<a target="_blank" rel="noopener noreferrer nofollow" href="https://user-images.github

Implementing Pattern Recognition for OCLL about puzzle-gen HOT 5 CLOSED

adiwastu commented on June 11, 2024

Implementing Pattern Recognition for OCLL

from puzzle-gen.

Comments (5)

adiwastu commented on June 11, 2024 1

Your lead worked wonders nevertheless! And I did get that done just several hours ago. Since I only cared about the corners of the last layers, that is 4 * 3 stickers, I pulled those info and evaluated whether they got the "U color" value. And then it's just the matter of joining them and comparing them to some literals (the 7 * 4 OCLL patterns/angles) and that's it! Here's a pic:

Cheers!

from puzzle-gen.

tdecker91 commented on June 11, 2024

Thanks for the feedback.

It is certainly possible and I'm considering what is the best way to expose that. I think I'd like to know the use case a little bit more before knowing how to handle it.

For some context, one of my goals with this library is to provide the puzzle rendering capability and be pretty agnostic to use case so the library can be utilized in different applications.

My first thought is that the application asking for an OLL render already knows the case in mind. So perhaps just providing a way to add a text label to the image itself would be sufficient?

Or perhaps the application is looking at a random scramble and trying to detect the case, I think this is more appropriate to include the logic outside of the puzzle-gen library, but expose the internal state of the puzzle so it can be used. I'm assuming this is more what you're after?

The good news is that the puzzle state is already exposed somewhat.

The modules are arranged kind of like this

rendering
- PNG
- SVG
- Canvas
visualizer
- simulator
- renderer

with PNG SVG Canvas just convenience methods to build a visualizer and output the image to the HTML document. Which is the core purpose of the library. A visualizer keeps the state of the puzzle in the simulator and then uses w/e renderer to output depending on the mode selected (png, svg, canvas)

the SVG and Canvas methods return the visualizer object and the puzzle state can be inspected to do what you're looking for.

from puzzle-gen.

adiwastu commented on June 11, 2024

Hey, thank you! I would love to explain my use case by quoting your comments and perhaps explain how mine might be a bit too niche:

My first thought is that the application asking for an OLL render already knows the case in mind.

So what I have now is around 25k of algs that I'm sure what it does (it affects OLL) but not quite sure what OLL it gave. The algs are generated by combining 2 algs, and that is why the case is not known to the application. I tried bruteforcing the algs, but could only handle filtering images for like a couple hundred combinations before getting tired (lol). That brings me to my next small code, basically I used puzzlegen.min.js to do the work of displaying the algs!

  <head>
    <script type="text/javascript" src="puzzleGen.min.js"></script>
  </head>
  <body>
    <div>
    // some div ...
    </div>
    <script type="text/javascript">
      function generatePuzzles(algorithmsId, containerId) {
        let algorithms = document.getElementById(algorithmsId).value.split("\n");
        let puzzleContainer = document.getElementById(containerId);
        puzzleContainer.innerHTML = "";

        for (let i = 0; i < algorithms.length; i++) {
          let options = {
            "width": 100,
            "height": 100,
            "strokeWidth": 0.03,
            puzzle:
            {
              alg: algorithms[i]
            }
          };
          let puzzleDiv = document.createElement("div");
          puzzleDiv.id = "cube" + (i + 1);
          puzzleContainer.appendChild(puzzleDiv);
          puzzleGen.PNG("#" + puzzleDiv.id, puzzleGen.Type.CUBE, options);

          // Add an indicator text to show the current index
          let indicatorText = document.createElement("p");
          indicatorText.innerHTML = "Generating puzzle " + (i + 1) + " of " + algorithms.length;
          puzzleDiv.appendChild(indicatorText);
        }
      }
    </script>
  </body>
</html>

As you can see, a super simple one. If I could somehow pull the state of the cube (like what color UFR is, or basically every sticker) I could very easily make logic to differentiate the case. And as you said here:

the SVG and Canvas methods return the visualizer object and the puzzle state can be inspected to do what you're looking for.

could you please guide me on how to pull those state from the objects? Thanks a bunch!

from puzzle-gen.

adiwastu commented on June 11, 2024

Hi Tyson,

Looks like I've found the way to expose the stickers that could work—by messing around on the console and from your lead (visualizer). This code shows the map and I am able to use the get() method to extract the stickers that matter, with LL_state being an object that is returned by the Canvas method.

LL_state.simulator.stickers;

Closing this issue. I might document what I did when my project is done, and for whoever reading this and find this pokes your interest, please expect some proof of concept codes on my page in the future. Thank you!

Best,
Adi

from puzzle-gen.

tdecker91 commented on June 11, 2024

Yup exactly.
Sorry, I was planning to write out a full example / description on the weekend when I had some time. But you get the idea.

To be honest it might not be very straightforward but you should be able to accomplish what you need. The simulator stuff wasn't really built with that in mind, but if it works that's great!

from puzzle-gen.

Implementing Pattern Recognition for OCLL about puzzle-gen HOT 5 CLOSED

Comments (5)

Related Issues (5)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent