Add GraphQL client API

[cafca]

Replaces the "ping" placeholder with a GraphQL API for p2panda clients.

• Add a query endpoint for next entry arguments
• Add tests
• Move request/response structs from rpc module and finally delete that
• Interface with storage provider
• Modules for serialising LogId and SeqNum as String

#60

Next

• Add mutation for publishing entries
• Add query for requesting documents and/or schemas

📋 Checklist

• Add tests that cover your changes
• Add this PR to the Unreleased section in CHANGELOG.md
• Link this PR to any issues it closes
• New files contain a SPDX license header

[sandreae]

Hey, thanks for this, all the changes look good! However I think it makes sense to base this PR off the development branch. As there have been/will be a series of breaking changes (removing JSON rpc etc...) we are merging into there for now. In development you already have access to the initial storage provider implementation, so you would be able to use next_entry_args() and publish_entry() and all other existing higher level storage methods there.

[cafca]

Oops, forgot something

[cafca]

I added a utility in u64_string.rs: it let's you do something like this

pub struct EntryArgsResponse {
    #[serde(with = "log_id_string_serialisation")]
    pub log_id: LogId,

    #[serde(with = "seq_num_string_serialisation")]
    pub seq_num: SeqNum,

    pub backlink: Option<Hash>,

    pub skiplink: Option<Hash>,
}

and then log_id and seq_num are serialised as string values without having to change their type on this struct or having to implement a custom serialiser. I am not sure whether this will be helpful outside of testing, maybe it should also go into p2panda-rs. What do you think @sandreae ?

[sandreae]

This all looks good! Great to have the client GraphQL being introduced and finally completely removing JSON RPC.

Interesting to see what we can do with the string serialisation definitions too 👍

[sandreae]

This is very cool! It's similar to what I was trying to do in the context of storing and retrieving p2panda types from the db where they are mostly strings: #89

I prefer your approach as it seems more generally useful. I don't know if it is a drop-in solution to for the db stuff though. Definitely let's keep an eye on it and see if we want to implement it in p2panda-rs and then try to use for the db models.