Just sharing the updates.
I made the WIP version for google datastore on
https://github.com/SteamShon/incubator-s2graph/tree/datastore.
To test this please reference on DatastoreStorageTest.
I tested it only on my local environment as follow.
1. startup local datastore HTTP server using emulators.
gcloud beta emulators datastore start --host-port localhost:8080
--consistency 1.0 --project async-test --data-dir project-test
2. checkout above branch.
3. startup s2graphql server using following options.
sbt -DschemaCacheTTL=-1 -Dhttp.port=8000 -DenableMutation=true
-Ds2graph.storage.backend=datastore -Ddatastore.host="http://localhost:8080;
-Ddatastore.dataset=async-test -Ddatastore.version=v1beta3
-Ddatastore.namespace=test 'project s2graphql' '~re-start'
then you can follow up
https://github.com/apache/incubator-s2graph/tree/master/s2graphql
If anyone with interest and time, please review. I would appreciate any
feedback.
On Sat, Nov 3, 2018 at 7:24 AM DO YUNG YOON wrote:
> Hi Hwansung.
>
> I think implementing DatastoreFetcher/DatastoreMutator,
> DynamoDBFetcher/DynamoDBMutator class which implement our Fetcher/Mutator
> interface will help you to work with data in that storage.
>
> The purpose of exposing Fetcher/Mutator interface is for others to
> implement in their way to work with their storage, only focusing on how to
> fetch vertex/edge from storage when the S2Query is given, and how to store
> vertex/edge to storage when S2Edge/S2Vertex are given.
>
> Also if you are deploying s2graph rest on the public cloud, then we can
> provide docker image for s2graph rest(not sure if we are already providing
> it, let me check it out).
>
> I have very little experience on the public cloud, such as GCP and AWS, so
> I might think the wrong direction. I guess we can list what are the
> necessary parts, then create issues regarding it.
>
> Please feel free to advise on my comments.
>
> On Fri, Nov 2, 2018 at 10:09 AM Hwansung Yu wrote:
>
>> Hi, folks.
>>
>> I have recently been working in a new environment.
>> I'm considering running s2graph on the Google Cloud Platform (GCP) and
>> AWS.
>> But operating HBase in a cloud platform environment is not easy.
>> Of course, AWS provides Hbase based on EMR, but if it supports
>> nosql basically provided such as DynamoDB or Datastore(GCP), is it
>> possible to operate s2graph more easily in the cloud environment?
>>
>> Regards.
>>
>