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
Due to ray-project/ray#21424, Ray (and thus Modin) is making pytest reports very noisy.
Presumably because of Ray printing these things on a separate server, something like money-patching print as a no-op doesn't work. I tried playing around with pytest's *Report objects, which get exposed in various hooks, however it seems none of these actually hold these captures (i.e. in report.sections). Interestingly pytest is capturing these prints and reporting them in the stdout/stderr capture sections in the terminal report, but none of the relevant hooks seem to expose those for us to remove. Note terminal summary (i.e. pytest_terminal_summary) comes after captured stdout/stderror, .
I'll have to have another look as these prints make testing with Modin/Ray rather impractical. I focused on Ray as the engine as that seems to be the more common and supported use case, but I might just see how Dask plays as an engine (or indeed test with both).
The text was updated successfully, but these errors were encountered:
Due to ray-project/ray#21424, Ray (and thus Modin) is making pytest reports very noisy.
Presumably because of Ray printing these things on a separate server, something like money-patching
print
as a no-op doesn't work. I tried playing around with pytest's*Report
objects, which get exposed in various hooks, however it seems none of these actually hold these captures (i.e. inreport.sections
). Interestingly pytest is capturing these prints and reporting them in the stdout/stderr capture sections in the terminal report, but none of the relevant hooks seem to expose those for us to remove. Note terminal summary (i.e.pytest_terminal_summary
) comes after captured stdout/stderror, .I'll have to have another look as these prints make testing with Modin/Ray rather impractical. I focused on Ray as the engine as that seems to be the more common and supported use case, but I might just see how Dask plays as an engine (or indeed test with both).
The text was updated successfully, but these errors were encountered: