bug: labware offsets not loaded when protocol crashes during local run

[bensong04]

Overview

Offsets for labware used in protocols are reset to their defaults whenever a protocol fails to finish running locally after being loaded onto the opentrons app. Note to researchers experiencing similar issues: protocols that run fine on the root installation of python3, and finish with no unexpected behavior on the actual machine, may not necessarily run without crashing locally after being loaded in on the opentrons app. This is because opentrons actually uses its own installation of python3 to testrun protocols. If you're using any non-standard modules in your protocols, make sure to pip install on the opentrons copy of python3.

Steps to reproduce

• write protocol that throws an exception when being run locally (or at all, even)
• load protocol into opentrons app; it should not be able to find offset data
• go through labware setup and save offsets
• cancel run
• reload the same protocol. lab offsets will not have been saved from the previous labware setup.

Current behavior

• opentrons app uses its own installation of python3 to testrun protocols
• opentrons app does not load saved labware offsets when an exception is thrown during a local testrun of a protocol

Expected behavior

• opentrons app should use the root installation of python3instead, since modules are often installed to that installation
• opentrons app should load saved labware offsets regardless of whether or not the testrun finishes successfully

Operating system

Mac

System and robot setup or anything else?

No response