL url, List ids, int readTimeout) throws IOException { - // long now = System.currentTimeMillis(); - 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.setRequestProperty("Accept", getResponseMimeType()); + int retriesLeft = MAX_RETRIES; + HttpURLConnection connection = null; + int responseCode = 0; - connection.setUseCaches(false); - connection.setDoInput(true); - connection.setDoOutput(multipleIds); - - connection.setConnectTimeout(CONNECT_TIMEOUT_MS); - connection.setReadTimeout(readTimeout); - - if (multipleIds) + while (retriesLeft > 0) { - writePostBody(connection, ids); + connection = tryConnection(url, ids, readTimeout); + responseCode = connection.getResponseCode(); + if (responseCode == HTTP_OVERLOAD) // 429 + { + retriesLeft--; + checkRetryAfter(connection); + } + else + { + retriesLeft = 0; + } } - - int responseCode = connection.getResponseCode(); - - if (responseCode != 200) + if (responseCode != HTTP_OK) // 200 { /* * note: a GET request for an invalid id returns an error code e.g. 415 @@ -290,65 +302,89 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher 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; } /** - * Inspect response headers for any sign of server overload and respect any - * 'retry-after' directive + * @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(); + + /* + * 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()); + connection.setRequestProperty("Accept", getResponseMimeType()); + + connection.setDoInput(true); + connection.setDoOutput(multipleIds); + + if (!Jalview.isJS()) + { + connection.setUseCaches(false); + connection.setConnectTimeout(CONNECT_TIMEOUT_MS); + connection.setReadTimeout(readTimeout); + } + + if (multipleIds) + { + writePostBody(connection, ids); + } + return connection; + } + + /** + * 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)); - } } /** @@ -361,25 +397,11 @@ 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 @@ -449,14 +471,13 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher */ private void checkEnsemblRestVersion() { - EnsemblInfo info = domainData.get(getDomain()); + EnsemblData info = domainData.get(getDomain()); JSONParser jp = new JSONParser(); 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) { @@ -520,11 +541,12 @@ 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); + url = new URL(getDomain() + "/info/data" + CONTENT_TYPE_JSON); + br = getHttpResponse(url, null); if (br != null) { JSONObject val = (JSONObject) jp.parse(br); @@ -536,6 +558,18 @@ abstract class EnsemblRestClient extends EnsemblSequenceFetcher { System.err.println( "Error checking Ensembl data version: " + t.getMessage()); + } finally + { + if (br != null) + { + try + { + br.close(); + } catch (IOException e) + { + // ignore + } + } } }