Data API for Jupyter notebooks

[rizo]

Create a POC implementation of a Jupiter notebook fetching the data from our benchmarking backend.

[rizo]

@gs0510 have you started any work on this?

[gs0510]

I did some for tezos benchmarks but that's unrelated to fetching data from the benchmarking backend. I can work on this after getting the tree benchmarks in. :)

[Deepthi562]

@rizo hi!. I am an outreachy applicant. Can I be assigned with this?

[art-w]

Hello @Deepthi562 ! Thanks you for your interest in this issue :)

The outreachy mentors are currently on holidays until monday, we might have to check with them that this is a good fit (cc @Sudha247 @shakthimaan @gs0510 ) But I would be glad if you want to work on this! Feel free to ask any questions you might have regarding Jupyter or the GraphQL api to get access to the benchmarks results (or anything else).

You probably want to start by looking at the network requests to https://autumn.ocamllabs.io/_hasura/v1/graphql on the website to understand which data are available, and try to reproduce those requests from a Jupyter notebook :)

[Sudha247]

Hey @Deepthi562, thanks for your interest! Thank you @art-w for the details.

I think this issue is a bit open ended, I'm happy for you continue working on this issue if you have some idea about how to proceed along the lines @art-w mentioned. Given the initial contribution period for outreachy ends in a few days, if you're looking for a straightforward one you might want to have a look at https://github.com/ocaml-bench/sandmark-nightly/issues/14.

[Deepthi562]

@Sudha247, thanks for the message. Could you please brief me more about the OCaml issue you mentioned? I am unable to understand that. Thank you.

[gs0510]

@Deepthi562 You can ask any question you have about the ocaml-bench issue on the issue link itself. (here: https://github.com/ocaml-bench/sandmark-nightly/issues/14)

[gs0510]

To add more context to this issue, we already have a graphql endpoint, something to explore would be how to host a jupyter notebook online that can fetch the data from this endpoint so if someone wants to make any kind of graphs they can just go to that website. It might involve some wrangling with Jupyterhub. :)