feat: add loggers and change prints to logs #27
Conversation
|
Hey @portellaa this is a big PR in terms of diffs but small in terms of core changes, your insight is useful on the following key aspects:
Notice on the logger class that despite us doing get logger and retrieving always the same logger, we were adding a new handler over each time the function was run (exactly once per engine, 8x warnings on the full package class 💥 ). That is why the add handler is conditioned on the number of existing handlers of the logger. We are still gonna iterate this over time but still persisting and most in need ASAP fix is finding a clever way to print the module name with the logs. Ideas are much appreciated! Cheers 🍻 |
src/ydata_quality/utils/logger.py
Outdated
|
|
||
| def get_logger(name, stream: TextIO = sys.stdout, level: str=logging.INFO): | ||
| acceptable_levels = [None]+list(_nameToLevel.keys()) | ||
| assert level in acceptable_levels, "Valid levels for warning severity are {}. Defaults to info level.".format(acceptable_levels) |
There was a problem hiding this comment.
why not use fstr formatting?
assert level in acceptable_levels, f"Valid levels for warning severity are {acceptable_levels}. Defaults to info level."
Changed all prints to logs
Defined a base logger for DQ
Changed lingering target variable names to label
TODO: Find a good way to print module names in logs
Fix a persisting runtime bug