Check whether `re.escape()` can be used in equal comparisons

[strogonoff]

I think re.escape() is supposed to be used with PostgreSQL’s like_regex operator, when we want to make a literal string part of a regular expression safely.

It is mostly used this way.

However, there are a couple of places in this service where it is also used in literal comparisons, for example:

https://github.com/ietf-ribose/bibxml-service/blob/b49a3614a92b4f5db03f4b38d25074930dc0a5f6/xml2rfc_compat/fetchers.py#L39-L40

https://github.com/ietf-ribose/bibxml-service/blob/b49a3614a92b4f5db03f4b38d25074930dc0a5f6/main/tests/test_query.py#L94

I suspect this could introduce undesirable behavior. For example, re.escape() escapes dots (re.escape('"foo.bar"') == '"foo\\.bar"'), meaning if docid contains a dot then PostgreSQL may not match it with == operator.

Could we check whether re.escaping is safe there, and if not removing or replacing it with other way of escaping (e.g., using json.dumps() on those strings)?