Updating repo so that classes extend/implement those defined at bio4j main repository #1

eparejatobes · 2014-03-17T16:51:25Z

Hey!

We just realized yesterday that we forgot at some point to adapt/update model classes here so that they implement or extends those defined at the main repository by means of the ones in the new and shiny blueprints-specific repository.

So far classes here have been extending those that were defined in the obsolete project BioinfoNeo4j - which were included temporarily in bioinfo-utils repo.

I'm going to create a new set of classes for all this here so that we have something equivalent to what we already got for the Titan implementation.

Let me know if you have any concern/insight on this 😉

eparejatobes · 2014-02-20T08:52:44Z

fine but do it on a different branch

pablopareja · 2014-02-20T08:53:52Z

OK... I like branches so much... 💩

I forgot to ask you, what names should they have?
So far they're called:

BasicEntity (for nodes)
BasicRelationsihp
Neo4jManager

We could keep Neo4jManager name as it is.
For the other two maybe we can change them to: BasicVertex/BasicVertexNeo4j and BasicEdge/BasicEdgeNeo4j ❓

eparejatobes · 2014-02-20T08:54:58Z

why not just Vertex and Edge?

pablopareja · 2014-02-20T09:07:30Z

I'm OK with that but in that case we should probably change those at blueprints repository too:

pablopareja · 2014-02-20T09:09:33Z

I just realized that the edge class had such a funny name... perhaps it's the perfect time to name all this with some coherence along Bio4j repositories... 😄

eparejatobes · 2014-02-20T09:12:23Z

OK

pablopareja · 2014-02-20T10:17:18Z

slight problem here:

http://stackoverflow.com/questions/21903263/how-can-i-convert-a-neo4j-node-to-a-blueprints-vertex/21903483?noredirect=1#21903483

Any ideas?

eparejatobes · 2014-02-20T10:18:13Z

but why this is an issue?

pablopareja · 2014-02-20T10:20:48Z

Since in principle I need to find a way to go from a Neo4j node as parameter for a constructor to the underlying blueprints Vertex object in the case we're extending the BasicVertex class at blueprints repo...

eparejatobes · 2014-02-20T10:26:32Z

mmm I don't know if I understand you. Just make your neo4j nodes Vertices

pablopareja · 2014-02-20T10:27:53Z

but that would mean that we would have to change everything so that we would be using 100% Blueprints stuff ! 😧

eparejatobes · 2014-02-20T10:37:50Z

what? I see two (or more) options here

make the neo4j implement just the abstract model, and forget about blueprints stuff. It could be easy to implement a wrapper on top of it if needed
create a wrapper class implementing both the Blueprints interfaces and the Neo4j stuff
use the Neo4j Blueprints implementation (a bit funky for my taste)

eparejatobes · 2014-02-20T10:41:57Z

footnote: I'm bashing Scala a lot lately, but when you see this things in Java, you get some perspective

pablopareja · 2014-02-20T10:42:42Z

I think I'd go for the first option.
In any case, could you explain a bit more the second one? (I'm not sure I understand what you mean)

eparejatobes · 2014-02-20T10:49:45Z

public abstract class Node implements org.neo4j.graphdb.Node, com.tinkerpop.blueprints.Vertex {
  // blahblahblah
}

pablopareja · 2014-02-20T10:52:20Z

mmm OK
maybe now it'd easier to go for the first option?

eparejatobes · 2014-02-20T10:59:06Z

not only easier but it actually makes more sense, we want to make use of funky neo4j stuff just as labels and the like

laughedelic · 2014-02-20T15:51:06Z

I'm not sure that I understand your conclusions. The blueprints layer will become useless with neo4j?

eparejatobes · 2014-02-20T15:57:30Z

@laughedelic not useless, just not used 👍

laughedelic · 2014-02-20T15:58:01Z

why is it ok?

eparejatobes · 2014-02-20T16:05:41Z

The thing is that in general we will need to change how things are stored at the graph level when implementing technology-specific distributions; given how closely connected to the store level Blueprints is, I don't know if it is possible to have an abstract works-for-everything Blueprints model. And even if it is, that's more or less what the abstract model is for.

Right now, I think of the Blueprints layer as a default works-for-every-graph-db distribution.

laughedelic · 2014-02-20T16:07:26Z

ok

eparejatobes · 2014-02-20T16:20:58Z

anyway, I'm open to all sort of suggestions about all this

pablopareja · 2014-02-24T09:38:51Z

Summing up then, right now we should update neo4j implementation code so that it implements the interfaces defined at bio4j repo. We'll see in the future what we do so that we can have a tighter direct integration with blueprints repo.

Do you both agree on this?

eparejatobes · 2014-02-24T09:39:54Z

👍

laughedelic · 2014-02-24T09:39:56Z

yes, I do

pablopareja · 2014-02-24T09:41:44Z

OK, so I get to work with this

eparejatobes · 2014-03-03T19:08:15Z

I just updated the package names in line with bio4j/model. I will finish this tomorrow; @pablopareja once I have this we should start checking if everything goes well with the import etc

pablopareja · 2014-03-04T08:24:06Z

ok 😉

eparejatobes · 2014-03-17T17:02:20Z

hey @pablopareja what I just pushed contains the abstract part for interacting with Neo4j. I implemented a node (not completely) and a rel in this style: NewTaxonNode and NewTaxonomyRel. The idea is do the same for all nodes and rels. Take a look at the diff and tell me how do you see this

pablopareja · 2014-03-18T08:50:50Z

I cannot get the branch contents.... 😕

ppareja@PPAREJATABLET /c/era7/ppareja/proyectos/JAVA/bio4j-neo4jdb (master)
$ git checkout refactor/classes/implement/model
error: pathspec 'refactor/classes/implement/model' did not match any file(s) known to git.

I also executed this command trying to get the contents into that new branch but I might have ended up putting them overriding my master contents?

ppareja@PPAREJATABLET /c/era7/ppareja/proyectos/JAVA/bio4j-neo4jdb (master)
$ git pull origin refactor/classes/implement/model
remote: Counting objects: 24, done.
remote: Compressing objects: 100% (19/19), done.
remote: Total 24 (delta 1), reused 19 (delta 1)
Unpacking objects: 100% (24/24), done.
From github.com:bio4j/bio4j-neo4jdb

branch refactor/classes/implement/model -> FETCH_HEAD
Updating 8423e31..9d63c72
Fast-forward
build.sbt | 2 +-
.../java/com/bio4j/neo4j/BasicRelationship.java | 2 +-
.../com/bio4j/neo4j/Bio4jRelationshipType.java | 9 +
src/main/java/com/bio4j/neo4j/Neo4jNode.java | 497 ++++++++++++++++++++
.../com/bio4j/neo4j/Neo4jPropertyContainer.java | 19 +
.../java/com/bio4j/neo4j/Neo4jRelationship.java | 35 ++
src/main/java/com/bio4j/neo4j/TypedNeo4jNode.java | 8 +
.../java/com/bio4j/neo4j/model/nodes/CityNode.java | 2 +-
.../com/bio4j/neo4j/model/nodes/NewTaxonNode.java | 121 +++++
.../com/bio4j/neo4j/model/nodes/TaxonNode.java | 3 +-
.../model/relationships/NewTaxonParentRel.java | 22 +
.../com/bio4j/neo4j/model/util/Bio4jManager.java | 4 -
12 files changed, 715 insertions(+), 9 deletions(-)
create mode 100644 src/main/java/com/bio4j/neo4j/Bio4jRelationshipType.java
create mode 100644 src/main/java/com/bio4j/neo4j/Neo4jNode.java
create mode 100644 src/main/java/com/bio4j/neo4j/Neo4jPropertyContainer.java
create mode 100644 src/main/java/com/bio4j/neo4j/Neo4jRelationship.java
create mode 100644 src/main/java/com/bio4j/neo4j/TypedNeo4jNode.java
create mode 100644 src/main/java/com/bio4j/neo4j/model/nodes/NewTaxonNode.java
create mode 100644 src/main/java/com/bio4j/neo4j/model/relationships/NewTaxonParentRel.java

eparejatobes · 2014-03-18T10:05:52Z

git remote update
git checkout <branchname>

And yes, you probably merged that branch into your local master. The fastest way out is (if you don't have any local commits) to make a fresh clone

laughedelic · 2014-03-18T10:54:56Z

or git reset --hard <commit before this pull>

eparejatobes · 2014-03-25T11:49:09Z

@pablopareja related with #6 what do you think of this? it will go in the next release, but I don't want to delay it

eparejatobes · 2014-03-25T11:51:18Z

Anyway, probably better to wait for the abstract model updates in bio4j/bio4j#36

pablopareja self-assigned this Feb 24, 2014

create Neo4j adapter code, samples for how to impl the abstract model

9d63c72

eparejatobes mentioned this pull request Mar 17, 2014

trying not to wrap Neo4j nodes #5

Closed

review code, get rid of commented leftovers

114093f

eparejatobes mentioned this pull request Mar 18, 2014

Relationship here clashes with Neo4j relationships bio4j/bio4j#33

Closed

Updating repo so that classes extend/implement those defined at bio4j main repository #1

Are you sure you want to change the base?

Updating repo so that classes extend/implement those defined at bio4j main repository #1

Conversation

eparejatobes commented Mar 17, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 20, 2014

pablopareja commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

laughedelic commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

laughedelic commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

laughedelic commented Feb 20, 2014

eparejatobes commented Feb 20, 2014

pablopareja commented Feb 24, 2014

eparejatobes commented Feb 24, 2014

laughedelic commented Feb 24, 2014

pablopareja commented Feb 24, 2014

eparejatobes commented Mar 3, 2014

pablopareja commented Mar 4, 2014

eparejatobes commented Mar 17, 2014

pablopareja commented Mar 18, 2014

eparejatobes commented Mar 18, 2014

laughedelic commented Mar 18, 2014

eparejatobes commented Mar 25, 2014

eparejatobes commented Mar 25, 2014