Add feature for one rule to inhibit another

[jdbaldry]

I'm thinking of something similar to the Prometheus Alertmanager's inhibition: https://prometheus.io/docs/alerting/latest/alertmanager/#inhibition.

A concrete example is that I'd like the ProductPossessive rule that catches "Grafana's" to inhibit the spelling error that fires for that word. In this case, I can allow the word by updating the dictionary to include a possessive affix and that may be the best solution.

I'm not sure of how exactly inhibition should behave but here are some ideas:

1. Use start and end locations: one rule would inhibit another if they have the same start and end locations.
2. Use tokens: one rule would inhibit another if they match the same token.
3. Use a list of messages: one rule could inhibit any messages that match a list of regexps defined in the rule.