Hearing Transcription Backfill

## Problem

We now have a pipeline that automatically generates transcriptions of legislative hearings shortly after the hearing video is posted to the Legislature's website. This gives us access to transcriptions of new hearings going forward, but there are years worth of hearings from previous sessions that still need to be transcribed.

## Success Criteria
- [ ] New script that iterates over all hearings without transcriptions (from courts 192, 193, and 194) and:
- [ ] Check if there is a video associated with that hearing (by scraping the legislature's website - the logic for this is in the HearingsScraper linked below)
- [ ] If so, submit that video to AssemblyAi for transcription
- [ ] We should be able to largely re-use the logic from the current HearingsScraper for processing and submitting hearings for transcription

## Related Links
* Hearings Scraper (Where most of the relevant logic starts): https://github.com/codeforboston/maple/blob/ae476e4ac564553f1369f20a411cdfebab1acf5a/functions/src/events/scrapeEvents.ts#L300
* Transcription Webhook (The cloud function webhook that actually saves AssemblyAI's transcriptions to our database): https://github.com/codeforboston/maple/blob/ae476e4ac564553f1369f20a411cdfebab1acf5a/functions/src/webhooks/transcription.ts