The audio file link in the minutes assumes the date is the directory but for task forces such as vc-education, the link should be (example) 2021-07-26-vc-education/audio.ogg instead of 2021-07-26/audio.ogg

In /scribe-tool/index.html the fetchLog function searches by data for a Jitsi url path that is specific to CCG like:

https://meet.w3c-ccg.org/archives/w3c-ccg-weekly-"+dateString+"-irc.log

Task force URLs vary such as for VC-EDU like:

https://meet.w3c-ccg.org/archives/w3c-ccg-education-"+dateString+"-irc.log

but fetchLog isn't aware of this.

If someone provides the Jitsi urls for the task forces in this issue, we can update scribe tool to search for task force minutes which will make generating and sharing the meetings much more efficient for the task forces.

This repo is massive because of all of the audio files, and it takes ages to fetch it the first time, and is pretty slow to fetch subsequent updates. When I just want to tweak scribe-tool, it's quite burdensome to download all of these audio files to my local machine.

I think it would be a suitable separation of concerns to keep this repo only for minutes/recordings, and have the 'plumbing' elsewhere. We could still keep a github pages scribe-tool page here (ie at w3c-ccg.github.io/meetings/scribe-tool) if that's desired by including the JS from another repo.

Unable to upload 10/27 audio file due to 25mb upload limit

Heather is too tired to address this now, but trouble shoot error message for minutes for solid extra meeting and the regular meeting on 3-10. https://github.com/w3c-ccg/meetings/actions/runs/640488484

Getting the following error when attempting to generate VC-Edu meeting minutes:

The Twitter notification CI step is failing, because I don't have correct credentials, or something.
Do we /need/ a Twitter CI step?

4/21 minutes got an "access denied" when downloading the log, but we ended up publishing empty minutes. Change the download step to detect if this happened, and fail fast

These items would help:

• configurable tweet
• configurable email title (e.g. edu vc vs regular calls)
• include link to instructions about cleaning up minutes

Github pages can take up to 5 minutes to build, so often the link to generated minutes is sent out by email and twitter before the page is ready. Maybe we can add a delay in the github actions script to minimise chances of people clicking the link immediately and getting a 404.

Admittedly the only people likely to do that are the people generating the minutes, to check it worked, so not high priority.. but just in case..

For the IFTF to discuss.

Ted's comment here points out that task force minutes for the DID resolution group looks different from the main CCG group (and VC-EDU). Truth is, it takes some wrangling to work with these scripts if you're not using the default CCG tools.

Whether the IFTF wants to support task forces that don't use Jitsi is one question they should consider. If they do want to, I wanted to share my experience based on the VC-EDU task force.

The index.js and scrawl.js scripts are quite tangled, generating both the summary index.html and individual index.html off of the irc.log file (and they'll try to recreate, making it tricky in the current form to directly write .html), so I found it was easiest to force my formats into the irc.log format; e.g

<date>\t<alias>\t<message>

Within that message, the scrawl script will look for magic words like "Action:", etc.

Going that path of least resistance, I ended up creating 2 command line tools:

1. vtt2irc: takes a zoom auto-transcribed transcript and converts it to irc.log. I stopped using it because these transcripts are very noisy.
2. txt2irc: takes a simpler txt file that basically contains <message> and converts it to an irc.log the scripts are happy with.

More about txt2irc: note that, in VC-EDU, our task force members decided that we wanted the audio recording and a much shorter summary: decisions + resolutions. Because we are very "working session" oriented, most of our work happens in github anyway, so the scribe function is sort of a distraction.

I also experimented with a cleaner solution involving mustache templates, but it because too messy because of the above-mentioned script entanglement, so I abandoned it.

What's interesting is that @peacekeeper seems to have landed on this same pattern as VC-EDU did; i.e meetings are working session oriented. Most of the work happens in github, and minutes are only needed for very high-level things.

This issue is mostly to capture my thinking on the topic while it's fresh, so that the IFTF can decide how they want to pursue this, including if they want to not support this path. But because our two current active TFs seem to be using patterns where CCG scripts are cumbersome, I thought you might want to be aware.

