Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
if anyone has improvements please do a pull request
what i've noticed is that there is alot of repeats in the final product so eventually someone will generate your ship/freighter with my script. and then it will be in the git repo.
alternatively you could run it and then do a pull request yourself.
also i'm not sure what you are asking in the last sentence... you want me to manually visit a system and take a picture? That isn't the purpose of this project... nothing is done manually :D
we give the wheel to RNGesus
https://www.nmseeds.club
thats for manual seed gathering though. not sure they would appreciate 10s of 1000s of force generated seeds as I can't link them to a system or source aside from the seeds themselves.
edit:
update.. the author of that site asks people to contact him/her and has their contact information on a subpage hidden in a hard to see blurb. the contact address is info@nmseeds.club
:) (as replied per email) would love to try and get this working but we do need to find a way to also get the data: type, seed, at least 1 colour, but 2-3 would be ideal. description isnt so important, but i dont mind if you want to drop a link to your github page in there.
First lets see how much data can be grabbed in the process, OCR could be a good approach. For everything else, i'd be happy to set up a "crowd sourcing" section of the website where users can browse through any items missing information and can then submit the missing information before moving on to the next image.
ok great let's keep the thread alive for now i'm going to think about a few solutions.
regarding colours... i can propose a statistical approach to determining colour but that might require too much processing per image hahah.
solution 2 is to redo everything and rename screenshots as-they-are-generated to have the seed name.. this is more possible and can be done with a .bat file passing an argument. this is easier than solution 1 but still will require some effort and will require regenerating the (not many imho) 100s of seeds i have done so far.
solution 3 is to just take the files as-is and have a separate section with "generative seeds not submitted by users" and then let people "claim" them and add details. the value proposition here is that only the seeds that people actually want will be part of the main-searchable DB.. best thing here would be to just link directly to the github folders so you dont use your bandwidth. no need to have indexed the eventualy 10s of 1000s of seeds on your site if people don't like everyhting (also there are alot of repeats). github has thumbnail folder and the images so you can easily set things up for previews etc.
solution 2 vs solution 3 is really the question of
Do you want 10s of 1000s of items ? many of which will be 100% identical to each other? that is solution 2.
Do you want the 10s of 1000s of items but have a curation layer? that is solution 3
might also look at automatically cropping the generated images a bit on import so they are more focused on the actual item :)
also, 4k isnt needed, but HD would be nice ;-)