Merge pull request #213 from Swirrl/rdf4j-and-updates

Update drafter to RDF4j and latest grafter
Swirrl · Oct 19, 2017 · 352d5d3 · 352d5d3
2 parents bd78be4 + 9224497
commit 352d5d3
Show file tree

Hide file tree

Showing 33 changed files with 320 additions and 259 deletions.
diff --git a/env/dev/clj/drafter/routes/dump.clj b/env/dev/clj/drafter/routes/dump.clj
@@ -5,7 +5,8 @@
             [drafter.rdf.drafter-ontology :refer :all]
             [grafter.rdf :refer [add statements]]
             [grafter.rdf.formats :refer [rdf-trig]]
-            [grafter.rdf.io :refer [rdf-serializer default-prefixes]]
+            [grafter.rdf.io :refer [default-prefixes]]
+            [grafter.rdf4j.io :refer [rdf-writer]]
             [ring.util.io :as rio]
             [ring.util.response :refer [not-found]]
             [grafter.url :as url]))
@@ -20,7 +21,7 @@
   the RAW database as a Trig String for debugging.  Don't use on large
   databases as it will be loaded into memory."
   [db ostream]
-  (add (rdf-serializer ostream :format rdf-trig :prefixes drafter-prefixes)
+  (add (rdf-writer ostream :format rdf-trig :prefixes drafter-prefixes)
        (statements db)))
 
 (defn build-dump-route [backend]

diff --git a/project.clj b/project.clj
@@ -1,4 +1,4 @@
-(defproject drafter "2.1.6-SNAPSHOT"
+(defproject drafter "2.1.7-SNAPSHOT"
   :description "Backend PMD service"
   :url "http://github.com/Swirrl/drafter"
   :license {:name "Proprietary & Commercially Licensed Only"
@@ -23,10 +23,6 @@
                  [org.clojure/clojure "1.9.0-beta2"]
                  [org.clojure/math.combinatorics "0.1.4"]
 
-                 [org.openrdf.sesame/sesame-queryrender "2.8.11"]
-                 [org.openrdf.sesame/sesame-runtime "2.8.11"]
-                 [org.openrdf.sesame/sesame-queryresultio-sparqlxml "2.8.11"]
-
                  [cognician/dogstatsd-clj "0.1.2"]
 
                  [clj-yaml "0.4.0"] ;; for loading our Swagger schemas
@@ -49,7 +45,7 @@
 
                  [com.sun.mail/javax.mail "1.6.0"]
                  ;;[com.taoensso/tower "2.0.2"]
-                 [grafter "0.9.0"]
+                 [grafter "0.10.0"]
                  [grafter/url "0.2.5"]
                  ;[grafter/vocabularies "0.1.3"]
                  [lib-noir "0.9.9" :exclusions [compojure org.clojure/java.classpath org.clojure/tools.reader org.clojure/java.classpath]]
@@ -66,9 +62,7 @@
                  [org.apache.jena/jena-iri "3.4.0" :exclusions [org.slf4j/slf4j-api]]
 
                  [org.mindrot/jbcrypt "0.4"]
-                 [org.openrdf.sesame/sesame-queryrender "2.8.11"]
-                 [org.openrdf.sesame/sesame-queryresultio-sparqlxml "2.8.11"]
-                 [org.openrdf.sesame/sesame-runtime "2.8.11"]
+
                  [org.slf4j/slf4j-log4j12 "1.7.25" :exclusions [log4j org.slf4j/slf4j-api]]
                  [prismatic/schema "1.1.7"]
 

