Suggestions to improve usability of generator

[pharapeti]

Hey, love the project. I've got a few suggestions, let me know what you think.

1. Create a sample config.json which contains default values, all set to null. This allows a user of the package to set config values before executing the program. After running the generator, any null or unset config values are populated when the --save-dev flag is set.

2. Update the setNames method to use the filename as name for each file, instead of requiring the user to input a name for each file. For example:

async function setNames(trait) {
  const files = fs.readdirSync(basePath + '/' + trait);
  files.forEach((file, i) => {
    names[file] = file.split('.')[0];
  });
}

3. It is fairly tedious to have to pre-calculate the number of images containing each trait, and entering an invalid number can result in the program crashing. To remedy, the weighting algorithm can adopt a percentage weight system for determining the weightings of each trait.

For example:
What percentages of ' + names[file] + ' ' + trait + ' should there be?'. To which the user would input 5 as in 5%.

Followed by What is the total number of images that should be generated?. To which the user would input 1000 (sum).

Using these values, you would be able to determine the number of images containing each trait as a percentage of the sum provided by the user.

Cheers!

[NotLuksus]

Sound like great ideas! I'm currently working on percentage based inputs since that is where most people have issues. I will check on the other ideas ones I'm completed with them, if you want you can fork it and create a pull request ones you are done!

[werksone]

Would absolutely love to be able to use the file name as the setName

[AkoZ]

yeahhh good good !
thank you for this generator, even for now not used.
some Jupyter to be found for ez use, and take dir with traits

• your last explanations about the 50 bckground and 25x2 as in fact 2 layers of 25 traits is not all clear ;)

• also there is : https://github.com/benyaminahmed/nft-image-generator which may be identical in jupyter mode.
  Well ! best regards ! and may it appears ;)

AkoZoOm

[FrostyWasTaken]

Sound like great ideas! I'm currently working on percentage based inputs since that is where most people have issues. I will check on the other ideas ones I'm completed with them, if you want you can fork it and create a pull request ones you are done!

Bro I would really find this useful have you gotten around to completing it?

[dcardines-design]

Guys whenever I run it it throws an error at the end of the process:

Error: Couldn't load image
(node:196704) UnhandledPromiseRejectionWarning: Error: Couldn't load image at img.onerror

Anyone has any idea how it can be fixed? Much appreciated

[dcardines-design]

Sorry solved it, it was just the count of traits, the totals of each trait group didn't match

[smehrjerdian]

1. develop program to read inputs from a csv template where each trait folder matches a sheet name in the csv and all trait names are in column A and values in column B

2. support a mix of gif + png layers and gif outputs to generate animate collections

[tucq88]

Some traits are must have, some are should have which have lesser than must ones. Now collection is generated based on the lesser. The (3) will resolve such problem.