Test utilities for libscran

This repository contains utilities for testing other libscran libraries. To include this in a project, just add the following chunk to the test-specific CMakeLists.txt:

include(FetchContent)
 
FetchContent_Declare(
  scran_tests
  GIT_REPOSITORY https://github.com/libscran/scran_tests
  GIT_TAG master # or any version of interest
)
 
FetchContent_MakeAvailable(scran_tests)
 
target_link_libraries(mylib INTERFACE scran_tests)

This will automatically pull in GoogleTest via FetchContent so downstream projects don't have to do it themselves. Then, to use the library, we just have to add the header:

#include "scran_tests/scran_tests.hpp"

We can now simulate some random data easily:

auto res = scran_tests::simulate_vector(100, scran_tests::SimulationParameters());
 
auto sparse_res = scran_tests::simulate_vector(100, []{
    scran_tests::SimulationParameters params
    params.density = 0.1;
    return params;
}());
 
auto int_res = scran_tests::simulate_vector(100, []{
    scran_tests::SimulationParameters<uint16_t> params
    params.density = 0.1;
    return params;
}());

Comparison of almost-equal floating-point numbers, given a relative tolerance:

scran_tests::compare_almost_equal(1.0, 1.0000000001);
 
std::vector<double> v1{1.0, 2.000000001};
std::vector<double> v2{1.000000001, 2.0};
scran_tests::compare_almost_equal(v1, v2);

Quick construction of vectors for use in EXPECT_EQ():

std::vector<double> big_array(1000);
auto subset = scran_tests::vector_n(big_array.data() + 100, 500); // slice from [100, 600)

Expect specific error messages:

scran_tests::expect_error([&]() { throw std::runtime_error("foobar"); }, "foo");

Test against zero-initialized assumptions:

std::vector<double> tmp(10, scran_tests::initial_value());
std::vector<double> tmp2(10, scran_tests::initial_value());

Check out the documentation for more details.