search

tim1mw / pycon_capture

Automated conference video capture and upload for PyconUK (and others)
4 stars 3 forks source link

Naming the things #19

Open IanTaylorFB opened 5 years ago

IanTaylorFB commented 5 years ago

Currently the project is called pycon_capture. But it is also now uploading the video files. Soon, it will create the playlists. The Google Project is called auto uploader. The script for capture is called code.py, while the script for upload is called upload_completed_videos.py.

I suggest the following: PyCon Video Manager (pycon_video_manager) as the project name, both in GitHub and Google. Use the pretty one in Google and for the GitHub title / documentation and the Pythonic name for the code repo.

With these top-level scripts for the main function steps: v-capture.py capture the video stream v_cut.py wrapper around ffmpeg-script.sh cutting the live stream v_verify.py verify that the session file content matches the id v_upload.py upload the session files v_playlist.py create / update playlists v_web_update.py update the main website data

More detail: v-capture.py would be the new name for code.py called from the local web page.

v_cut.py would be a wrapper around ffmpeg-script.sh cutting the live stream into separate session files. Or maybe, the script writes and then runs ffmpeg-script.sh so you don’t have to press a button on the web page and then run a script as well?

v_verify.py would iterate over the separate session video files, loading each in turn into the VLC player for human checking. After each video is reviewed, the script confirms whether it is ok to upload it. If the operator responds yes, then the file is moved from edited_sessions to ready_for_upload and if the response is no, the file is moved to more_editing_needed.

v_upload.py would upload the session video files and captures the video_id returned by YouTube. (See issue .)

v_playlist.py would create or update the YouTube playlist for the day and room combination. This needs to be a separate script from v_upload.py because we need to know all the YouTube video_id and have them in the correct order to create the playlist. Also, because we might have a video missing (say because some fix is required first) we need to be able to update (or recreate) the playlist to include the missing item. All this requires that we capture the video_id from YouTube and associate it with the session_id (see issue ).

v_web_update.py (future capability) once we have captured the video_id we should write these back to the main website, so that anyone browsing the site later can jump straight to the video. This additionally needs a POST endpoint for website schedule metadata updates.

IanTaylorFB commented 5 years ago

Note that v_playlist.py would be dependent on #12 to get the YouTube id.