Make fuzzy-rule configurable with an external data source

[benoit74]

Currently, fuzzy rules are configured in a YAML (/JSON) file and transformed into code.

Mid-term goal is to share these rules with WebRecorder team and other contributors. This probably means that at some point we will need to source this information from an online source.

Even before that, we would benefit from being able to source these rules from an external data source so that they can be updated without needing a new warc2zim release.