You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We previously attempted to improve the error reporting of the Psi4Harness via #266. However, now most of the errors we observe in production use of the Psi4Harness called within geomeTRIC are of the catchall variety, namely: "Unknown error, error message is not found"
Is it possible that the keys we are attempting to access in Psi4Harness._handle_errors are simply not given by the output structures of the psiapi? Are we trying to access the error messages incorrectly?
To Reproduce
This is difficult to reproduce, as the errors are likely legitimate and occur in our use case from problematic geometry optimizations. Someone perhaps more familiar with the structures of the psiapi in psi4 may be well-positioned to validate the usage of the output structure to extract error messages in Psi4Harness._handle_errors. This could be as simple as a typo in a dictionary key name.
Expected behavior
We expect more detailed error messages to be found in many cases, but a large fraction of our Psi4 errors via geomeTRIC are of the "Unknown error" variety. This is unexpected and largely not useful, since this catchall error message is meant to be the exception, not the rule.
The text was updated successfully, but these errors were encountered:
Describe the bug
We previously attempted to improve the error reporting of the
Psi4Harness
via #266. However, now most of the errors we observe in production use of thePsi4Harness
called within geomeTRIC are of the catchall variety, namely: "Unknown error, error message is not found"Is it possible that the keys we are attempting to access in
Psi4Harness._handle_errors
are simply not given by the output structures of thepsiapi
? Are we trying to access the error messages incorrectly?To Reproduce
This is difficult to reproduce, as the errors are likely legitimate and occur in our use case from problematic geometry optimizations. Someone perhaps more familiar with the structures of the
psiapi
inpsi4
may be well-positioned to validate the usage of the output structure to extract error messages inPsi4Harness._handle_errors
. This could be as simple as a typo in a dictionary key name.Expected behavior
We expect more detailed error messages to be found in many cases, but a large fraction of our Psi4 errors via geomeTRIC are of the "Unknown error" variety. This is unexpected and largely not useful, since this catchall error message is meant to be the exception, not the rule.
The text was updated successfully, but these errors were encountered: