WebSocketWithOpenCV Project

This project is a WebSocket server that uses Boost, OpenCV, and C++ to perform image processing. The server processes images, detects edges, and sends a confirmation message to the client upon completion. The project includes support for displaying processed images on the backend and can be configured to save output images instead.

Installation

With Nix

The preferred way to build this project is with the Nix package manager, which builds in a hermetically sealed and reproducible build environment as well as providing critical build dependencies. Nix will automatically manage the build and orchestrate meson and ninja.

Note that Nix is only supported on macOS and Linux. Windows users should use first install the Windows Subsystem for Linux (WSL) and then follow the below instructions.

1. Install Nix:

curl --proto '=https' --tlsv1.2 -sSf -L https://install.determinate.systems/nix | \
  sh -s -- install

Important

If you already have Nix installed and did not use the above installer, you must enable the experimental "flakes" and "nix-command" features.

2. Build the program

nix build

3. Run the program

result/bin/WebSocketWithOpenCV

Alternatively, build and run it in one step with

nix run

Hacking

To hack on the code, it is not convenient to re-run nix build upon every code change. A devenv development shell is provided so you can directly execute meson, ninja, and take advantage of incremental compilation.

Important

It is highly recommended that you use the methods described below if you are going to be hacking (developing) on the code. Directly invoking ninja vastly speeds up subsequent compile times, since Nix cannot take advantage of incremental compilation.

Follow the instructions to install nix as detailed above, then run

nix develop --impure

Alternatively, if you want to retain your current shell environment and seamlessly layer the development environment over it, you can install direnv, then run

direnv allow

Now, your shell will automatically be hooked into the development environment as soon as you enter the project root.

The development shell provides various pieces of tooling like meson, ninja, pkg-config, etc. It also provides the libraries and dependencies necessary to build the project.

Once you are in the development environment, run

meson setup build

This will use meson to generate the build/ directory containing build orchestration files. If you are using a language service like clangd, this also provides the compile_commmands.json required for it to work properly.

Enter the directory and execute ninja:

cd build
ninja

This will produce the WebSocketWithOpenCV binary.

Important

If you add new libraries to Nix and meson.build, it is likely you will have to rm -rf the build directory and regenerate it with the commands above, due to some quirks with the Nix environment.

Additionally, you should exit the nix develop shell and re-run nix develop --impure if you modify any nix files.

Formatter

The treefmt autoformatter has been configured in treefmt.toml to automatically format the entire tree.

To invoke it, please install Nix, then run:

nix fmt

We recommend you run this commmand before any commit to ensure your code meets quality guidelines. Continuous integration will run on all commits to ensure lints and formatting checks pass.

Solution without package managers, only Meson

Warning

This approach is not recommended and should only be used if Nix is strictly not available for whatever reason.

Before setting up this project, ensure that the following libraries and tools are installed on your system:

Note

These are only required when compiling manually. When building with Nix, all dependencies are provided automatically.

• Boost (including Boost.Asio and Boost.Beast)
• OpenCV (version 4.0 or higher)
• libtorch (v2.5.0 or higher)
• liboai (v0.4.1 or higher)
• A C++ Compiler (supporting C++17 or higher, preferably clang)
• CMake (version 3.10 or higher, required to find dependencies)
• Ninja
• Meson
• nlohmann-json (version 3.11.3 or higher)
• libcurl 7.78.0 (version 3.10 or higher)

Using this approach, you will have to manually provide the dependencies above, placing them somewhere that pkg-config or CMake or Meson can find them.

1. Clone the Repository

   git clone <repository-url>
   cd WebSocketWithOpenCV
   

2. Install dependencies If you already have dependencies on your computer I still recommend you to have a libs folder that includes the cloned version of all the required dependencies(Both headers and the compiled code). It is then set up in the CmakeList to then find it relative to your project directory. If you don't then it still works, I'll made it so cmake accounts for that!

3. Build and server (1) meson setup build && cd build (2) ninja