Skip to content

Latest commit

 

History

History
120 lines (87 loc) · 4.89 KB

VIDEOS.md

File metadata and controls

120 lines (87 loc) · 4.89 KB
Raw

Videos

Major tasks with videos:

  • Add the playlists to the events files
  • Add more videos
  • Add twitch streams
  • Add the speakers
  • Add tags, update descriptions

Trailers: There are some conferences that upload two version of their talk the full version and one called -trailer which is a 3 min snippet. The trailers should be skipped. (We should probably also remove them later on.)

Add the playlists to the events files

Each event has a field called youtube that can hold the id of the playlist in YouTube of the videos of that specific event. Running python3 video.py will list all the event that do NOT have a playlist yet. Check the conference web page and/or search for the name of the conference in YouTub and add the ID of the playlist to the appropriate data/events/* file.

Add more videos

We index the videos recorded at events. The videos of event abc-2000 which is stored at data/events/abc-2000.json are located in data/videos/abc-2000. Each video is represented by a JSON file in that directory. Optionally each video can also have a .html file (with the same filename as the JSON file has). The content of this HTML file will be the "description" of the video. The reason we allow external files is to make it easy to spread the text to multiple lines and to include html tags.

For now we are only indexing videos on YouTube.

Add videos from YouTube

Using a forked version of the youtube.py from https://github.com/pyvideo/data/tree/master/tools we can fetch the list of videos from Youtube and create the JSON files.

python3 tools/youtube.py -k YOUTUBE_API_KEY --list YOTUBE_PLAYLIST_ID -d 2016-08-03

This will create the a directory called category-TIMESTAMP inside that directory is a subdirectory called videos which needs to be moved to conferences/data/videos/abc-2000.

TODO: further change the Python script to save the length of the video.

Update the videos from YouTube

  1. in the speakers field add the name of the speaker or speakers in "full-name" format. The format should only include lower case a-z letters and dash (-).

The name of the speaker usually can be found in title of the talk and/or in the description of the talk. In many cases the filename of the video also contains the name of the speaker(s). Sometimes however this is missing or only part of the name can be found. In such cases it is harder to track down who is the speaker.

Check if the file date/people/full-name.json exist, if not create it and fill it with the appropriate fields. You can usually find some of the details of the speaker on the web site of the event. From there you can usually find the other values as well.

name:
twitter:
github:
home:
country:
  1. Remove the name of the speaker and the name (and year) of the conference from the title. Keep just the real title of the talk. For example if this was the title "Foo Bar - Frobnicating the FuzzBazz - Conference 2016" then it looks like the name of the speaker is "Foo Bar" and the event is "Conference 2016". We would like the title to be: "Frobnicating the FuzzBazz"

If possible add a field called 'abstract' that will lead to the original page of the talk on the web site of the conference.

If you can find the slides of the talk, add a field called 'slides' linking to those slides.

Run python3 generate.py --html to generate the web site and to check if the values in the speakers field match the filenames.

Add the speakers

The videos of these conferences were already added, but the speaker information has not been added yet:

grep speaker * | grep '\[\]'

Add tags, update descriptions