Hello Flume community, Wanted to bring up a few things regarding design and implementation of the graphsink: 1) I am currently investigating the use of spring-data-neo4j to wire in the infrastucture code to that adds nodes, repositories and relationships, I was wondering if there are strong opinions or objections to wiring in any spring dependencies against any sink, if it turns out this is adding too many layers I will just use the java rest binding (https://github.com/neo4j/java-rest-binding) 2) For the initial release I am embedding the schema for the graph database inside a json file and parsing that to construct the database/nodes and relationships, an example of this is shown below towards the bottom of the page:
https://github.com/moviepilot/neoscout the workflow behind the graphsink would read the chunk of json, construct the graph db ndoes and relationships and then wait for future events to perform CRUD on nodes and edges 3) As mentioned in the jira ticket the basic functionality of the sink would be to handle events for db initialization, node and edge creation and updates and deletion 4) For the first release I am going to be strictly focusing on neo4j I am planning on shipping the first release of this post 1.4 but wanted to get some feedback or additional thoughts on features, look forward to hearing from folks. Regards