diff --git a/src-java/drafter_sparql_repository/drafter-sparql-repository.iml b/src-java/drafter_sparql_repository/drafter-sparql-repository.iml
diff --git a/src-java/drafter_sparql_repository/pom.xml b/src-java/drafter_sparql_repository/pom.xml
@@ -40,9 +40,9 @@
                 NOTE: this version of sesame should be kept in sync with the version in the parent
                 leiningen project.clj.
             -->
-            <groupId>org.openrdf.sesame</groupId>
-            <artifactId>sesame-runtime</artifactId>
-            <version>2.8.11</version>
+            <groupId>org.eclipse.rdf4j</groupId>
+            <artifactId>rdf4j-runtime</artifactId>
+            <version>2.2.2</version>
             <!--<exclusions>-->
                 <!--<exclusion>-->
                     <!--<groupId>org.openrdf.sesame</groupId>-->

diff --git a/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSPARQLConnection.java b/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSPARQLConnection.java
@@ -1,14 +1,16 @@
 package drafter.rdf;
 
-import org.openrdf.query.*;
-import org.openrdf.repository.RepositoryException;
-import org.openrdf.repository.sparql.SPARQLConnection;
-import org.openrdf.repository.sparql.SPARQLRepository;
-import org.openrdf.http.client.SparqlSession;
+import org.eclipse.rdf4j.http.client.SPARQLProtocolSession;
+import org.eclipse.rdf4j.query.*;
+import org.eclipse.rdf4j.repository.RepositoryException;
+import org.eclipse.rdf4j.repository.sparql.SPARQLConnection;
+import org.eclipse.rdf4j.repository.sparql.SPARQLRepository;
+import org.eclipse.rdf4j.http.client.SPARQLProtocolSession;
+
 
 public class DrafterSPARQLConnection extends SPARQLConnection {
 
-    public DrafterSPARQLConnection(SPARQLRepository repository, SparqlSession sparqlSession) {
+    public DrafterSPARQLConnection(SPARQLRepository repository, SPARQLProtocolSession sparqlSession) {
         super(repository, sparqlSession);
     }
 

diff --git a/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSPARQLRepository.java b/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSPARQLRepository.java
@@ -2,42 +2,57 @@
 
 import org.apache.http.client.HttpClient;
 import org.apache.http.impl.client.HttpClients;
-import org.openrdf.http.client.SesameClient;
-import org.openrdf.model.vocabulary.SESAME;
-import org.openrdf.repository.RepositoryConnection;
-import org.openrdf.repository.RepositoryException;
-import org.openrdf.repository.sparql.SPARQLRepository;
+import org.eclipse.rdf4j.http.client.HttpClientSessionManager;
+import org.eclipse.rdf4j.http.client.SesameClient;
+import org.eclipse.rdf4j.repository.RepositoryConnection;
+import org.eclipse.rdf4j.repository.RepositoryException;
+import org.eclipse.rdf4j.repository.sparql.SPARQLRepository;
 
 public class DrafterSPARQLRepository extends SPARQLRepository {
 
-    private SesameClient sesameClient;
+    private HttpClientSessionManager httpClientManager;
     private Integer maxConcurrentHttpConnections;
 
     public DrafterSPARQLRepository(String queryEndpoint) { super(queryEndpoint); }
     public DrafterSPARQLRepository(String queryEndpoint, String updateEndpoint) {
         super(queryEndpoint, updateEndpoint);
     }
 
-    @Override public RepositoryConnection getConnection() throws RepositoryException {
-        if(!this.isInitialized()) {
-            throw new RepositoryException("SPARQLRepository not initialized.");
-        } else {
-            return new DrafterSPARQLConnection(this, createHTTPClient());
-        }
-    }
+     @Override public RepositoryConnection getConnection() throws RepositoryException {
+         if(!this.isInitialized()) {
+             throw new RepositoryException("SPARQLRepository not initialized.");
+         } else {
+             return new DrafterSPARQLConnection(this, createHTTPClient());
+         }
+     }
 
-    @Override public synchronized SesameClient getSesameClient() {
-        if (this.sesameClient == null) {
-            HttpClient httpClient = newHttpClient();
-            this.sesameClient = new DrafterSesameClientImpl(httpClient);
-        }
-        return this.sesameClient;
-    }
+     public synchronized HttpClientSessionManager getHttpClientSessionManager() {
+         if (this.httpClientManager == null) {
+             HttpClient httpClient = newHttpClient();
+             this.httpClientManager = new DrafterSesameClientImpl(httpClient);
+         }
+         return this.httpClientManager;
+     }
 
