[jira] [Commented] (TINKERPOP-1578) Support writable distributed graphs using spark as the underlying engine
[ https://issues.apache.org/jira/browse/TINKERPOP-1578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15728804#comment-15728804 ] Dovid Kopel commented on TINKERPOP-1578: My approach added on the `spark-gremlin` module. I provided mutations using spark as the underlying storage mechanism. If one wants to use the `SparkGraph` they may do so without using the graph computer directly. If you look at my `SparkTest#testSparkGraph()` test you can see this functional. > Support writable distributed graphs using spark as the underlying engine > > > Key: TINKERPOP-1578 > URL: https://issues.apache.org/jira/browse/TINKERPOP-1578 > Project: TinkerPop > Issue Type: Improvement > Components: hadoop >Affects Versions: 3.2.3 >Reporter: Dovid Kopel > -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TINKERPOP-1578) Support writable distributed graphs using spark as the underlying engine
[ https://issues.apache.org/jira/browse/TINKERPOP-1578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15728642#comment-15728642 ] Marko A. Rodriguez commented on TINKERPOP-1578: --- Your work only works for {{SparkGraphComputer}}. You made your changes at the Spark-level, not at the {{GraphComputer}}-level. Thus, its not general. We need to be able to have OLAP mutations working for ANY {{GraphComputer}} implementation: {{GiraphGraphComputer}}, {{TinkerGraphComputer}}, {{SparkGraphComputer}}, etc. This means figuring out the interface additions and {{GraphComputerTest}} infrastructure needed to validate the semantics of OLAP mutations across all {{GraphComputer}} implementations. Its a massive amount of work, but would be a huge contribution. As your contribution stands right now, it can't be merged as its not provider-agnostic. > Support writable distributed graphs using spark as the underlying engine > > > Key: TINKERPOP-1578 > URL: https://issues.apache.org/jira/browse/TINKERPOP-1578 > Project: TinkerPop > Issue Type: Improvement > Components: hadoop >Affects Versions: 3.2.3 >Reporter: Dovid Kopel > -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TINKERPOP-1578) Support writable distributed graphs using spark as the underlying engine
[ https://issues.apache.org/jira/browse/TINKERPOP-1578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15728436#comment-15728436 ] Dovid Kopel commented on TINKERPOP-1578: The changes I made didn't affect the `GraphComputer`...can you explain what you mean? > Support writable distributed graphs using spark as the underlying engine > > > Key: TINKERPOP-1578 > URL: https://issues.apache.org/jira/browse/TINKERPOP-1578 > Project: TinkerPop > Issue Type: Improvement > Components: hadoop >Affects Versions: 3.2.3 >Reporter: Dovid Kopel > -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TINKERPOP-1578) Support writable distributed graphs using spark as the underlying engine
[ https://issues.apache.org/jira/browse/TINKERPOP-1578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15712441#comment-15712441 ] Dovid Kopel commented on TINKERPOP-1578: So far I haven't touched graph computer at allit's the original one Dovid Kopel do...@dovidkopel.com 443-291-9212 Sent from my Android. > Support writable distributed graphs using spark as the underlying engine > > > Key: TINKERPOP-1578 > URL: https://issues.apache.org/jira/browse/TINKERPOP-1578 > Project: TinkerPop > Issue Type: Improvement > Components: hadoop >Affects Versions: 3.2.3 >Reporter: Dovid Kopel > -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TINKERPOP-1578) Support writable distributed graphs using spark as the underlying engine
[ https://issues.apache.org/jira/browse/TINKERPOP-1578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15711896#comment-15711896 ] Marko A. Rodriguez commented on TINKERPOP-1578: --- Can you generalize to make work for {{GraphComputer}} and not just {{SparkGraphComputer}} ? This way its not a Spark-specific thing? > Support writable distributed graphs using spark as the underlying engine > > > Key: TINKERPOP-1578 > URL: https://issues.apache.org/jira/browse/TINKERPOP-1578 > Project: TinkerPop > Issue Type: Improvement > Components: hadoop >Affects Versions: 3.2.3 >Reporter: Dovid Kopel > -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (TINKERPOP-1578) Support writable distributed graphs using spark as the underlying engine
[ https://issues.apache.org/jira/browse/TINKERPOP-1578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15709687#comment-15709687 ] Dovid Kopel commented on TINKERPOP-1578: https://github.com/dovidkopel/tinkerpop/tree/TINKERPOP-1578 I spent around two days and have basic functionality. I did not make a pull request yet. I haven't created ample tests and coverage. I intentionally abstracted the RDD storage from the graph to allow for future changes. I had fully functional traversals using the built-in strategies. Please let me know what you think so far. If you think this is a reasonable approach I can clean things up and make an official pull request. > Support writable distributed graphs using spark as the underlying engine > > > Key: TINKERPOP-1578 > URL: https://issues.apache.org/jira/browse/TINKERPOP-1578 > Project: TinkerPop > Issue Type: New Feature >Reporter: Dovid Kopel > -- This message was sent by Atlassian JIRA (v6.3.4#6332)