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
In order to remove the Log object, we need to print messages directly to stdout.
This is straightforward except that we will then need to have a way of sorting messages by verbosity. The approach we are taking is to prepend messages with a known string (ie 5:) that contains information about verbosity. Higher numbers mean more verbose messages. Users could then grep messages directly (ie. "I want all messages with prefixes less than 5). Or we could later redirect stdout to a streamIO and parse the numbers ourselves. In either case, this would take place at the level of glue-code or higher.
The text was updated successfully, but these errors were encountered:
One challenge with this approach is when printing things like Numpy arrays, or other multi-line objects from non-HORTON code.
There are two solutions I can think of: One is to modify the grep so that un-prefixed output retains the same level as the last prefix. Another solution is to print to a streamIO, modify the output with the correct prefixes, and then pass it to stdout. In either case, it's a small change which can be deferred for later.
I like getting rid of the logger; I think this is more-or-less what Toon had recommended earlier. He had talked about passing a log object as an argument also, but I guess this may be preferable; it is more straightforward I think.
In order to remove the Log object, we need to print messages directly to stdout.
This is straightforward except that we will then need to have a way of sorting messages by verbosity. The approach we are taking is to prepend messages with a known string (ie
5:
) that contains information about verbosity. Higher numbers mean more verbose messages. Users could then grep messages directly (ie. "I want all messages with prefixes less than 5). Or we could later redirect stdout to a streamIO and parse the numbers ourselves. In either case, this would take place at the level of glue-code or higher.The text was updated successfully, but these errors were encountered: