Fixing issues with BethelParkSpider (prod failures)

[will-snavely]

The BethelParkSpider was failing in production for a few reasons

• I wasn't assigning an ID to each meeting, which is required in production
• The start/end datetime set in the meeting needs to be tz-unaware, I think (see this line in the city-scrapers-core code; it compares the start time of the meeting with datetime.now(), a tz-unaware date. If the start time is tz-aware, this comparison fails. I'm not sure their code is correct, though).
• The meeting location entry needs to be a dict, not a string
• The meeting description can't be None

I made these changes and ensured the scraper now produces output in a production setting (by running scrapy crawl and dumping to my own s3 bucket). Also updated tests to check these new requirements.

[will-snavely]

Note: I propose allowing a dev to stand up their own config find in the city_scrapers folder, called dev.py (opposed to the current prod.py). That's what I did for the purposes of testing, here. This is why this commit includes a .gitignore for that file.

[ben-nathanson]

This looks good. Thanks @will-snavely !