Code named Gadzorp, the entire point is to serve as a reference with scripts from every episode as well as all the GIFs that I have created for the show. The biggest issue at this point is that I need to make time to manually go though each script and revise them as needed (as well as add scripts for the deleted scenes). But in the meantime, please check it out and give me some feedback! Tell me how it breaks, try to hack into it, all that stuff.
Looks great. Seems like it’ll be handy for searching quotes (which is hit or miss with most search engines; I mostly start directly with Yarn these days)
How are you doing the transcripts? Just pulling from the extracted
.srtfiles?How are you doing the transcripts? Just pulling from the extracted
.srtfiles?Almost. My Blu-ray rips are subtitled with PGS/SUP, which apparently uses an image for subtitles instead of simple plain text like SRT. So I found a program called sup2srt which uses an OCR engine to convert the subtitle images to SRT, which I can then load into the database that this site uses. But, there are some issues with that:
- Obviously, OCR is imperfect. It did a great job overall, but there were still a lot of things that need to be fixed
- Most seasons, the subtitles are ALL CAPS.
- There are a few scenes with hardcoded subtitles (such as the scene in Queen of Jordan in which Jack/Liz are whispering). These lines are not part of the subtitle file
I made some quick Python scripts that go through all the script files to fix common mistakes, but I am also slowly going through each script individually to manually revise them.
Wow! Very impressive. I can’t give feedback beyond that yet.
Oh, one thing, loading the Animations page slowed my whole browser down for a minute or two. (Laptop, Zen Browser 1.13.2b, Linux Mint 22.1.)
Did it slow down just while the page was still loading, or for a minute or two after the page had finished loading? The Animations page does download 17MB of thumbnail images, which may be a bit much.
Just while the page was loading.
I think I should make it not download all 17MB of thumbnails when the page is loaded, and instead download each thumbnail individually as it is scrolled into view.
Sounds good.
