https://github.com/willf/nanogenmo-2021
https://github.com/willf/nanogenmo-2021
Last synced: about 1 year ago
JSON representation
- Host: GitHub
- URL: https://github.com/willf/nanogenmo-2021
- Owner: willf
- License: apache-2.0
- Created: 2021-11-02T23:16:16.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2021-11-03T00:05:52.000Z (over 4 years ago)
- Last Synced: 2025-03-24T12:52:25.408Z (about 1 year ago)
- Language: Python
- Size: 438 KB
- Stars: 2
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# NaNoGenMo 2021 Entry
## Will Fitzgerald
## Brainstorming Movies
**First of all: [The Results](nanogenmo2021.md)**
Here's the basic idea:
1. Get a big list of movie names, say, from [American films from 1920](https://en.wikipedia.org/wiki/List_of_American_films_of_1920)
2. Generate a prompt from this information to send to OpenAI's generate API.
3. Turn this all into a story.
Most of the code here is about managing text. Basically, things go through these stages:
1. First a [text file](table.txt) was created from [Wikipedia](https://en.wikipedia.org/wiki/List_of_American_films_of_1920). This was done manaually.
2. Then this is parsed, and a [JSONL-format](https://jsonlines.org) file is created, with parsed information, one per line: [table.json](table.json).
3. Then, this is enriched by creating a prompt which is sent to the Open API and then collecting the result and adding it to the dictionaries, again, one per line: [synopsis.json](synopsis.json).
4. Then, a script is run to create the result, which is called [Possible Movies](nanogenmo2021.md).
```bash
❯ python nanogenmo2021.py --help
usage: nanogenmo2021.py [-h] [--table TABLE] [--json JSON]
[--synopsis SYNOPSIS] [--story STORY] [--n N]
[--create_json] [--create_story] [--create_synopses]
@willf nanogenmo2021
optional arguments:
-h, --help show this help message and exit
--table TABLE input wiki file
--json JSON JSON file pre-synopses
--synopsis SYNOPSIS JSON file with Synopsis
--story STORY story file
--n N number of stories to generate
--create_json create json file
--create_story create story file
--create_synopses create synopsis file
```
That is, something like:
```bash
python nanogenmo2021.py --create_json
python nanogenmo2021.py --create_synopses
python nanogenmo2021.py --create_story
```
The prompt, for an original line like this:
```wikipedia
[[The Adventurer (1920 film)|The Adventurer]] || [[J. Gordon Edwards]] || [[William Farnum]], [[Estelle Taylor]] || Historical || [[Fox Film]]
```
is:
> The Adventurer is a historical movie directed by J. Gordon Edwards. It stars William Farnum and Estelle Taylor. Give a synopsis of the movie."