Visualize a sample of the alignments in the task alignments task

[gregtatum]

When training languages it can be hard to know if alignments are generated correctly and behaving as expected. They can be visualized with a tool like word-alignment-visualization.

During the alignment step, it would be nice to sample the corpus, and generate a set of alignments so the we can manually verify that the alignments are working as expected.

This could be an alignments.html file that is generated in the artifacts.