Introducing conflict resolution strategies in the UMA server

[woutslabbinck]

Introduces a Strategy component in the ODRLAuthorizer such that the strategy is less hardcoded.

[woutslabbinck]

There is a small issue which is related to componentsjs and the package policy-conflict-resolver.
As a result, building the packages results into an error:

[@solidlab/uma]: Resource https://linkedsoftwaredependencies.org/bundles/npm/policy-conflict-resolver/^0.0.0/dist/EvaluatorHandler.jsonld#EvaluatorHandler is not a valid component, either it is not defined, has no type, or is incorrectly referenced by https://linkedsoftwaredependencies.org/bundles/npm/policy-conflict-resolver/^0.0.0/dist/WrappedEvaluatorHandler.jsonld#WrappedEvaluatorHandler.
[@solidlab/uma]: Process exited (exit code 1), completed in 6s 275ms

All tests do work however if you remove all aspects of components.js in node_modules (which is dirty, but I wanted to check if I had to change more tests) from the policy-conflict-resolver.

There are two more interesting aspects that could be introduced in this PR:

• make strategies truly configurable via componentsjs (not much work)
• ODRL evaluator 0.6.0 introduced eyeling. We could make eyeling the default N3 Reasoner (though maybe that is better for a different branch + PR)