script_buddy_v2

Script Buddy v2 is a film script text generation tool built using film scripts from the world's most popular film scripts.

Read more about the project on Medium

I'm also a bot on twitter, I tweet a new script snippet every hour or so, follow me at @script_buddy

Project Structure

• imsdb_scraper contains the scrapy spider built to crawl IMSDB.com and download the film script data that is used to train our text generation model.

  • To run the scraper change directory to imsdb_scraper with cd imsdb_scraper. As I haven't built any pipelines currently to run the scraper and save it's output simply run scrapy crawl scriptSpider -o scripts.json -t json and for now manually move this output file to script_buddy/data/ and you'll be good to go.

• script_buddy contains the model fine tuning notebook, utility script parsing functions, tweepy bot code, and streamlit application code.

  • If you'd like to run the streamlit app locally

    • git clone https://github.com/cdpierse/script_buddy_v2.git
    • pip install -r requirements.txt
    • cd script_buddy
    • streamlit run app.py

  • This should open an instance of the app for you to experiment with. Running it for the first time may take a while as the model needs to be downloaded from huggingface's model hub.