-    @Override public synchronized void setSesameClient(SesameClient client) {
-        this.sesameClient = client;
+//     @Override public synchronized SesameClient getSesameClient() {
+//         if (this.httpClientManager == null) {
+//             HttpClient httpClient = newHttpClient();
+//             this.httpClientManager = new DrafterSesameClientImpl(httpClient);
+//         }
+//         return this.httpClientManager;
+//     }
+
+
+    @Override public synchronized void setHttpClientSessionManager(HttpClientSessionManager httpMan) {
+        this.httpClientManager = httpMan;
     }
 
+
+//
+//    @Override public synchronized void setSesameClient(SesameClient client) {
+//        this.httpClientManager = client;
+//    }
+
     /**
      * Gets the maximum number of concurrent TCP connections created per route
      * in the HTTP client used by this repository.
@@ -60,6 +75,7 @@ public synchronized void setMaxConcurrentHttpConnections(Integer maxConcurrentHt
         this.setSesameClient(null);
     }
 
+    // Fix for: https://github.com/eclipse/rdf4j/issues/367
     private synchronized HttpClient newHttpClient() {
         //the 'system' HTTP client uses the http.maxConnections
         //system property to define the size of its connection pool (5 is the default

diff --git a/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSesameClientImpl.java b/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSesameClientImpl.java
@@ -1,27 +1,36 @@
 package drafter.rdf;
 
 import org.apache.http.client.HttpClient;
-import org.apache.http.impl.client.HttpClients;
-import org.openrdf.http.client.SesameClientImpl;
-import org.openrdf.http.client.SparqlSession;
+import org.eclipse.rdf4j.http.client.SPARQLProtocolSession;
+import org.eclipse.rdf4j.http.client.SharedHttpClientSessionManager;
+import org.eclipse.rdf4j.http.client.SparqlSession;
 
 import java.util.concurrent.ExecutorService;
 import java.util.concurrent.Executors;
 
-public class DrafterSesameClientImpl extends SesameClientImpl {
+//public class DrafterSesameClientImpl extends SesameClientImpl {
+public class DrafterSesameClientImpl extends SharedHttpClientSessionManager {
 
     private static final ExecutorService QUERY_EXECUTOR = Executors.newCachedThreadPool();
 
     public DrafterSesameClientImpl(HttpClient httpClient) {
         this.setHttpClient(httpClient);
     }
 
-    @Override public synchronized SparqlSession createSparqlSession(String queryEndpointUrl, String updateEndpointUrl) {
-        //NOTE: The two-argument constructor (HttpClient, ExecutorService) of SesameClientImpl does not seem to be
-        //used and the executor is created by the initialize() method called from the default constructor. initialize()
-        //sets the executor to Executors.newCachedThreadPool()
-        //ExecutorService executor = Executors.newCachedThreadPool();
+    @Override public SPARQLProtocolSession createSPARQLProtocolSession(String queryEndpointUrl, String updateEndpointUrl) {
+        DrafterSparqlSession session = new DrafterSparqlSession(queryEndpointUrl, updateEndpointUrl, this.getHttpClient(), QUERY_EXECUTOR);
 
-        return new DrafterSparqlSession(queryEndpointUrl, updateEndpointUrl, this.getHttpClient(), QUERY_EXECUTOR);
+        //session.setQueryURL(queryEndpointUrl);
+        //session.setUpdateURL(updateEndpointUrl);
+        return session;
     }
+
+//    @Override public synchronized SparqlSession createSparqlSession(String queryEndpointUrl, String updateEndpointUrl) {
+//        //NOTE: The two-argument constructor (HttpClient, ExecutorService) of SesameClientImpl does not seem to be
+//        //used and the executor is created by the initialize() method called from the default constructor. initialize()
+//        //sets the executor to Executors.newCachedThreadPool()
+//        //ExecutorService executor = Executors.newCachedThreadPool();
+//
+//        return new DrafterSparqlSession(queryEndpointUrl, updateEndpointUrl, this.getHttpClient(), QUERY_EXECUTOR);
+//    }
 }
diff --git a/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSparqlSession.java b/src-java/drafter_sparql_repository/src/main/java/drafter/rdf/DrafterSparqlSession.java
@@ -13,15 +13,16 @@
 import org.apache.http.params.HttpConnectionParams;
 import org.apache.http.params.HttpParams;
 import org.apache.http.util.EntityUtils;
-import org.openrdf.OpenRDFException;
-import org.openrdf.http.client.SparqlSession;
-import org.openrdf.http.protocol.UnauthorizedException;
-import org.openrdf.http.protocol.error.ErrorInfo;
-import org.openrdf.http.protocol.error.ErrorType;
-import org.openrdf.query.*;
-import org.openrdf.repository.RepositoryException;
-import org.openrdf.rio.RDFParseException;
-import org.openrdf.rio.UnsupportedRDFormatException;
+import org.eclipse.rdf4j.OpenRDFException;
+import org.eclipse.rdf4j.http.client.SPARQLProtocolSession;
+import org.eclipse.rdf4j.http.client.SparqlSession;
+import org.eclipse.rdf4j.http.protocol.UnauthorizedException;
+import org.eclipse.rdf4j.http.protocol.error.ErrorInfo;
+import org.eclipse.rdf4j.http.protocol.error.ErrorType;
+import org.eclipse.rdf4j.query.*;
+import org.eclipse.rdf4j.repository.RepositoryException;
+import org.eclipse.rdf4j.rio.RDFParseException;
+import org.eclipse.rdf4j.rio.UnsupportedRDFormatException;
 
 import java.io.IOException;
 import java.lang.reflect.Field;
@@ -30,12 +31,12 @@
 import java.util.List;
 import java.util.concurrent.ExecutorService;
 
-public class DrafterSparqlSession extends SparqlSession {
+public class DrafterSparqlSession extends SPARQLProtocolSession /*SparqlSession*/ {
     /**
      * The longest URL length accepted by stardog. SPARQL queries which result in a URL longer than this length
      * should be sent as POST requests instead.
      */
-    public static final int STARDOG_MAXIMUM_URL_LENGTH = 4083;
+    //public static final int STARDOG_MAXIMUM_URL_LENGTH = 4083;
 
     public DrafterSparqlSession(String queryEndpointUrl, String updateEndpointUrl, HttpClient client, ExecutorService executor) {
         super(client, executor);
@@ -79,7 +80,7 @@ private HttpParams getHttpParams() {
     }
 
     private HttpClientContext getHttpContext() {
-        return readField(SparqlSession.class, this, "httpContext");
+        return readField(SPARQLProtocolSession.class, this, "httpContext");
     }
 
     /**
@@ -144,9 +145,9 @@ private static void removeTimeoutQueryParams(List<NameValuePair> queryPairs) {
      * @param query The SPARQL query
      * @return Whether to use a POST request to submit the query
      */
-    @Override protected boolean shouldUsePost(String query) {
-        return query.length() > STARDOG_MAXIMUM_URL_LENGTH;
-    }
+    // @Override protected boolean shouldUsePost(String query) {
+    //     return query.length() > STARDOG_MAXIMUM_URL_LENGTH;
+    // }
 
     @SuppressWarnings("deprecation")
     @Override protected HttpResponse execute(HttpUriRequest method) throws IOException, OpenRDFException {

diff --git a/src/drafter/backend/endpoints.clj b/src/drafter/backend/endpoints.clj
@@ -10,7 +10,7 @@
              [protocols :as proto]
              [repository :as repo]]
             [schema.core :as s])
-  (:import org.openrdf.model.URI))
+  (:import org.eclipse.rdf4j.model.URI))
 
