Hi All, I'd like to access the Dagster DAG from outside Dagster (to do some analysis) Where can I find it? I imagine there is some file that is created during the "reload code" action that stores what asset depends on what other assets

It's stored in the database that Dagster uses. If you'd like, the cleanest interface to grab the data you're looking for is the GraphQL API that comes with every Dagster instance. Here's a rudimentary GraphQL query to grab all your assets, its dependencies, and what depends on it:

query GetLineage {
  assetNodes {
    assetKey {
      path
    }
    dependedBy {
      asset {
        assetKey {
          path
        }
      }
    }
    dependencies {
      asset {
        assetKey {
          path
        }
      }
    }
  }
}

Thank you, I'm new to the GraphQL api - do you have a python snippet to pull it?

https://docs.dagster.io/concepts/dagit/graphql they have a pretty good resource on their gql api

Here's a gist I keep on how to query a GraphQL API in Python https://gist.github.com/tacastillo/c4a7a1b55028fcdda1cc07961bd40ba8

@Moody Edghaim - I got as far as

from dagster_graphql import DagsterGraphQLClient
import dagster 
import warnings

warnings.filterwarnings("ignore", category=dagster.ExperimentalWarning)
gql_client = DagsterGraphQLClient('localhost', port_number=3000)

q = '''
query GetLineage {
  assetNodes {
    assetKey {
      path
    }
    dependedBy {
      asset {
        assetKey {
          path
        }
      }
    }
    dependencies {
      asset {
        assetKey {
          path
        }
      }
    }
  }
}
'''

Just trying to figure out how to run the query

tbh, someone from my own team can correct me, but I think the

DagsterGraphQLClient

is a wrapper that runs specific pre-made queries. If you'd like to query the endpoint on your own outside of Dagster, ( such as in another python script or a notebook), then you can copy+paste the python code in the gist I sent you and install its dependencies.

Thank you, I succeeded