How is Memgraph dealing with deduplication? Especially between restarts. I've read the blog post but I couldn't find details on this topic.
When data is transferred from Memgraph to Elasticsearch asynchronously (i.e., the user explicitly invokes a method to transmit data), there is no issue to discuss. However, if a user aims to synchronize Memgraph with Elasticsearch using triggers and Memgraph experiences a restart, the identical data will be serialized and transmitted to Elasticsearch. This will overwrite any previous data in Elasticsearch, as both vertex and edge IDs are employed as document IDs and preserved during restarts through snapshots. It is important to note that Elasticsearch internally versions the data, but always presents the most recent version.