... is a thin, header-only C++ wrapper for matio..

• modern C++ semantics implemented: all the allocation and freeing of memory happens via constructor and smart pointers
• uses boost.regex to filter for channels/columns/vectors/whatever you want to call them
• uses exceptions to gracefully handle things
• is really drop-in and use; makes you feel like C++ is Python (except thanks to matio you can use one solution for all versions of mat files)

Have a look at the tests, ideally, but it's all meant to work as simple as:

//
// create bmc::mat object
//
bmc::mat my_data{filename /* std::string */, channel_regex /* std::string */};
auto no_of_channels = my_data.number_of_channels();
auto temp_sensor = my_data.at("temp_sensor");  // or my_data["temp_sensor"]
for (const auto & it : my_data) {
  // do something with the channel
}
if (my_data.find("pressure") != my_data.end()) {
  std::cout << "pressure column found" << '\n';
}
auto size_of_columns = my_data.size();
if (my_data.empty()) {
  std::cout << "no data found in file" << '\n';
}

The regex string is optional; you can leave it empty to just load everything.

Data is returned as std::vector<float>.

There are two ways to do this. Option A is: clone the repo, copy the header to your project and include it. That's probably the easiest thing to to if you're not using CMake.

If you are using CMake, I recommend you clone the repo to your thirdparty directory or whereever you want it to live and add the following to your CMakeList files:

add_subdirectory (${YourProject_SOURCE_DIR}/path/to/matiopp)
add_exectuable (your_executable your_sources.cc)          # or library
target_link_libraries (your_executable matiopp::matiopp)  # or PRIVATE etc

That's it. Have fun.