Pretty representations of Transcripts via embedded JSON/YAML in PDFs

[chriswblake]

The current specs from the model only declare that the transcript must be provided as a JSON or YAML. However, this format is naturally unfriendly for everyday use.

Recommendation:

• Create tooling to attach the JSON/YAML file to a PDF.
• Create a PDF generation tool that provides options for visualizing the topic scores.

Thoughts:

• Use a templating engine, maybe like LaTeX so users can define their own styles.
• Add support for using the official score interpretations.
• Add restricions to prevent misleading visuals.