Use `dateutil` for Date Parsing Everywhere

In some parts of the codebase, we are using the dateutil package, while in others we are using pandas (!) and in some places, I think we might still be calling strptime(). (Noticed this issue while looking at #356.) We should make audit all our files and make sure to use dateutil everywhere.

Additionally, in places where we know a date is an ISO 8601 date (like parsing responses in web_monitoring.db), we should use dateutil.parser.isoparse() to make sure there’s never any ambiguity.

This requires a little spelunking through the codebase (note: we don’t need to worry about the .ipynb [Jupyter/iPython Notebook] files for this), but should mostly be a straightforward search for date parsing locations and replace with calls to dateutil.parser.parse().

Update: No need to go searching for where this needs to be changed. Here are the places that need updating:

Parsing CLI arguments in cli.py:

web-monitoring-processing/web_monitoring/cli.py

Lines 682 to 687 in 5d1a86a

    
           try: 
        
               parsed = pandas.to_datetime(date_string) 
        
               if not pandas.isnull(parsed): 
        
                   return parsed 
        
           except ValueError: 
        
               pass

Parsing “unplaybackable” cache records:

web-monitoring-processing/web_monitoring/cli.py

Lines 566 to 567 in 5d1a86a

if isinstance(date, str):

date = datetime.strptime(date, '%Y-%m-%dT%H:%M:%SZ')

There are some spots where we use strptime() in internetarchive.py, but we are extracting that module into a separate package (see #477), so it should not be changed.

	try:
	parsed = pandas.to_datetime(date_string)
	if not pandas.isnull(parsed):
	return parsed
	except ValueError:
	pass

	if isinstance(date, str):
	date = datetime.strptime(date, '%Y-%m-%dT%H:%M:%SZ')