[X] My Assistant

I have a very important update on this topic.

Ghost Trick does indeed extract the images correctly. However, there are some cases where the extracted images are NOT of the highest available resolution.

.......

After writing the entire explanation and how to solve it, I realized that I can't post links because it's my first post....I will continue my explanation in another post. (IMG:[invalid] style_emoticons/default/cry.gif)

......

Now then....

I have a very important update on this topic.

Ghost Trick does indeed extract the images correctly. However, there are some cases where the extracted images are NOT of the highest available resolution.

If the .dlst you have looks like this in DLsite Play: [ibb.co] Example Everything is fine. You can download it and use Ghost Trick to extract the images. You will notice that the resolution of the extracted images corresponds to the ones listed on the website. (845x1200 in this case).

On the other hand, if the .dlst you have looks like this in DLsite Play: [ibb.co] Example Ghost Trick (for some reason) will extract images with lower resolution. I don't know if this is a problem with Ghost Trick or if DLsite is giving us .dlst files that don't contain the highest resolution.

I was able to solve this by developing an alternative method to extract the images. This method only works if what you want to extract looks like this in DLsite Play: [ibb.co] Example (like a "book")

The method consists of using the Chrome or Firefox development panel to log all network transactions made during the viewing of the work. Then we export this log in a .har file and open it in a Python script. This script extracts the protected (visually altered) images and can reorder them in the original position. The resulting image is of the highest available quality.

I compare this method using this material: [www.dlsite.com] Test
The extraction of Ghost Trick generates images of 1057x1500 and the method I propose generates images of: 1353x1920.

I don't have much time to develop the script better, but here you can find what I have been working on: [pastebin.com] Script

Notes:
1) You must ensure that the exported .har file records all navigation. That is, you must enter to view the manga, then open the developer panel, disable the cache, start the recording, clear the log, and refresh the page. After this, go through the entire manga. Only in this way can we ensure that the .har file has everything.
2) In the script you must replace the name of the .har file with the one you downloaded.
3) In the script you must replace the values of "frameW" and "frameH" with the width and height of the disordered fragments: [ibb.co] This dimensions. For this, you will probably have to first download one of the protected images and use some tool to calculate the dimension.


I will make a more detailed guide soon.

PS: This is my first post on the forum, although I have been using e-hentai and collaborating on works for years; I believe that this discovery is important to share.

I guess we still need a stable solution for melonbooks, since the github project doesn't work now. melonbooks prevent screenshot in browser and that firefox extension just simply won't give anything output. So genl now becomes the only option to break it. We need a second choice, once genl becomes unavailable there will be no way to get it through.

...

Hey, thanks for doing this! I just bought a book that's browser viewer exclusive and have been trying to figure out how to rip it. I tried your script, but it seems to get hung up right at the beginning when it reads the har archive:

CODE

Traceback (most recent call last):
  File ".\test.py", line 9, in <module>
    contenido = archivo.read()
UnicodeDecodeError: 'cp932' codec can't decode byte 0x81 in position 17838: illegal multibyte sequence

Any idea what it could be? I have bs4 and Pillow installed.

I figured out the issue. The open command had to be done in binary format for me, so "rb" instead of "r". I've managed to dump the images.

It took me a bit of tinkering with the frame dimension variables to get it to run at all. It seems if the dimensions go a bit over the intended then the program doesn't know how to handle it, and returns errors like this.

CODE

Traceback (most recent call last):
  File ".\test.py", line 96, in <module>
    imagen.paste(sections[realSectionHere], (i*frameW,j*frameH))
IndexError: list index out of range

Thankfully I managed to find a range where it started working, and I got results like this:
[i.imgur.com] https://i.imgur.com/QxqDNkg.png

I tried using the method you suggested to find the dimensions of the frame but sadly didn't work for me here. I measured 144x144, which was not it.

I did some manual trial and error and eventually got there. For this book it was 208 x 296. Warning: nsfw.
[i.imgur.com] https://i.imgur.com/edCWEhs.jpeg

I really hope you can figure out a nicer way to figure out the dimensions. Aside from that it works great.

This post has been edited by Zinbieel: Apr 20 2023, 01:58

anyone got a ripper for dlsite cypherguard pdf?

genl can do it, but it seems they are inactive for now.
I hope it's temporal, though.

Now then....

I have a very important update on this topic.

Ghost Trick does indeed extract the images correctly. However, there are some cases where the extracted images are NOT of the highest available resolution.

If the .dlst you have looks like this in DLsite Play: [ibb.co] Example Everything is fine. You can download it and use Ghost Trick to extract the images. You will notice that the resolution of the extracted images corresponds to the ones listed on the website. (845x1200 in this case).

