Log the pages that generated no snippets, for further analysis.

[eggpi]

At a high level, our workflow for going through Wikipedia to find snippets is:

1. Enumerate all articles that have a {{citation needed}} template.
2. Process each of those articles trying to extract a snippet around those templates.
3. Figure out the categories of each article and a few more details.

Often, in step 2, we look at an article but fail to extract any snippets, even though the template is there. We should log these occurrences somewhere, so eventually we can tweak the code to get more snippets.

I think this could be as simple as inserting a log statement here in parse_live.py, once we've invoked the snippet parser and saw that it gives us no results.

We could also get a little more clever and log articles where we didn't extract enough snippets: e.g. there were 5 occurrences of the template, but only 1 snippet was produced.