Add verbosity flag for MPB #1302
Conversation
…into place with a simple `make`
|
In addition...
|
…ribute. This will help with transitioning while the feature rolls out.
|
I think the remaining test failures are all due to using a libmpb that does not have a |
|
When running interactively from Python, e.g. in Jupyter notebooks, we probably want to set verbosity to 1 by default. |
* Return the prior verbosity level setting. Change the parameter name to `level` * Add verbosity level global variable, and a Python function for setting it * Switch the verbosity value and related code to a class * switch parameter name to `level`, prep for reuse with other cvars * Move verbosity module to meep * Rename module to avoid name conflicts with names of instances * Restore returning the former level from set() * Use the Verbosity class in meep * Add ability to set the initial level * Make sure we use mpb's verbosity flag, and not colliding with meep's * Move the creation of the verbosity obj so it's available when submodules are imported * Add explicit rules for the meep.mpb .py files so they will be copied into place with a simple `make` * Add verbosity checks in solver.py * Add verbosity checks in libpympb/pympb.cpp * Revert a Makefile change * Handle the case where a cvar object does not yet have a verbosity attribute. This will help with transitioning while the feature rolls out. * Python2 doesn't have f-strings * empty commit to trigger build * Support building with an older MPB which doesn't have the mpb_verbosity global * Set the verbosity default to 1 * match MPB Co-authored-by: Steven G. Johnson <[email protected]>
See NanoComp/mpb#116 and NanoComp/mpb#119
Currently all verbosity checks are checking for
>=2so more fine-tuning will be needed to adjust the levels tested for the specific messages.