Transcripts for every available Harmontown episode, transcribed using OpenAI Whisper.

todo:

• replace rss_parser library

• download episodes

• create a version of this that can run on any rss feed

  • deal with weird urls used for marketing (how?, maybe keep only string after the last "https://")

• script that can transcribe any episode

  • semantic search
  • shell script (or something?) that runs script that can transcribe any episode such that you can stop it and start it whenever
  • make this the main file to run, that does everything in "lexicap for any podcast" version
    • requirements.txt
    • create a logfile for which podcasts have been done, which podcasts are in progress, etc
      • basic search (is semantic search feasible on device?)
    • for lexicap-any (come up with a better name):
      • basic checks of whether directories involved exist for clientside
      • check if the podcast mp3 actually downloads
      • how do you actually build stuff like this that you just plug and play run kind of deal?
    • Add description and if there are shownotes, create something to find closest link div to that shownote.

• standardize use of " vs ' in the files