Closed badisa closed 3 months ago
Does it make sense to consolidate max_visits
and max_node_visits
into just max_node_visits
(or even rename max_node_visits
to max_visits
to avoid breaking API)? I don't see much of a use for a per-layer semantic of max_visits
at this point.
Does it make sense to consolidate
max_visits
andmax_node_visits
into justmax_node_visits
? I don't see much of a use for a per-layermax_visits
at this point.
In https://github.com/proteneer/timemachine/pull/1067 we introduced the warning. The intention was to return a core if we had hit the limit. I would be alright with making it so that the warning is thrown if len(mcs_result.all_maps) > 0
and an exception otherwise. Which would avoid the additional parameter.
@proteneer I changed the logic in https://github.com/proteneer/timemachine/pull/1357/commits/49157dc669822106daf8826ecefe779406b7ef82 to raise an exception only if no cores were found, otherwise have a warning that the core may suboptimal
- The default value is set incredibly high, to preserve the old behavior
- Raise NoMappingError exception rather than assertion to make the errors raised more uniform
I think this also changes the meaning of max_visits
to accumulate over outer loop iterations (?) Should we mention that in the description and/or PR title?
- The default value is set incredibly high, to preserve the old behavior
- Raise NoMappingError exception rather than assertion to make the errors raised more uniform
I think this also changes the meaning of
max_visits
to accumulate over outer loop iterations (?) Should we mention that in the description and/or PR title?
Updated description (https://github.com/proteneer/timemachine/pull/1357#issue-2440795079) to indicate that it is the max visits across all thresholds (again think this edges, but want to get a better understanding of code before committing) and that this allows more deterministic wall clock times.
max_visits
now means the total number of nodes visited rather than the max number of visits per threshold (which I believe to be edges). This enables users to enforce more consistent wall clock times