I suspect this information is somewhere in the forum, but if so I cannot find it. Apologies if I'm duplicating.

I want to build a patch to shuffle-play a number of audio tracks, with no repeats until they have all been played, and then start again with a new shuffled list.

I have already made this 'shuffle' subpatch to avoid consecutive repeats:

What would be the best way to build something that will shuffle numbers 0-3, then only allow repeats of any of these numbers once all of them have been produced? It would then need to begin again from the beginning and run indefinitely in this cycle.

Thanks!

Check vanillaUrn from here https://github.com/alexdrymonitis/miscellaneous_abstractions.

I've always liked motex/shuffle because it allows you to specify how soon a repetition can occur when you generate a new permutation. See the third argument:

@jameslo said:

I've always liked motex/shuffle because it allows you to specify how soon a repetition can occur when you generate a new permutation. See the third argument

can you better explain how this works?

[else/rand.hist] allows an unrepeated sequence but also allows a bigger sequence with 2 or more equal repetitions of each item.

But else also offers [rand.u] for a one shot unrepeated sequence

@porres Here's a comparison with cyclone/urn. Both are generating 4 element permutations, but urn sometimes generates a permutation that starts with the same number that the previous permutation ended with. Setting shuffle's 3rd arg to 0.25 guarantees that won't happen with 4+ element permutations.
motex shuffle explanation.pd

nice, thanks, I will add this functionality to else/rand.u now

Just added this parameter. Unfortunately motex is not available for me on apple silicon...

Thanks to all who have already replied here! Apologies I didn't think to specify this in my original post but an important detail here is that I need to make sure that no consecutive files play ever. With many of the objects helpfully suggested above, there is a risk that the last number in a random sequence will be the same as the first number of the next random sequence. I understand this has already been discussed above but seems worth mentioning here.

Sadly the example random-wo-repitition.pd produces occasional repeats when it restarts the sequence.

Something like the 3rd argument on motex/shuffle would be ideal. However, I am using pd 0.56.2 on a raspberry pi 5, and motex is not available to me via deken for this architecture.

@porres thanks for adding the hold feature to else/rand.u - I downloaded the latest else library and I don't think the external has been updated yet for pi/linux? I don't see the hold argument on the list in the help window for the object.

If this isn't going to be feasible on my architecture I think a good alternative would be for me to create a long list of numbers without repeats and prompt these in order with a bang: can anyone recommend an object that will do this and loop back around to begin the sequence again once it has finished? This would at least create the illusion of randomness, as I could make sure it doesn't repeat.

Thanks again to anyone who can help with this!

@Dizzy-Dizzy said:

I downloaded the latest else library and I don't think the external has been updated yet for pi/linux?

nope, not yet... but you can get it from github actions

a real update might come in a month or so

@lacuna Clever! That [t f b] is doing all the work!

Thank you @lacuna for sharing this patch. I built it according to the image you sent and it doesn't seem to produce all of the numbers before permitting a repeat:

For example, the run of numbers 1, 0, 2, 3, 2, 0, 1. Maybe I'm just getting confused- thinking about it, it's possible for any specific number to have up to 6 other numbers between them, if they were randomly selected as the first in sequence of four, and then in the last in the next sequence of four.
Thank you! You are all waaay ahead of me on this.
@porres I should probably learn how to get externals from github. It's a little complicated as in this instance I'm running the patches I'm building on 3 raspberry pis all connected via wireless internet: I often find it challenging to build on my laptop and then migrate the patches smoothly onto the pis.
Thank you for your help everyone, I really appreciate it.

I don't quite understand @Dizzy-Dizzy's issue description, but I did notice that @lacuna's patch doesn't generate complete permutations after the first one. I think you have to add the excluded # back into the pool at the right point. Sadly, my way of doing that significantly spaghettis-up the patch, but maybe someone can find a simpler way to do the same thing.
lacuna permutation wo repetition.pd

I also wonder if it's not the worst idea to just generate random permutations and simply throw out the ones you don't like, e.g. ones that start with the same # as the end of the previous permutation. I've done that in my own work, especially when I think the sound is more interesting than the programming.

randomwmemory.pd

There are many ways to do this ... @Dizzy-Dizzy Please read the patches and try to understand them.