Refactor transcript tooling into modular quality API

[scottjoyner]

Summary

• introduce a structured quality_api package with configuration, database, embedding, and service layers powering a refreshed transcript search experience
• replace the inline Flask app in compare_transcripts.py with an application factory that renders a modern UI, exposes JSON endpoints, and surfaces speaker/location context and transcript drilldowns
• refactor quality validation logic into reusable services, add Docker and docker-compose definitions, and document gunicorn dependency for containerized deployment

Testing

• python -m compileall quality_api compare_transcripts.py quality_report.py

---

https://chatgpt.com/codex/tasks/task_e_68d69ab947308325a0fe4763f8e4149f

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting

• @codex fix this CI failure
• @codex address that feedback

[chatgpt-codex-connector]

[P1] Handle missing transcripts without raising driver error

The new get_transcript method calls session.run(...).single() and then checks if not record. single() raises neo4j.exceptions.ResultNotSingleError when the query returns zero rows, so a request to /transcripts/<id> with an invalid ID will crash the endpoint instead of returning a 404. Catch the exception or use single(strict=False) so the method can return None and the view can issue the expected 404 response.

Useful? React with 👍 / 👎.