Can you explain more how you went about this process of extracting Unity assets? Did you "unbundle" the Unity game assets to peek at the coding? Aren't these things precompiled?
I haven't played your port yet (about to), but judging from people's comments this is an impressive work. Congratulations.
Sure, I don't know how much detail you want, just ask about anything you specifically want to know more about. I used a bunch of different things to get the parts I needed for the game.
Images first as that was the most straight forward. Up until version 2.9 I was able to unpack the data.unity3d file via AssetStudio. This is nothing new, people have been doing it for this and other Unity games for a long time to share the art without the game. Then I got to sorting the 6500+ image files. After that, I assigned each image one at a time matching the running the game on one screen for parity and my IDE on another.
I also used uTinyRipper to pull most of the unity project out so I could see what scenes were in the game, as well as structure. I was really hoping it would let me grab the dialogue, but, nope. Additionally, I was still trying to figure out how to get Real2D to get the animations and naïve hoping I could just plug it into a Unity project and have it magically work. Sadly, this method doesn't preserve the Unity Mono code links, so while I got the scenes, no scripts were hooked into them, making them mostly worthless other than getting the scene names, which did help later on.
For the animations, I just used windows+G to capture the video, then edited into a loop and bam. Removing the text and icons is WAY more technical, but I can go into details if you want, short version is it required editing the C# files via dnSpy.
Version 3.9 however wouldn't unpack, so, I had to come up with a way to get the images out. I ended up writing an auto screen cap program that advanced the dialogue one step, captured, advance, capture, and so on. The I had to go through and throw out all the dups, or most of them. As with the animations, I hide the text and icons while capturing.
To get the text, originally I typed it by hand, since I only wanted to do a few sense. As I moved on to do the full game, this was unstainable. So I wrote my own OCR application, then ran it through each scene to get the dialogue. It wasn't perfect and it transcribed a lot of errors that I had to go fix, but it saved me a lot of typing.
For character poses, I did them all hand by hand, one line at a time. The characters are texture files, broken into their sub parts like left eye, right eye, mouth 1, mouth 2, etc... all in one big image for each character. I cut the parts out then reassembled every character, plus outfits in my art application (stuido paint) and exported them in a more friendly renpy format for layered images.
Early on, I kept a running save file, so that I could help track event times, but this sadly kept breaking, because the Unity version is buggy. So I started modifying my save file to replay scenes, but some scenes were so broken they still wouldn't play. Eventually, I just directly altered the C# code to jump directly to scene names.
If you have more questions, feel free to ask, and I'll hit you up with another wall of text