On the other hand, if the .dlst you have looks like this in DLsite Play: [ibb.co] Example Ghost Trick (for some reason) will extract images with lower resolution. I don't know if this is a problem with Ghost Trick or if DLsite is giving us .dlst files that don't contain the highest resolution.

I was able to solve this by developing an alternative method to extract the images. This method only works if what you want to extract looks like this in DLsite Play: [ibb.co] Example (like a "book")

The method consists of using the Chrome or Firefox development panel to log all network transactions made during the viewing of the work. Then we export this log in a .har file and open it in a Python script. This script extracts the protected (visually altered) images and can reorder them in the original position. The resulting image is of the highest available quality.

I compare this method using this material: [www.dlsite.com] Test
The extraction of Ghost Trick generates images of 1057x1500 and the method I propose generates images of: 1353x1920.

I don't have much time to develop the script better, but here you can find what I have been working on: [pastebin.com] Script

Notes:
1) You must ensure that the exported .har file records all navigation. That is, you must enter to view the manga, then open the developer panel, disable the cache, start the recording, clear the log, and refresh the page. After this, go through the entire manga. Only in this way can we ensure that the .har file has everything.
2) In the script you must replace the name of the .har file with the one you downloaded.
3) In the script you must replace the values of "frameW" and "frameH" with the width and height of the disordered fragments: [ibb.co] This dimensions. For this, you will probably have to first download one of the protected images and use some tool to calculate the dimension.
I will make a more detailed guide soon.

PS: This is my first post on the forum, although I have been using e-hentai and collaborating on works for years; I believe that this discovery is important to share.

Hi bro
do you have a readme how this works. I like to make viewport capture of magazine much easier.

I have a very important update on this topic.

Ghost Trick does indeed extract the images correctly. However, there are some cases where the extracted images are NOT of the highest available resolution.

.......

After writing the entire explanation and how to solve it, I realized that I can't post links because it's my first post....I will continue my explanation in another post. (IMG:[invalid] style_emoticons/default/cry.gif)

......

how do you dump the jpg from DLsite?
I have tried many ways, but I failed

I guess we still need a stable solution for melonbooks, since the github project doesn't work now. melonbooks prevent screenshot in browser and that firefox extension just simply won't give anything output. So genl now becomes the only option to break it. We need a second choice, once genl becomes unavailable there will be no way to get it through.

Use Tempermonkey, in this link:
https://e-hentai.org/g/2604589/fba9093bd4/
可以用油猴插件，在我上传的这本里面有教程。

Is there a knife for Melonbook's digital book viewer thing?

Is there a knife for Melonbook's digital book viewer thing?

I did some research over these time and I hope my notes would be helpful.

I did some research over these time and I hope my notes would be helpful.

Your notes are leading to a dead end.

Your notes are leading to a dead end.

yes because it's a personal note written inside subforum GSC so you won't be able to see it if you aren't a GSC member.

yes because it's a personal note written inside subforum GSC so you won't be able to see it if you aren't a GSC member.

Ah, I see (goddammit).

I did some research over these time and I hope my notes would be helpful.

Btw the Melon ripper has some wrong page orders and duplicate page issues. Just to report.

I did some research over these time and I hope my notes would be helpful.

yes because it's a personal note written inside subforum GSC so you won't be able to see it if you aren't a GSC member.

I sort of get it, due to the specific topic at hand, but you are knowingly gatekeeping, and I hate gatekeepers. Would have been better to not say anything.

Just my two cents.

Use Tempermonkey, in this link:
可以用油猴插件，在我上传的这本里面有教程。

Seems to be OKish however the file sizes are bloated and some detail is lost since this script is what i am assuming is ripping from the canvas.

Is there a knife for Melonbook's digital book viewer thing?

The github thing still works fine for me, not sure if this is what you are having a problem with in the reply above.

Btw the Melon ripper has some wrong page orders and duplicate page issues. Just to report.

No, this was on me. What happened was my viewer when opened was on like page 20 or whatever, so when I actually had the ripper going I had to back track and was fumbling the book so to speak as it was snatching everything it saw.

Note, make sure you have single page view and you're on page 1. Also make sure you "flip" accordingly through and don't backtrack. Otherwise it's fine for the most part if you're looking for a quick fix.

The github thing still works fine for me, not sure if this is what you are having a problem with in the reply above.

Were you using a different one? I tried the one that works on Firefox only but I wasn't sure how to fully operate it.

This post has been edited by Super Shanko: Sep 1 2023, 05:48

Were you using a different one? I tried the one that works on Firefox only but I wasn't sure how to fully operate it.

Dont know of the firefox one, so yeah i guess using a different one.