wildbg-c.md

Crate wildbg-c

This crate contains a small C library to access some functionality of wildbg.

In contrast to the web API this library offers fewer features and requires more manual work to set up. Only use it if you have existing C code and no other way to connect that to wildbg.

You can see the API in the header file: crates/wildbg-c/wildbgh.h.

How to use this library from your C code

Execute the following from the project's root folder.

1. Build the library

cargo build --package wildbg-c --lib --release

2. Copy the library to your C project

The shared library file created may be called wildbg.dll or wildbg.so, depending on your operating system.

cp target/release/libwildbg.a $YOUR_C_PROJECT_FOLDER

3. Copy the library header

Replace $YOUR_C_PROJECT_FOLDER with the correct path):

cp crates/wildbg-c/wildbg.h $YOUR_C_PROJECT_FOLDER

4. Copy the neural nets

cp -r neural-nets $YOUR_C_PROJECT_FOLDER

5. Include the header file in your C file and use wildbg from there.

#include <stdio.h>
#include "wildbg.h"

int main() {
  // Initialize engine:
  Wildbg *wildbg = wildbg_new();
  
  // Define position and print winning probability:
  int starting[] = {0, -2, 0, 0, 0, 0, 5, 0, 3, 0, 0, 0, -5, 5, 0, 0, 0, -3, 0, -5, 0, 0, 0, 0, 2, 0,};
  CProbabilities p = probabilities(wildbg, &starting);
  printf("The estimated probability to win is %.2f percent.\n", 100 * p.win);
  
  // Specify 1 pointer game type:
  BgConfig config = { .x_away = 1, .o_away = 1 };
  
  // Find and print best move:
  CMove move = best_move(wildbg, &starting, 3, 1, &config);
  printf("The computer would make the following moves:\n")
  for (int i = 0; i < move.detail_count; i ++){
    printf("\tfrom %d to %d.\n", move.details[i].from, move.details[i].to);
  }
  
  // Deconstruct the engine and free the memory:
  wildbg_free(wildbg);  
}

6. Link the library into your binary

cd $YOUR_C_PROJECT_FOLDER
gcc main.c libwildbg.a