Synthetic article for runner checks

This synthetic paragraph exists to exercise the local comparison runner. It is authored test data, has no real site or human quality labels, and cannot justify changing an extraction default. Both profiles receive the same saved HTML bytes. The evaluator records whether the independently listed fixture snippets occur in each profile output.

A second synthetic paragraph provides enough connected prose for a smoke observation of the optional package. Runtime and traced Python allocations describe this single process only. Import overhead, interpreter caches, native allocations and unmeasured resident memory prevent treating this fixture as a resource benchmark. Real authorized French and English pages still require independent annotation.