search

bmeg / ophion

Language for making graph queries from data
4 stars 3 forks source link

Ophion differences #21

Open kellrott opened 6 years ago

kellrott commented 6 years ago

We need to resolving differences from the different versions of the Ophion API.

Current version: https://github.com/bmeg/ophion/blob/master/proto/ophion.proto

Proposed Next Version (named AQL to avoid namespace collision) https://github.com/bmeg/arachne/blob/master/aql/aql.proto

buchanae commented 6 years ago
prismofeverything commented 6 years ago

What is the benefit of google.protobuf.ListValue over repeated?

Seems like the service definitions could be ported straight over. The main difference seems to be the addition of a graph parameter to queries to support multiple graphs, which is a pretty straightforward change.

buchanae commented 6 years ago

As I understand it, ListValue exists to support dynamically typed list items. Not sure if it applies here though.

buchanae commented 6 years ago

Seems like potentially the Ophion protobuf doesn't match the actual query schema used by the clients. @prismofeverything Is that true? If so, we'll need to resolve that before AQL can be synced up.

prismofeverything commented 6 years ago

It may have drifted a bit since I wrote the schema, I'll take a look. What differences are you seeing?

On Dec 22, 2017 10:58 AM, "Alex Buchanan" notifications@github.com wrote:

Seems like potentially the Ophion protobuf doesn't match the actual query schema used by the clients. @prismofeverything https://github.com/prismofeverything Is that true? If so, we'll need to resolve that before AQL can be synced up.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/bmeg/ophion/issues/21#issuecomment-353656642, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAjd11vfH_FZGkx4JWZz94xTJ5_mW6Aks5tC_u9gaJpZM4RI2zn .

buchanae commented 6 years ago

I was looking for Traversal.statement.hasLabel. Possibly I'm just reading it wrong though. My plan was to get familiar with the clients so I can get a better feel for the diff.

I guess another approach would be to encode a bunch of queries using only the Protobuf and see what fails.

Is there a definite source of truth I can look at?

prismofeverything commented 6 years ago

I would go with the clients as people are successfully using them to make queries. I can show you in the code too where the queries are processed, but that may be harder to determine.

On Dec 22, 2017 11:14 AM, "Alex Buchanan" notifications@github.com wrote:

I was looking for Traversal.statement.hasLabel. Possibly I'm just reading it wrong though. My plan was to get familiar with the clients so I can get a better feel for the diff.

I guess another approach would be to encode a bunch of queries using only the Protobuf and see what fails.

Is there a definite source of truth I can look at?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/bmeg/ophion/issues/21#issuecomment-353659073, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAjd_-RP2_MuC5-Bnh6NgtV670s_ZQZks5tC_-BgaJpZM4RI2zn .

buchanae commented 6 years ago

Here are the items I propose we resolve. I'm viewing AQL as a minimal, experimental proof-of-concept, rather than a full replacement for Ophion, so the items here are aimed at merging interesting features into Ophion. If that's not true, let's discuss that.

Keep in mind that a proposal should...