Stop using custom logging setup for workflows

[oyvindeide]

Using:

subscript/src/subscript/__init__.py

Line 12 in 5e159d1

def getLogger(module_name="subscript"):

results in flooding the stdout of ert when used in workflows where subscript is essentially used as a library. In those cases using logging is fine, but ert should be responsible for setting up the logger. In those cases the standard logging module should be used directly:

import logging
...
logger = logging.getLogger(__name__)
logger.info("msg")

for forward models using the subscript logger is fine, because they have their own entry points and is the module in charge of the excution.

[berland]

For e.g. csv_merge, is this a matter of only resetting the logger object (and its level) around

subscript/src/subscript/csv_merge/csv_merge.py

Lines 64 to 72 in 5e159d1

	def run(self, *args):
	# pylint: disable=no-self-use
	"""Parse with a simplified command line parser, for ERT only,
	call csv_merge_main()"""
	parser = get_ertwf_parser()
	args = parser.parse_args(args)
	logger.setLevel(logging.INFO)
	globbedfiles = glob_patterns(args.csvfiles)
	csv_merge_main(csvfiles=globbedfiles, output=args.output)

?