[
https://issues.apache.org/jira/browse/TINKERPOP-1438?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15943330#comment-15943330
]
ASF GitHub Bot commented on TINKERPOP-1438:
-------------------------------------------
Github user dpitera commented on a diff in the pull request:
https://github.com/apache/tinkerpop/pull/569#discussion_r108179063
--- Diff:
gremlin-server/src/main/java/org/apache/tinkerpop/gremlin/server/GraphManager.java
---
@@ -21,139 +21,98 @@
import org.apache.tinkerpop.gremlin.process.traversal.TraversalSource;
import org.apache.tinkerpop.gremlin.structure.Graph;
import org.apache.tinkerpop.gremlin.structure.Transaction;
-import org.apache.tinkerpop.gremlin.structure.util.GraphFactory;
-import org.apache.tinkerpop.gremlin.util.iterator.IteratorUtils;
-import org.slf4j.Logger;
-import org.slf4j.LoggerFactory;
import javax.script.Bindings;
-import javax.script.SimpleBindings;
-import java.util.HashSet;
import java.util.Map;
import java.util.Set;
-import java.util.concurrent.ConcurrentHashMap;
-import java.util.function.Predicate;
-
-/**
- * Holder for {@link Graph} and {@link TraversalSource} instances
configured for the server to be passed to script
- * engine bindings. The {@link Graph} instances are read from the {@link
Settings} for Gremlin Server as defined in
- * the configuration file. The {@link TraversalSource} instances are
rebound to the {@code GraphManager} once
- * initialization scripts construct them.
- */
-public final class GraphManager {
- private static final Logger logger =
LoggerFactory.getLogger(GremlinServer.class);
-
- private final Map<String, Graph> graphs = new ConcurrentHashMap<>();
- private final Map<String, TraversalSource> traversalSources = new
ConcurrentHashMap<>();
+import java.util.function.Supplier;
+public interface GraphManager {
/**
- * Create a new instance using the {@link Settings} from Gremlin
Server.
+ * Get a list of the {@link Graph} instances and their binding names
+ *
+ * @return a {@link Map} where the key is the name of the {@link
Graph} and the value is the {@link Graph} itself
*/
- public GraphManager(final Settings settings) {
- settings.graphs.entrySet().forEach(e -> {
- try {
- final Graph newGraph = GraphFactory.open(e.getValue());
- graphs.put(e.getKey(), newGraph);
- logger.info("Graph [{}] was successfully configured via
[{}].", e.getKey(), e.getValue());
- } catch (RuntimeException re) {
- logger.warn(String.format("Graph [%s] configured at [%s]
could not be instantiated and will not be available in Gremlin Server.
GraphFactory message: %s",
- e.getKey(), e.getValue(), re.getMessage()), re);
- if (re.getCause() != null) logger.debug("GraphFactory
exception", re.getCause());
- }
- });
- }
+ public Map<String, Graph> getGraphs();
--- End diff --
> Seems like we would want to use GraphManager to enforce the logic of the
implementation, thus getGraphs() should go away?
Agreed.
I am confused by your suggestion to "lessen the breaking change", but I'll
give a stab at what I think you meant, and we can reiterate
> Consider GraphManager as an interface
> -------------------------------------
>
> Key: TINKERPOP-1438
> URL: https://issues.apache.org/jira/browse/TINKERPOP-1438
> Project: TinkerPop
> Issue Type: Improvement
> Components: server
> Affects Versions: 3.2.2
> Reporter: stephen mallette
> Priority: Minor
> Labels: breaking
>
> If {{GraphManager}} were an interface it would make embedding Gremlin Server
> easier as {{Graph}} instances could be more easily supplied by the host
> application. In doing this, It also might be good to force a
> {{TraversalSource}} to be referred to by both the {{Graph}} name and
> {{TraversalSource}} name.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
