Package Usage

[EyalAbadi]

Hello,

I'm trying to understand how to use this package. Do I need to have a regular GraphQL running app and then use this package there or is this standalone?

If so, how do I integrate it?

Thanks!

[coady]

I'm trying to understand how to use this package. Do I need to have a regular GraphQL running app and then use this package there or is this standalone?

It's standalone. pip install graphique[server] includes everything needed to run it as a service: env PARQUET_PATH=... uvicorn graphique.service:app

If so, how do I integrate it?

It could also be used as a package, but I'd have to know more about the use case.

[EyalAbadi]

Hi Coady,

I'm trying to expose a parquet file stored on disk and expose it through a GraphQL type of API. Allowing the client to specify their required schema and return/filter the files. I'm trying to stand up a basic instance of it, but getting confused as I'm used to using something like apollo with its specific graphql terms and structure.

[coady]

For apollo federation, the schema can be output for composition with: python3 -m graphique.schema <parquet-path>.

[coady]

b677fb9 adds a FEDERATED configuration option.

[coady]

v0.7 releases the federated option.

But I am also interested in this project being usable as extensible package: multiple tables, multiple indexes, and custom fields. Converting to a disussion for follow-up.

[coady]

v1.0 is more extensible. A custom GraphQL app can be created, as in tests/federated.py. Or it can be used as a package, extending the Dataset interface.