search

meta-automata-nix / actions-batch

Time-sharing supercomputer built on GitHub Actions
https://blog.alexellis.io/github-actions-timesharing-supercomputer/
MIT License
1 stars 0 forks source link

Sweep: Create a script that adds unlimited storage via the infinite-storage-glitch #25

Open ar4s-gh opened 1 week ago

ar4s-gh commented 1 week ago

Familiarize yourself with the below README.md, consider the logic needed to build the required "add-to-isg-store.sh" script, and impliment the directives that follow the readme.

Reference repo: https://github.com/Nick4421/ISG-2.0

1. First, step-by-step, draft a bash script in /examples named "build-n-run-isg.sh" that does the following:

List of all the Netpbm programs, search for pamscale and pamenlarge to see what they actually do and how they work.

(There is a Makefile in src/src_code, but the two source files encode.c and decode.c don't have any dependencies outside of the standard libraries, so you probably don't even need the Makfile when compiling.)

2. Install Ollama, "Docker for AI"

sweep-ai[bot] commented 1 week ago
Sweeping

0%

Actions

Thank you for submitting this issue. Unfortunately, the scope of this request is too broad for our AI assistant Sweep to handle directly. Sweep is designed to make targeted code changes within an existing codebase to fix bugs or add small features.

This issue requires cloning repositories, building and replacing executables, installing new dependencies, and making system-level configurations. These types of large-scale changes and external operations are outside of Sweep's current capabilities.

To move forward, I would recommend breaking this down into smaller tasks and handling the system setup and dependencies separately. Once you have a specific code change that is needed within the codebase itself, feel free to open a new issue and Sweep can assist with that part.

Please let me know if you have any other questions! I'm happy to provide further guidance.

Report a bug.

[!TIP] To recreate the pull request, edit the issue title or description.

This is an automated message generated by Sweep AI.

ar4s-gh commented 1 week ago 

This is a way to embed files in a video. Taking inspiration from DvorakDwarf's infinite storage glitch project, I decided to make my own version using C as well as some shell scripting. The basic idea is to embed a file in a video, upload that video somewhere, such as YouTube, and essentially use that as a file system. Because YouTube has no limit to how many videos a user can upload, you can theoretically store as many files there as you want (although this is probably against their terms of service, so use with caution).
Dependencies and getting started.
I using MacOS in a Linux environment, and making it run on other peoples machines the same can be difficult, so if your version doesn't work, here are a few things to try:
Download the correct version of ffmpeg here and replace the existing executable located in src/executables.
Build your own versions of pamscale and pamenlarge from the Netpbm library here and replace the existing executables in src/executables.
There are instructions on how to build it in the library itself.
List of all the Netpbm programs, search for pamscale and pamenlarge to see what they actually do and how they work if you are interested.
Compile your own executables from the source code in src/src_code and replace the existing executables in src/executables.
There is a Makefile in src/src_code, but the two source files encode.c and decode.c don't have any dependencies outside of the standard libraries, so you probably don't even need the Makfile when compiling.
The shell script uses ANSI format in some places to make the terminal output look nicer, so if you don't have this, it might look screwed up, but it won't affect the functionality of the program.
How to Use
Simply run the executable run, and you'll be prompted to encode or decode a file.
Encoding a File:
You'll be prompted to enter the file path to your desired file, with the current directory being ISG-2.0.
If the file path does not lead to a file, you can either enter another file path or go back to the main menu.
You'll be prompted to enter the name you want the encoded video to have.
If you enter a name, it must end in ".mp4". If it doesn't, you'll be prompted to re-enter a name.
If you just press enter and leave the name blank, it will default to being named "output.mp4".
You'll finally be prompted to enter a path to where the final video will be stored.
If the file path leads to an invalid directory, you will be prompted to re-enter a path to a valid directory.
If you press enter and leave this blank, the final video will be placed in the current directory, which is ISG-2.0.
Decoding a Video:
Very similar to encoding a file.
You'll be prompted to enter a path to a video.
If it is an invalid file path or the file does not end in ".mp4", you can either enter another file path or go back to the main menu.
You'll be prompted to enter a name for the final decoded file.
If you leave this blank and just press enter, the name will default to "output".
You'll finally be prompted to enter a path to where the final file will be stored.
If the file path leads to an invalid directory, you will be prompted to re-enter a path to a valid directory.
If you press enter and leave this blank, the final file will be placed in the current directory, which is ISG-2.0.
After encoding or decoding a video, you will be returned to the main menu where you can either encode, decode, or quit.
How it Works:
Every piece of digital information can be broken down into a series of bits. We can use this to store information in a series of images using black and white pixels. In the raw pbm image format, a single bit represents a pixel. A 1 is a black pixel and a 0 is a white pixel. Using this information, we can store any file as a series of raw pbm images which can then be turned into a video. Turning the series of images into a video takes up more space than the original file, which may seem counter-intuitive, but we can store this video on websites online that have no limit on number of uploads per user to essentially gain unlimited storage of files.
Encoding:
Encoding a file into a video can be broken down into 3 main steps:
Create a series of raw pbm images (learn more about pbm images here).
Scale each image up by a factor of 2 using pamenlarge.
Make a video using each of the images as a single frame using ffmpeg.
Explanation:
When encode runs, it first finds the number of bytes in the desired file. It then gets each byte from the file as a char and copies it into a raw pbm. When one pbm image fills up, a new one is created, and the process continues until all the bytes of the original file have been copied over. Due to video compression, a single pixel getting flipped can mean corruption for the file. To combat this, we scale the image up by a factor of 2 using pamenlarge, so each pixel now becomes 4 pixels. This way when pixels do get flipped, the image does not become corrupted as it has been scaled up. Once each image has been scaled up, it is turned into a video using ffmpeg where each image is a single frame. Once this step has been completed, the process is complete and the file has successfully been embedded into a video.
Decoding:
Much like encoding a file, decoding a video can be split up into 3 main steps as well:
Separate each frame of the video into a raw pbm image file using ffmpeg.
Scale each image down by a factor of 0.5 using pamscale.
Copy the image raster data into a single file using decode.
Explanation:
Decoding a video first starts with turning every frame into it's own raw pbm image using ffmpeg. Since each image has been scaled up by a factor of 2, we use pamscale to scale each image back to it's original size. In order to eliminate the effects that video compression had on the images, we use the -nomix flag for pamscale. When scaling down, this option looks at each 4-pixel block and sets the correspoinding pixel in the scaled-down version to the color that shows up the most. This approach seems to work good enough to eliminate the effects of compression. Finally, decode goes through each images and copies the raster data over to a final file. Once this is done, the video has been successfully decoded.```