Waryyn's page

7 posts. 1 review. No lists. No wishlists.


RSS


Dr A Gon wrote:

If you want art for bestiary creatures it's in Nethys. The images in the books are higher resolution but they're often cropped to fit.

The easiest way to script getting art for monster manual books is to scrape google image search or use the bing api. I've done that myself. Those engines' algorithms are good enough that almost always grabbing the first 10 will get you what you were after. (EDIT: And of course, in publishing the script you aren't distributing anyone else's content).

I agree that this is by far the easiest solution, but I think normalizing this kind of script where users can get the art for free is dangerous, as it will be hard to transition users back to paying for art if they just get it for free. Paizo is extremely nice to allow AoN to host the art, but if everyone just starts scraping it for art, that may be reconsidered as it limits their ability to earn money on it. For this reason, I think it would be best for solutions for getting images on VTT should be based on PDFs, as these are at least bought from Paizo.


But would how easy the image extraction is, change if it is still within fair use? So far most of the examples given are rather time-consuming and cumbersome.

We can take as an example the bestiaries. Using CUP we are allowed to use the names of creatures for community use. It is therefore possible to create a simple mapping of (page number, image number) -> creature name, and make an automated script that takes these images out and gives them the correct name. This script could easily be shared, such that everyone who owns the Bestiaries easily could create a folder of named images. Would a script like these, which actually have the names included, be ok? And would such a script require watermark checking to be within fair use, or is that not an actual requirement?

If that is within fair use I would be happy to create such a script, I have most of the code, and the mapping can be done almost automatically using image similarity search.


I really think a lot of the needs of the community could be taken care of if some guidelines were developed with regards to what code we can share and develop when it comes to pdf extractions.

To give a very concrete example. Extracting the pawn images with correct names from the pawn pdfs, thereby getting the art people are requesting for running on VTT, is extremely simple with a script due to the design of having the name on top of the image. But the big problem is if we can develop and share this kind of script without breaking the fair use agreement from Paizo. Is it for example enough if a script has a watermark checker as PDF to Foundry package does, or is there a difference between maps and token art?.


Hi all, this is potentially not the correct forum, but I could not find one which was fully applicable.

Does anyone know what Paizo's stand is on sharing tools for automatic extraction of content from their pdfs intended for private games? With more and more games going online, there seem to be more people asking for automatic ways to get pawn art for monsters without having to manually copy all images from a pawn pdf. It is e.g. relatively simple to write a generic* scraper for the pawn pdf's which extracts the images and uses their names as written on top of the image for saving, so it is easy to use in an online game. The big question is if sharing code like this is seen as inappropriate by Paizo.

Does anyone know how Paizo has previously ruled on automatic extraction of content from their pdfs?

* With generic I mean a scraper that is not coded in any way towards Paizos pdfs specifically and can be used in a variety of different pdfs.


1 person marked this as a favorite.

Hi,

I am writing to be sure that I am not breaking any fair use agreement with regards to Paizos pdfs. I have written a script that given a Paizo Token pdf, automatically extracts all creature images, and saves them using the name of the creature as given by Archive of Nethys, and if no match is found save it using the creature name in the token pdf. This allows semi-automatic importing of images to Foundry VTT. If it changes anything, I can obfuscate the code and insert a watermark checker to ensure the validity of any provided token pdfs. Is it within the community guidelines to share code like this, or is it only allowed for personal use?

Just to be sure, the only Paizo content in the actual script is the name of creatures for name mapping, which as far as I understand is ok to use.


xNellynelx wrote:

I'm back! I looked ahead because something dawned on me worth pointing out. More spoiler tabs

** spoiler omitted **

Thanks alot for your very nice response ! Shared your response with my player today, and they decided to go with the flame Oracle :)


Hello all, I am the GM of a group who have just finished book 2 and going to start book 3 now. With the APG out, one of my players fell in love with the flame Oracle, but knowing the overall dragon theme of the adventure path, wanted to be sure that it was viable with regards to resistances. Anyone having GM'ed further than book 2, what are your experiences with the number of enemies with fire resistance, and would a flame Oracle have a bad experience ?
Thanks in advance :)