Scott Peters

06/28/2021, 4:20 PM
I have a fairy abstract question here, but maybe someone has some thoughts to point me in the right direction. Here are the take away questions ( this context to follow ) 1. is there any notion of a
dagster client
( maybe the graphql client could be this? ), for interacting with dagster ( submit, get output ) without having direct local access to
? ( currently, the
graphql client
is fairly limited
) 2. (related) ... what is the appropriate way to get output from a pipeline to other non-dagster python tools... ( aka - pipeline creates an s3 bucket that another script needs to send that s3 path back to the user )


06/28/2021, 5:59 PM
The graphql client is just a wrapper around some basic operations of our graphql api. Any metadata or operations done by dagit can also be done through this api - if you’re looking for other operations that aren’t currently supported by the graphql client, feel free to add a feature request or use the graphql api directly by constructing your own custom queries.