Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The Very Last Day #12

Open
frittate opened this issue Nov 1, 2022 · 6 comments
Open

The Very Last Day #12

frittate opened this issue Nov 1, 2022 · 6 comments
Labels

Comments

@frittate
Copy link

frittate commented Nov 1, 2022

Trying to generate a somewhat coherent story by

  • scraping episode summaries from Wikipedia for several TV shows, such as Game Of Thrones, Futurama, Lost, Friends etc. Detailled selection tbd.
  • cutting those summaries into bullet points for each plot point
  • (maybe replacing all the names with a single set of names as to have a coherent set of characters)
  • arraging random plot points into set of 10 or so to have one "episode"
  • using an AI to generate an episode summary from the plot points, or maybe to generate a text from this.
  • ???
  • profit!
@frittate
Copy link
Author

frittate commented Nov 1, 2022

I think I'll use a show will fewer characters, such as Brooklyn Nine-Nine instead of the massive GoT.

@frittate
Copy link
Author

frittate commented Nov 1, 2022

@frittate frittate changed the title The Best of All Shows The Very Last Day Nov 15, 2022
@frittate
Copy link
Author

I've updated my approach and have produced a 50K-ish document, titled at "INTERIOR.md".
https://github.com/frittate/2022-NaNoGenMo/blob/master/INTERIOR_2022-11-14--21-10.md
I'm not done and finished because eventually, I liked what happened and want to bring to a higher level still.

The code at interior.py does the following:

  • goes to sfy.ru and takes a number of movie scripts which are hosted there as text files. From the texts, it looks for movie scene headings: INT, EXT, INTERIOR, EXTERIOR. These are typically used to describe scenes in films, and I think this works here as well
  • goes to a news API and gets some news headlines and excerpts on the topic of "climate change"
  • Under the scene headings, it randomly adds several types of content. Either there is a variety of descriptors for emptyness and silence, or sources of sound which take content from the news in point 2
  • The result is a montage of single scenes from various "locations" (provided by the movies) that are devoid of human life, and some commentary on current climate change topics, played by automated systems, interspersed.

Some funny moments occur when well known movies have been scraped, and the scene headings are "SURFACE OF DEATH STAR", I'll decided whether to keep this in the future.

@frittate
Copy link
Author

I've counted about 3700 "shots" (or scene headings), if each shot is 2 seconds in a movie, the movie would run 123 hours, or about 5 days.

@frittate
Copy link
Author

Some excerpts (text from news segments)

INT. FIRE ACADEMY CHIEFS OFFICE - DAY
a radio repeats a line: ...never...

EXT. WHITE HOUSE - NIGHT
...the Rocks, Sharon Burke and Andrea Cameron discuss wargaming climate change. To get the most out of climate change wargames, planners should heed some lessons from the... is indistinctly said by a news program on a badly received TV channel

INT. TV NEWS SHOW- THREE WEEKS LATER
It's empty - almost inaudible insect buzzing.

INT. SMALL KITCHEN - SAME
an audio recording whispers something along the lines of: ...(Xinhua) -- Pakistani President Arif Alvi said on Tuesday that the adverse impacts of climate change are not limited at national borders and will affect the whole...

@hugovk hugovk added the preview label Nov 22, 2022
@greg-kennedy
Copy link

This is a good entry and I really liked the idea of cutting up movie scripts to come up with scene descriptions.

That said I could not help but laugh when the scenes of desolation suddenly jumped to "GOLD FIVE'S Y-WING COCKPIT" after some castles and indoor locations etc

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

3 participants