fancy-mouse-14245
01/10/2023, 1:17 PMancient-application-36103
01/10/2023, 4:39 PMancient-application-36103
01/10/2023, 4:40 PMfancy-mouse-14245
01/10/2023, 6:06 PMcalm-smartphone-49719
01/12/2023, 3:11 PM.flow/<flow_name>/<runid> in their start-step, which then lets you do something like
python flow_a.py run && python flow_b.py run
flow_b.py will then interrogate .flow/ to look for the highest .flow/FlowA<runid> file and pull the run-id, then construct Run('FlowA/<runid>') using the client API and from there all FlowA data is available.
We also standardise on a outputs_dict artefact on all flows we want to use this pattern with, so which makes figuring out "what outputs are available" a bit easier.calm-smartphone-49719
01/12/2023, 3:12 PM