 (require 'drafter.backend.repository)
 

diff --git a/src/drafter/backend/repository.clj b/src/drafter/backend/repository.clj
@@ -3,7 +3,7 @@
   type."
   (:require [drafter.backend.protocols :refer :all]
             [drafter.rdf.sesame :as ses])
-  (:import org.openrdf.repository.Repository))
+  (:import org.eclipse.rdf4j.repository.Repository))
 
 (extend-type Repository
   SparqlExecutor

diff --git a/src/drafter/backend/sesame/remote.clj b/src/drafter/backend/sesame/remote.clj
@@ -3,11 +3,11 @@
             [clojure.tools.logging :as log]
             [grafter.rdf.repository.registry :as reg])
   (:import drafter.rdf.DrafterSPARQLRepository
-           [org.openrdf.query.resultio.sparqljson SPARQLBooleanJSONParserFactory SPARQLResultsJSONParserFactory]
-           [org.openrdf.query.resultio.sparqlxml SPARQLBooleanXMLParserFactory SPARQLResultsXMLParserFactory]
-           org.openrdf.query.resultio.text.BooleanTextParserFactory
-           org.openrdf.rio.nquads.NQuadsParserFactory
-           org.openrdf.rio.ntriples.NTriplesParserFactory))
+           [org.eclipse.rdf4j.query.resultio.sparqljson SPARQLBooleanJSONParserFactory SPARQLResultsJSONParserFactory]
+           [org.eclipse.rdf4j.query.resultio.sparqlxml SPARQLBooleanXMLParserFactory SPARQLResultsXMLParserFactory]
+           org.eclipse.rdf4j.query.resultio.text.BooleanTextParserFactory
+           org.eclipse.rdf4j.rio.nquads.NQuadsParserFactory
+           org.eclipse.rdf4j.rio.ntriples.NTriplesParserFactory))
 
 (defn get-required-configuration-setting [var-key config]
   (if-let [ev (var-key config)]

diff --git a/src/drafter/errors.clj b/src/drafter/errors.clj
@@ -13,7 +13,7 @@ overly-specific in your interpretation of the error."
              [errors :refer [encode-error]]
              [responses :as r]]))
 
-(defmethod encode-error org.openrdf.query.QueryEvaluationException [ex]
+(defmethod encode-error org.eclipse.rdf4j.query.QueryEvaluationException [ex]
   ;; This exception should also be caught in specific routes concerned with
   ;; SPARQL endpoints, here we use this to unpack our ConnectionPoolTimeoutException.
 

diff --git a/src/drafter/middleware.clj b/src/drafter/middleware.clj
@@ -10,7 +10,7 @@
             [drafter.user :as user]
             [drafter.user.repository :as user-repo]
             [drafter.rdf.sesame :refer [read-statements] :as ses]
-            [grafter.rdf.io :refer [mimetype->rdf-format]]
+            [grafter.rdf.formats :refer [mimetype->rdf-format]]
             [buddy.auth :as auth]
             [buddy.auth.protocols :as authproto]
             [buddy.auth.backends.httpbasic :refer [http-basic-backend]]

diff --git a/src/drafter/rdf/content_negotiation.clj b/src/drafter/rdf/content_negotiation.clj
@@ -1,8 +1,8 @@
 (ns drafter.rdf.content-negotiation
   (:require [ring.middleware.accept :refer [wrap-accept]]
             [drafter.rdf.formats :refer [csv-rdf-format tsv-rdf-format]])
-  (:import [org.openrdf.query.resultio BooleanQueryResultFormat TupleQueryResultFormat]
-           [org.openrdf.rio RDFFormat]))
+  (:import [org.eclipse.rdf4j.query.resultio BooleanQueryResultFormat TupleQueryResultFormat]
+           [org.eclipse.rdf4j.rio RDFFormat]))
 
 (defn- format-preferences->mime-spec [format-prefs]
   (into {} (mapcat (fn [[f q]] (map (fn [m] [m [f q]]) (.getMIMETypes f))) format-prefs)))