[SL-2323] [Feature] Update `dbt-metricflow` dependencies to use `dbt-*` 1.8 packages about metricflow HOT 4 CLOSED

Whilst this is being worked on, what is the workaround please for successfully being able to run metric flow commands such as mf list metrics with a codebase.

I created a vanilla virtual environment where I installed just dbt-metricflow but running the mf list metrics commands complains about missing adapter. I guess dbt-metricflow is technically standalone but indirectly, it is not when you consider real world of what profiles.yml and artifact files look like. So then I went through a process of whack-a-mole installing dbt-adapters and then dbt-common etc. etc. to try and get to a place where I could run the command. Eventually, I installed everything together (incl dbt-core) but then you get the profobuf issue (TypeError: Couldn't build proto file into descriptor pool: duplicate file name types.proto) which is the one I assume updating dbt-metricflow's dependencies to not be incompatible with dbt 1.8 will solve.

But in the meantime, how can I get a stable environment?

from metricflow.

@plypaul please can you advise when a new version of dbt-metricflow will be released on pypi with this critical fix?

from metricflow.

@adamcunnington-mlg Sorry, missed your comment earlier. I tried a test run with a clean virtual environment, and things seemed to work with what's currently on master. Mind sharing the commands that you ran so I can reproduce it?

from metricflow.

@plypaul thank you - the problem can be reproduced by installing dbt-core and dbt-bigquery for 1.8 and then subsequently installing dbt-metricflow which will force a downgrade of dbt-core and then you'll run into a low-level Protobuf issue, basically caused by incompatibilities between the adapter's dependencies and dbt-core/dbt-metricflow.

You can get around this by doing pip install dbt-metricflow[bigquery] which will have the effect of downgrading both but the way I'd expect this to work is that dbt-metricflow's dependencies are more dynamic and controlled by my existing environment (dbt-core & dbt-bigquery).

from metricflow.