Much simpler way to implement required UUIDs.

[cheerfulstoic]

Just remove the methods/properties/constants every time it's called again. Thanks @subvertallchris !

[coveralls]

Coverage decreased (-0.15%) when pulling 1c37d85 on require_uuid into cdd0b69 on master.

[subvertallchris]

Simpler AND it can be merged automatically! 👏

Though I still think merge (or at least a release with this) should be accompanied be a rake task to add IDs. I can give that a shot tomorrow, if you want.

[subvertallchris]

Boom! Automating the process of adding IDs to existing nodes was trickier than I imagined. I wrote a bit about it at http://stackoverflow.com/questions/25952884/neo4j-assign-unique-values-to-all-nodes-matching-query/25954136#25954136.

[andreasronge]

@subvertallchris really cool

[coveralls]

Coverage decreased (-0.15%) when pulling fa71907 on require_uuid into cdd0b69 on master.

[coveralls]

Coverage decreased (-0.15%) when pulling 99105d9 on require_uuid into a00d124 on master.

[coveralls]

Coverage decreased (-0.15%) when pulling 99105d9 on require_uuid into a00d124 on master.

[coveralls]

Coverage decreased (-0.15%) when pulling de559ba on require_uuid into a00d124 on master.

[subvertallchris]

Just so I'm getting it right, it's dropping the constraint at load if one
is already defined, right? And that would happen if it's changed or if it's
inherited and then the subclass has its id property changed?

On Sunday, September 21, 2014, Coveralls notifications@github.com wrote:

[image: Coverage Status] https://coveralls.io/builds/1236692

Coverage decreased (-0.15%) when pulling de559ba
de559ba
on require_uuid into a00d124
a00d124
on master.

—
Reply to this email directly or view it on GitHub
#478 (comment).

[cheerfulstoic]

I'm pretty sure it's dropping the constraint just for the class/label concerned. Inheriting and calling id_property again should just act like more calls to id_property (which at each point drop the old constraint and undefined the old property)

[coveralls]

Coverage decreased (-0.05%) when pulling 43c5f0b on require_uuid into a00d124 on master.

[cheerfulstoic]

If these tests pass (which I believe they will) I'm going to merge it!

[subvertallchris]

But when does it drop it? It looks like every time the model is loaded if there was already an id property defined, right? Does that force a wipe and rebuild of all indexes, too, since indexes follow labels?

EDIT: Wait, that's a little drunk logic talking. I think it might have to reindex the ID properties of all nodes with that label, though.

[coveralls]

Coverage decreased (-0.05%) when pulling cb10870 on require_uuid into a00d124 on master.

[cheerfulstoic]

Yeah, that's a good point ;) I'll look at that now. I would think part of the problem is that there's no way to query for constraints (except, maybe, a separate endpoint), but maybe it's possible to make a single query which doesn't change anything if the constraint is already there.

[subvertallchris]

If you look in at data/graph.db/messages.log, it shows when it's dropping and rebuilding indexes.

If dropping the old indexes is just to try keeping the database a little cleaner, maybe that should be left alone for now and the user can maintain that on their own?

[cheerfulstoic]

Ok, just made a change so that it shouldn't remove the constraint if it already exists. Merging!

[coveralls]

Coverage decreased (-0.05%) when pulling cb6695e on require_uuid into a00d124 on master.

[subvertallchris]

That is brilliant! I didn't think you could get constraint info from
Cypher. Great work!

On Sunday, September 21, 2014, Coveralls notifications@github.com wrote:

[image: Coverage Status] https://coveralls.io/builds/1236772

Coverage decreased (-0.05%) when pulling cb6695e
cb6695e
on require_uuid into a00d124
a00d124
on master.

—
Reply to this email directly or view it on GitHub
#478 (comment).