|
--- |
|
license: other |
|
viewer: false |
|
--- |
|
License link: [https://www.gnu.org/licenses/agpl-3.0.en.html](https://www.gnu.org/licenses/agpl-3.0.en.html) |
|
|
|
## What is this |
|
|
|
2500+ Screenplays scraped from the scriptlab for science purposes |
|
|
|
## Different folders |
|
|
|
`pdfs` includes screenplays as PDFs, this is the raw scraped data. |
|
|
|
`text2` includes PDFs converted into .txt directly, some pdfs have formatting issues and some are outright blank or have obfuscated text. |
|
|
|
`texts3` includes OCRd PDFs which mitigate the previous issues but add issues with slight character or formatting errors (like inconsistent newlines) |
|
|
|
Note: Originally we were going to do OCR, however, some morons put image decompression bombs and I can't be bothered to fight that, for now around 1/6 pdfs with DRMs won't be available as .txt format, feel free to convert them to .txt yourself and send a merge request though! |
|
|
|
Obviously the TAR files are compressed versions of the folders. |
|
|
|
It could be helpful to use an LLM to clean it up! Make sure the LLM doesn't change any of the actual contents. Otherwise I don't have enough time to do it but still want to release this golden trove of data. |
