Oklahoma's `search_by_date` method is broken for counties without a dedicated latest filings page

[palewire]

When an Oklahoma county lacks a dedicated daily filings page, the scraper attempts to use the global search.

if self.place_id in DAILY_FILING_COUNTIES:
    search_obj = DailyFilings(self.place_id)
else:
    search_obj = Search(self.place_id)

That doesn't work, as shown in the test_site.test_by_date_kiowa test, which will fail. The traceback looks something like:

        # Search results contain an entry for every party
        # to a case, so we need to deduplicate
        results = {}
        # Only grab result rows (i.e. skip header)
>       for row in self.soup.table.find_all("tr", class_="resultTableRow"):
E       AttributeError: 'NoneType' object has no attribute 'find_all'

The bug can be recreated manually with the following snippet:

from court_scraper import Site
kiowa = Site("ok_kiowa")
kiowa.search_by_date()

The test that demonstrates the bug is currently commented out. If this bug is squashed, it can be restored.