If you find any of my utilities useful, I'll happily transfer them to the ccg org.

The organizers in index.html of the meeting notes are listed as "Wayne Chang", "Heather Vescent", "Mike Prorock" when they should be the co-chairs of the meeting topic/task force.

It looks like when the index.html in scribe-tool is run it initializes the scrawl with scrawl.chair = config[0].chairs forcing the first item in the array to be used. Based on other functionality, it looks like this should rely on the group.txt value?

Consider putting all 2021 in a separate folder under this repo to better manage files.

These might be appropriate for a different repo. I'm logging them here as "most likely repo" thanks to Dmitri's CCG sleuthing. Feel free to move or copy them to other repos. Please tag me if you do so, so I can track updates.

These probably need to be handled distinctly, but they're intimately connected so I'm raising one issue for both.

1. transcriber fails mid-meeting

Historically, this was handled by stopping and restarting the recording, but apparently that resulted in overwriting the existing recording (up to "now") with the new recording (starting "now"), so the earlier part of the meeting was lost.

This could be partially handled by naming the recording file with DATETIME instead of just DATE, perhaps rounding to the minute (dropping seconds and fractions thereof). That would make the admin functions a bit more challenging, but would at least keep as much recording as the day's chair was able to capture with timely re-starts.

After-meeting post-processing of the recording to generate a transcript could then be used against all the audio files, hopefully resulting in something more complete than the mid-meeting failure delivers.

2. transcriber doesn't hear some speakers

This is an ongoing problem, which might be related to browser-specific issues on the speaker's side, but the speaker isn't always able to drop-and-reconnect or to change browsers.

This would be less of an issue if the recording is complete, or if the post-processing described above can be run (assuming the recording captures all speakers, even when the transcriber doesn't), but it's still problematic during the meeting.

Opening this repo we see:

We found potential security vulnerabilities in your dependencies.
Some of the dependencies defined in scribe-tool/package.json have known security vulnerabilities and should be updated.

Only users who have been granted access to vulnerability alerts for this repository can see this message.

Learn more about vulnerability alerts

We just made some general improvements to the various GitHub Actions running here, but more is needed based on the logs from this most recent run: https://github.com/w3c-ccg/meetings/actions/runs/9387952522

Mainly, the file changes system we've been using hasn't been touched in sometime, so we should consider moving to https://github.com/tj-actions/changed-files

Also, set-output is deprecated, and we should move to using $GITHUB_STATE per https://github.blog/changelog/2022-10-11-github-actions-deprecating-save-state-and-set-output-commands/#examples

We may also want to look into using https://github.com/nektos/act for local GitHub Action testing (probably coupled with some mail trap system).

Never a dull moment. 😉

Look at Action logs in the week of 6/22-25 for failure messages. Essentially we can no longer send through the w3c-ccg gmail account, and it broke when we switched over ownership of the account. We need someone to investigate and fix

Extensions of @rhiaro's scribe_tool improvements:

• automate cleanup ui -> git - I think scribe-tool can already talk to
  the git API, just needs hooking up to the the frontend and some api
  token magic..
• have ccg-bot automatically offer up the cleanup URL in IRC once the
  call ends for the designated logs-cleaner to click on (in SocialWG this
  used to also be the scribe's job - and I made a similar web tool for
  that too - so it was a different person every week, though in practice I
  ended up doing it a lot of the time anyway).
• make the scribe-tool UI actually look nice.

We'll need to backfill audio recordings from the beginning of 2020 til whenever we automate audio as part of github actions

need to convert and trim out beginning/end.

The main CCG meeting minutes from 2022-09-20 onward are not correct. The irc logs are there, but the generated and sent emails and the web pages are wrong and usually just have one line. The other related groups seem fine.

https://github.com/w3c-ccg/meetings/tree/main/2022-09-20
https://github.com/w3c-ccg/meetings/tree/main/2022-09-27
https://github.com/w3c-ccg/meetings/tree/main/2022-10-04
https://github.com/w3c-ccg/meetings/tree/main/2022-10-11
https://github.com/w3c-ccg/meetings/tree/main/2022-10-18
https://github.com/w3c-ccg/meetings/tree/main/2022-10-25
https://github.com/w3c-ccg/meetings/tree/main/2022-11-01

4/20 minutes were not generated correctly. Note that irc.log is correct, but the generated email.log and index.html only picked up the last entry.

It looks like when there are two meetings on the same date, on the meetings index page the second overrides the first in the list.

https://github.com/w3c-ccg/meetings/blob/gh-pages/generate_minutes.md#step-1-fetch-raw-minutes-into-github-meetings-repo

From @msporny, we could improve the ease of creating meeting transcripts as follows:

1. A way to stream IRC logging to a public website.
2. A way to send the audio record to a public website.

Once we have that working, most anyone could clean up the minutes.

The specific capabilities that we need are:

1. A website that can keep logs (separated by CG/WG), so something that
   allows the IRC bot to send single IRC line POSTs to
   https://example.com/ccg/2017-07-28.log
2. The same for audio: https://example.com/ccg/2017-07-28-audio.wav

We need a simple server that is capable of doing that.

Files are getting generated but html url is serving a 404.

See error:

System.IO.IOException: No space left on device : '/home/runner/runners/2.309.0/_diag/Worker_20231003-184749-utc.log'
at System.IO.RandomAccess.WriteAtOffset(SafeFileHandle handle, ReadOnlySpan1 buffer, Int64 fileOffset) at System.IO.Strategies.BufferedFileStreamStrategy.FlushWrite() at System.IO.StreamWriter.Flush(Boolean flushStream, Boolean flushEncoder) at System.Diagnostics.TextWriterTraceListener.Flush() at GitHub.Runner.Common.HostTraceListener.WriteHeader(String source, TraceEventType eventType, Int32 id) at GitHub.Runner.Common.HostTraceListener.TraceEvent(TraceEventCache eventCache, String source, TraceEventType eventType, Int32 id, String message) at System.Diagnostics.TraceSource.TraceEvent(TraceEventType eventType, Int32 id, String message) at GitHub.Runner.Worker.Worker.RunAsync(String pipeIn, String pipeOut) at GitHub.Runner.Worker.Program.MainAsync(IHostContext context, String[] args) System.IO.IOException: No space left on device : '/home/runner/runners/2.309.0/_diag/Worker_20231003-184749-utc.log' at System.IO.RandomAccess.WriteAtOffset(SafeFileHandle handle, ReadOnlySpan1 buffer, Int64 fileOffset)
at System.IO.Strategies.BufferedFileStreamStrategy.FlushWrite()
at System.IO.StreamWriter.Flush(Boolean flushStream, Boolean flushEncoder)
at System.Diagnostics.TextWriterTraceListener.Flush()
at GitHub.Runner.Common.HostTraceListener.WriteHeader(String source, TraceEventType eventType, Int32 id)
at GitHub.Runner.Common.HostTraceListener.TraceEvent(TraceEventCache eventCache, String source, TraceEventType eventType, Int32 id, String message)
at System.Diagnostics.TraceSource.TraceEvent(TraceEventType eventType, Int32 id, String message)
at GitHub.Runner.Common.Tracing.Error(Exception exception)
at GitHub.Runner.Worker.Program.MainAsync(IHostContext context, String[] args)
Unhandled exception. System.IO.IOException: No space left on device : '/home/runner/runners/2.309.0/_diag/Worker_20231003-184749-utc.log'
at System.IO.RandomAccess.WriteAtOffset(SafeFileHandle handle, ReadOnlySpan`1 buffer, Int64 fileOffset)
at System.IO.Strategies.BufferedFileStreamStrategy.FlushWrite()
at System.IO.StreamWriter.Flush(Boolean flushStream, Boolean flushEncoder)
at System.Diagnostics.TextWriterTraceListener.Flush()
at System.Diagnostics.TraceSource.Flush()
at GitHub.Runner.Common.TraceManager.Dispose(Boolean disposing)
at GitHub.Runner.Common.TraceManager.Dispose()
at GitHub.Runner.Common.HostContext.Dispose(Boolean disposing)
at GitHub.Runner.Common.HostContext.Dispose()
at GitHub.Runner.Worker.Program.Main(String[] args)