base64_module.cpp

// base64_module.cpp
// C++ extension (built with pybind11) exposing a fast file -> base64 encoder
// to Python. Build with setup.py (see README).

#include <pybind11/pybind11.h>
#include <fstream>
#include <vector>
#include <string>
#include <stdexcept>

namespace py = pybind11;

static const char* kB64Chars =
    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
    "abcdefghijklmnopqrstuvwxyz"
    "0123456789+/";

static std::string base64_encode(const std::vector<unsigned char>& data) {
    std::string encoded;
    encoded.reserve(((data.size() + 2) / 3) * 4);

    size_t i = 0;
    while (i + 2 < data.size()) {
        unsigned int n = (data[i] << 16) | (data[i + 1] << 8) | data[i + 2];
        encoded += kB64Chars[(n >> 18) & 0x3F];
        encoded += kB64Chars[(n >> 12) & 0x3F];
        encoded += kB64Chars[(n >> 6) & 0x3F];
        encoded += kB64Chars[n & 0x3F];
        i += 3;
    }

    size_t remaining = data.size() - i;
    if (remaining == 1) {
        unsigned int n = data[i] << 16;
        encoded += kB64Chars[(n >> 18) & 0x3F];
        encoded += kB64Chars[(n >> 12) & 0x3F];
        encoded += "==";
    } else if (remaining == 2) {
        unsigned int n = (data[i] << 16) | (data[i + 1] << 8);
        encoded += kB64Chars[(n >> 18) & 0x3F];
        encoded += kB64Chars[(n >> 12) & 0x3F];
        encoded += kB64Chars[(n >> 6) & 0x3F];
        encoded += "=";
    }

    return encoded;
}

// Reads a file from disk and returns its contents as a base64 string.
// Equivalent to the Python read_b64() helper, just implemented in C++.
std::string read_file_b64(const std::string& path) {
    std::ifstream f(path, std::ios::binary);
    if (!f) {
        throw std::runtime_error("Cannot open file: " + path);
    }
    std::vector<unsigned char> buffer(
        (std::istreambuf_iterator<char>(f)),
        std::istreambuf_iterator<char>());
    return base64_encode(buffer);
}

PYBIND11_MODULE(fastb64, m) {
    m.doc() = "Fast base64 file encoder implemented in C++, used from Python";
    m.def("read_file_b64", &read_file_b64,
          "Read a file from disk and return its base64-encoded contents",
          py::arg("path"));
}