Feature Request: Graph Database

### Is your proposal related to a problem?

I hate join tables. They clutter everything and complicate db architecture a lot. They grind to a halt during friends-of-friends-of-friends queries.

I'm unsure of where their performance breaks down, but I love the idea of relationships being first-class citizens.

### Describe the solution you'd like

I think graph dbs are a great solution to a plethora of db architecture decisions but no one in the space seems to be doing it right. Neo4j seems slow compared to Dgraph, doesn't offer a nice GraphQL experience out of the box, and is slow to adopt features (they just got some half-baked version of subscriptions).

Dgraph has a _mountain_ of drama surrounding it, it seems development on it has slowed to a crawl, and the UI & their custom GraphQL query syntax are pretty ugly.

AWS Neptune is complicated by virtue of AWS being an entire... thing... And the Gremlin query lang is kind of gross.

I think Hasura can, and should, take over the entire graph db space.

### Describe alternatives you've considered

There is no alternative. I've kind of just been forced to stay on Postgres for my entire life.

### If the feature is approved, would you be willing to submit a PR?

Heck no. It seems like a lot of complicated work suited for a bunch of Haskell & db geniuses like yourselves. I would be willing to work on designs, beta test for as long as needed, and help brainstorm how this new offering could coincide with Hasura's existing services seamlessly.

I don't have high hopes for this happening, but I have been daydreaming of Hasura + graph dbs for years, so I figured I'd at least open a request for it on the off chance @coco98 sees this and thinks to himself _"this could be a billion dollar play..."_

Love you guys and will continue using/promoting your service wherever I can. ❤️