[jira] [Commented] (FLINK-2033) Add overloaded methods with explicit TypeInformation parameters to Gelly

2015-10-20 Thread PJ Van Aeken (JIRA)

[ 
https://issues.apache.org/jira/browse/FLINK-2033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=14964914#comment-14964914
 ] 

PJ Van Aeken commented on FLINK-2033:
-

Totally forgot about this one, sorry. Yes it's ok now. Gelly has the explicit 
declarations.

> Add overloaded methods with explicit TypeInformation parameters to Gelly
> 
>
> Key: FLINK-2033
> URL: https://issues.apache.org/jira/browse/FLINK-2033
> Project: Flink
>  Issue Type: Task
>  Components: Gelly
>Affects Versions: 0.9
>Reporter: PJ Van Aeken
>Assignee: PJ Van Aeken
> Fix For: 0.10
>
>
> For the implementation of the Scala API for Gelly (FLINK-1962), we need to 
> pass explicit TypeInformation since the Java TypeExtractor does not work for 
> all Scala Types (see FLINK-2023).
> To do this, the java Gelly API needs to be expanded with methods that allow 
> for explicit passing of TypeInformation.
> An example with mapVertices:
> {code}
>  public  Graph mapVertices(final MapFunction, 
> NV> mapper) {
> TypeInformation keyType = ((TupleTypeInfo) 
> vertices.getType()).getTypeAt(0);
> String callLocation = Utils.getCallLocationName();
> TypeInformation valueType = 
> TypeExtractor.getMapReturnTypes(mapper, vertices.getType(), callLocation, 
> false);
> TypeInformation> returnType = 
> (TypeInformation>) new TupleTypeInfo(
> Vertex.class, keyType, valueType);
> return mapVertices(mapper,returnType);
> }
> public  Graph mapVertices(final MapFunction, 
> NV> mapper, TypeInformation> returnType) {
> DataSet> mappedVertices = vertices.map(
> new MapFunction, Vertex>() {
> public Vertex map(Vertex value) throws 
> Exception {
> return new Vertex(value.f0, mapper.map(value));
> }
> }).returns(returnType);
> return new Graph(mappedVertices, this.edges, this.context);
> }
> {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (FLINK-2033) Add overloaded methods with explicit TypeInformation parameters to Gelly

2015-08-22 Thread Vasia Kalavri (JIRA)

[ 
https://issues.apache.org/jira/browse/FLINK-2033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14708099#comment-14708099
 ] 

Vasia Kalavri commented on FLINK-2033:
--

I can close this one now, right [~vanaepi]?

 Add overloaded methods with explicit TypeInformation parameters to Gelly
 

 Key: FLINK-2033
 URL: https://issues.apache.org/jira/browse/FLINK-2033
 Project: Flink
  Issue Type: Task
  Components: Gelly
Affects Versions: 0.9
Reporter: PJ Van Aeken
Assignee: PJ Van Aeken

 For the implementation of the Scala API for Gelly (FLINK-1962), we need to 
 pass explicit TypeInformation since the Java TypeExtractor does not work for 
 all Scala Types (see FLINK-2023).
 To do this, the java Gelly API needs to be expanded with methods that allow 
 for explicit passing of TypeInformation.
 An example with mapVertices:
 {code}
  public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper) {
 TypeInformationK keyType = ((TupleTypeInfo?) 
 vertices.getType()).getTypeAt(0);
 String callLocation = Utils.getCallLocationName();
 TypeInformationNV valueType = 
 TypeExtractor.getMapReturnTypes(mapper, vertices.getType(), callLocation, 
 false);
 TypeInformationVertexK, NV returnType = 
 (TypeInformationVertexK, NV) new TupleTypeInfo(
 Vertex.class, keyType, valueType);
 return mapVertices(mapper,returnType);
 }
 public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper, TypeInformationVertexK, NV returnType) {
 DataSetVertexK, NV mappedVertices = vertices.map(
 new MapFunctionVertexK, VV, VertexK, NV() {
 public VertexK, NV map(VertexK, VV value) throws 
 Exception {
 return new VertexK, NV(value.f0, mapper.map(value));
 }
 }).returns(returnType);
 return new GraphK, NV, EV(mappedVertices, this.edges, this.context);
 }
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (FLINK-2033) Add overloaded methods with explicit TypeInformation parameters to Gelly

2015-06-10 Thread PJ Van Aeken (JIRA)

[ 
https://issues.apache.org/jira/browse/FLINK-2033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14580225#comment-14580225
 ] 

PJ Van Aeken commented on FLINK-2033:
-

If the PR for FLINK-1962 gets accepted, this can be marked as resolved

 Add overloaded methods with explicit TypeInformation parameters to Gelly
 

 Key: FLINK-2033
 URL: https://issues.apache.org/jira/browse/FLINK-2033
 Project: Flink
  Issue Type: Task
  Components: Gelly
Affects Versions: 0.9
Reporter: PJ Van Aeken
Assignee: PJ Van Aeken

 For the implementation of the Scala API for Gelly (FLINK-1962), we need to 
 pass explicit TypeInformation since the Java TypeExtractor does not work for 
 all Scala Types (see FLINK-2023).
 To do this, the java Gelly API needs to be expanded with methods that allow 
 for explicit passing of TypeInformation.
 An example with mapVertices:
 {code}
  public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper) {
 TypeInformationK keyType = ((TupleTypeInfo?) 
 vertices.getType()).getTypeAt(0);
 String callLocation = Utils.getCallLocationName();
 TypeInformationNV valueType = 
 TypeExtractor.getMapReturnTypes(mapper, vertices.getType(), callLocation, 
 false);
 TypeInformationVertexK, NV returnType = 
 (TypeInformationVertexK, NV) new TupleTypeInfo(
 Vertex.class, keyType, valueType);
 return mapVertices(mapper,returnType);
 }
 public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper, TypeInformationVertexK, NV returnType) {
 DataSetVertexK, NV mappedVertices = vertices.map(
 new MapFunctionVertexK, VV, VertexK, NV() {
 public VertexK, NV map(VertexK, VV value) throws 
 Exception {
 return new VertexK, NV(value.f0, mapper.map(value));
 }
 }).returns(returnType);
 return new GraphK, NV, EV(mappedVertices, this.edges, this.context);
 }
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (FLINK-2033) Add overloaded methods with explicit TypeInformation parameters to Gelly

2015-05-26 Thread PJ Van Aeken (JIRA)

[ 
https://issues.apache.org/jira/browse/FLINK-2033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14559213#comment-14559213
 ] 

PJ Van Aeken commented on FLINK-2033:
-

I will add overloaded methods to the Java API on a need-to-have basis for 
implementing the Scala Gelly API. Progress can be followed here: 

https://github.com/PieterJanVanAeken/flink/tree/scala-gelly-api/

 Add overloaded methods with explicit TypeInformation parameters to Gelly
 

 Key: FLINK-2033
 URL: https://issues.apache.org/jira/browse/FLINK-2033
 Project: Flink
  Issue Type: Task
  Components: Gelly
Affects Versions: 0.9
Reporter: PJ Van Aeken
Assignee: PJ Van Aeken

 For the implementation of the Scala API for Gelly (FLINK-1962), we need to 
 pass explicit TypeInformation since the Java TypeExtractor does not work for 
 all Scala Types (see FLINK-2023).
 To do this, the java Gelly API needs to be expanded with methods that allow 
 for explicit passing of TypeInformation.
 An example with mapVertices:
 {code}
  public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper) {
 TypeInformationK keyType = ((TupleTypeInfo?) 
 vertices.getType()).getTypeAt(0);
 String callLocation = Utils.getCallLocationName();
 TypeInformationNV valueType = 
 TypeExtractor.getMapReturnTypes(mapper, vertices.getType(), callLocation, 
 false);
 TypeInformationVertexK, NV returnType = 
 (TypeInformationVertexK, NV) new TupleTypeInfo(
 Vertex.class, keyType, valueType);
 return mapVertices(mapper,returnType);
 }
 public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper, TypeInformationVertexK, NV returnType) {
 DataSetVertexK, NV mappedVertices = vertices.map(
 new MapFunctionVertexK, VV, VertexK, NV() {
 public VertexK, NV map(VertexK, VV value) throws 
 Exception {
 return new VertexK, NV(value.f0, mapper.map(value));
 }
 }).returns(returnType);
 return new GraphK, NV, EV(mappedVertices, this.edges, this.context);
 }
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (FLINK-2033) Add overloaded methods with explicit TypeInformation parameters to Gelly

2015-05-18 Thread Aljoscha Krettek (JIRA)

[ 
https://issues.apache.org/jira/browse/FLINK-2033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14547985#comment-14547985
 ] 

Aljoscha Krettek commented on FLINK-2033:
-

Yes, I think this is the right way to go.

 Add overloaded methods with explicit TypeInformation parameters to Gelly
 

 Key: FLINK-2033
 URL: https://issues.apache.org/jira/browse/FLINK-2033
 Project: Flink
  Issue Type: Task
  Components: Gelly
Affects Versions: 0.9
Reporter: PJ Van Aeken

 For the implementation of the Scala API for Gelly (FLINK-1962), we need to 
 pass explicit TypeInformation since the Java TypeExtractor does not work for 
 all Scala Types (see FLINK-2023).
 To do this, the java Gelly API needs to be expanded with methods that allow 
 for explicit passing of TypeInformation.
 An example with mapVertices:
 {code}
  public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper) {
 TypeInformationK keyType = ((TupleTypeInfo?) 
 vertices.getType()).getTypeAt(0);
 String callLocation = Utils.getCallLocationName();
 TypeInformationNV valueType = 
 TypeExtractor.getMapReturnTypes(mapper, vertices.getType(), callLocation, 
 false);
 TypeInformationVertexK, NV returnType = 
 (TypeInformationVertexK, NV) new TupleTypeInfo(
 Vertex.class, keyType, valueType);
 return mapVertices(mapper,returnType);
 }
 public NV GraphK, NV, EV mapVertices(final MapFunctionVertexK, VV, 
 NV mapper, TypeInformationVertexK, NV returnType) {
 DataSetVertexK, NV mappedVertices = vertices.map(
 new MapFunctionVertexK, VV, VertexK, NV() {
 public VertexK, NV map(VertexK, VV value) throws 
 Exception {
 return new VertexK, NV(value.f0, mapper.map(value));
 }
 }).returns(returnType);
 return new GraphK, NV, EV(mappedVertices, this.edges, this.context);
 }
 {code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)