Zoom File Handling

[jonthegeek]

This will likely go into a different repo, but for now this feels like the best place for me to do a brain dump.

This is what I do with the video + chat transcript that Zoom saves for each book club meeting.

Videos

• [ ] After a given meeting, wait for the video to process (I batch these and do them when things are available, we can probably set our GHA to do similar batching)
• [ ] Check that status == "completed"
• [ ] Download
• [ ] Delete video file on Zoom (we should wait until we confirm that the file is up on YouTube before we delete)
• [ ] Upload to YouTube (see settings below; after this step we need to wait or YouTube to process the video)
• [ ] Once the video upload is processed: Manually edit the video to trim off the beginning and end, and occasionally to cut out breaks in the middle. I don't see any reasonable way for us to automate this step. I add things to an "Edited" playlist at this point, so I can ~quickly see that I've processed them. I also make sure the description is accurate at this point (ie, that the club covered what they said they were gonna cover; sometimes they're off by a week or three).
• [ ] Once the edit is processed: Make the video public, choose a title screen (see below), and share the link in the club channel. Remove from the "Edited" playlist.
• [ ] Occasionally: Embed the video at the appropriate point in the bookdown (usually I do this a few times per club and do several chapters at once). In theory we could automate this, probably after the edit (using the title/description to know for sure where the video goes).

Video settings

When I do this manually, I copy settings from the previous video for this club (or the closest equivalent for a new club), and a lot of these are auto-set. But I'll log the important stuff here so we have them. A lot of this information comes from the individual club's sign-up sheet, which isn't always up-to-date (we can assume it is for the automation and then fix things as necessary during manual editing/review).

• [ ] Title:

  {book_title_short}: {chapter_title} ({book_abbr}{cohort_num} {chapter_num})

• [ ] Description:

  {presenter} presents Chapter {chapter_num} ("{chapter_title}") from {book_title_full} by {book_authors} on {meeting_date}, to the R4DS {book_abbr} Book Club, Cohort {cohort_num}.

  Read along: {book_url} Join the conversation: r4ds.io/join

• [ ] Playlist: {book_abbr}{cohort_num}
• [ ] "No, it's not made for kids"
• [ ] Tags: "rstats"
• [ ] Recording date: {meeting_date}
• [ ] License: Creative Commons - Attribution
• [ ] Category: Science & Technology ("Education" would be better but there are more settings I haven't messed with yet)

Chat Transcripts

• [ ] After a given meeting, wait for the chat transcript to be available.
• [ ] Check that status == "completed"
• [ ] Download
• [ ] Delete chat transcript file from Zoom (move this to after we process it in the future)
• [ ] Check that there's something more than "hi!" in the transcript. When we automate, it's fine if we don't do this filter.
• [ ] Post transcript in the appropriate channel on R4DS, with something like "Cohort 03 from 2023-01-11:". The transcript goes inside tickmarks to preserve code and formatting and whatnot.
• [ ] Right now: Copy transcript to a "Chat logs" folder to deal with in batches. We'll need to store it somewhere for the next step when it's automated, but theoretically we could use the R4DS thread as the record; that way I could manually edit those if necessary and before we commit them to the repo.
• [ ] Occasionally: Embed the log at the appropriate point in the bookdown, and delete the local copy. When automated, this should happen at the same time that the video is PRed into the bookdown, and needs to happen within 90 days so it's still in the available Slack data.

[jonthegeek]

Oh, I forgot to include the note I intended to include about the video thumbnail!

I created special thumbnails for some of the clubs. For example, this is one of the ones from rpkgs.

I don't love those but they show the idea We could theoretically either create these for everything, or even automate it (book cover centered + text at various places based on chapter number, that sort of thing).