Add split_n & split_once; some tests

src/include/split.hpp

@@ -22,17 +22,42 @@
 #include <vector>
 
 namespace jessilib {
+namespace impl {
+
+template<typename FirstArgT, typename...>
+struct first_arg {
+	using first_type = FirstArgT;
+};
+
+// Using a function's return type because both `using` and classes require parameter packs be at the end
+template<template<typename...> typename ContainerT, typename DefaultMemberT, typename... ArgsT,
+	typename std::enable_if<sizeof...(ArgsT) == 0>::type* = nullptr>
+constexpr auto split_container_helper_f() -> ContainerT<DefaultMemberT> {
+	return {};
+};
+
+template<template<typename...> typename ContainerT, typename DefaultMemberT, typename... ArgsT,
+	typename std::enable_if<sizeof...(ArgsT) != 0>::type* = nullptr>
+constexpr auto split_container_helper_f() -> ContainerT<ArgsT...> {
+	return {};
+};
+
+template<template<typename...> typename ContainerT, typename... ArgsT>
+using split_container_helper_t = decltype(split_container_helper_f<ContainerT, ArgsT...>());
+
+} // namespace impl
 
 /** Splits an input string into substrings */
-template<typename InStringT, typename ResultT = std::vector<InStringT>>
-constexpr ResultT split(const InStringT& in_string, typename InStringT::value_type in_delim) {
-	ResultT result;
+template<template<typename...> typename ContainerT = std::vector, typename... ContainerArgsT, typename InputT>
+constexpr auto split(const InputT& in_string, typename InputT::value_type in_delim) {
+	using MemberT = typename impl::first_arg<ContainerArgsT..., std::basic_string<typename InputT::value_type>>::first_type;
+	using container_type = impl::split_container_helper_t<ContainerT, MemberT, ContainerArgsT...>;
+	container_type result;
 	// Nothing to return
 	if (in_string.empty()) {
 		return result;
 	}
 
-	// If only MSVC didn't suck, we could use begin() and end()...
 	auto begin = in_string.data();
 	auto end = in_string.data() + in_string.size();
 	for (auto itr = begin; itr != end; ++itr) {
@@ -43,21 +68,100 @@ constexpr ResultT split(const InStringT& in_string, typename InStringT::value_ty
 		}
 	}
 
-	// begin == end only if last character was in_delim
-	if (begin == end) {
-		result.emplace_back();
+	// Push final token to the end; may be empty
+	result.emplace_back(begin, end - begin);
+
+	return result;
+}
+
+/**
+ * Splits an input string into 2 substrings at and omitting an input delimiter. Returns:
+ * An empty pair if in_string is empty,
+ * otherwise if the delimiter is not present, a pair who's `second` member is empty and `first` member is equal to `in_string`,
+ * otherwise, a pair split at first instance of the delimiter
+ * Complexity: O(in_string.size())
+ *
+ * @param in_string string to split
+ * @param in_delim delimiter to split on
+ * @return A pair representing `in_string` split at a delimiter, with first half stored in `first` and second in `last`
+ */
+template<typename InStringT, typename ResultMemberT = InStringT>
+constexpr std::pair<ResultMemberT, ResultMemberT> split_once(const InStringT& in_string, typename InStringT::value_type in_delim) {
+	// Nothing in, nothing out
+	std::pair<ResultMemberT, ResultMemberT> result;
+	if (in_string.empty()) {
+		return result;
 	}
-	else {
-		result.emplace_back(begin, end - begin);
+
+	auto begin = in_string.data();
+	auto end = in_string.data() + in_string.size();
+	for (auto itr = begin; itr != end; ++itr) {
+		if (*itr == in_delim) {
+			// in_delim found; split upon it
+			result.first = ResultMemberT{ begin, itr };
+			result.second = ResultMemberT{ itr + 1, end };
+			return result;
+		}
 	}
 
+	// in_delim not found
+	result.first = in_string;
 	return result;
 }
 
+/**
+ * Splits a string up to a specified number of times and returns the result
+ *
+ * @tparam ContainerT Container type to store the results in
+ * @tparam ContainerArgsT Optional template parameters for ContainerT
+ * @param in_string String to split
+ * @param in_delim Delimiter to split upon
+ * @param in_limit Maximum number of times to split
+ * @return ResultT containing to up `in_limit` + 1 substrings
+ */
+template<template<typename...> typename ContainerT = std::vector, typename... ContainerArgsT, typename InputT>
+constexpr auto split_n(const InputT& in_string, typename InputT::value_type in_delim, size_t in_limit) {
+	using MemberT = typename impl::first_arg<ContainerArgsT..., std::basic_string<typename InputT::value_type>>::first_type;
+	using container_type = impl::split_container_helper_t<ContainerT, MemberT, ContainerArgsT...>;
+
+	container_type result;
+	if (in_string.empty()) {
+		return result;
+	}
+
+	auto begin = in_string.data();
+	auto end = in_string.data() + in_string.size();
+	for (auto itr = begin; itr != end && in_limit != 0; ++itr) {
+		if (*itr == in_delim) {
+			// Push token to result
+			result.emplace_back(begin, itr - begin);
+			begin = itr + 1;
+			--in_limit;
+		}
+	}
+
+	// Push final token to the end; may be empty
+	result.emplace_back(begin, end - begin);
+
+	return result;
+}
+
+/** Splits an input string into view substrings; cannot specify element return type */
+template<template<typename...> typename ContainerT = std::vector, typename... ContainerArgsT, typename InStringT, typename MemberT = std::basic_string_view<typename InStringT::value_type>>
+constexpr ContainerT<MemberT, ContainerArgsT...> split_view(const InStringT& in_string, typename InStringT::value_type in_delim) {
+	return split<ContainerT, MemberT, ContainerArgsT...>(in_string, in_delim);
+}
+
 /** Splits an input string into view substrings (same as split, but different default return type) */
-template<typename InStringT, typename ResultT = std::vector<std::basic_string_view<typename InStringT::value_type>>>
-constexpr ResultT split_view(const InStringT& in_string, typename InStringT::value_type in_delim) {
-	return split<InStringT, ResultT>(in_string, in_delim);
+template<typename InStringT, typename ResultT = std::basic_string_view<typename InStringT::value_type>>
+constexpr std::pair<ResultT, ResultT> split_once_view(const InStringT& in_string, typename InStringT::value_type in_delim) {
+	return split_once<InStringT, ResultT>(in_string, in_delim);
+}
+
+/** Splits an input string into a specified number of view substrings */
+template<template<typename...> typename ContainerT = std::vector, typename... ContainerArgsT, typename InStringT, typename MemberT = std::basic_string_view<typename InStringT::value_type>>
+constexpr auto split_n_view(const InStringT& in_string, typename InStringT::value_type in_delim, size_t in_limit) {
+	return split_n<ContainerT, MemberT, ContainerArgsT...>(in_string, in_delim, in_limit);
 }
 
 } // namespace jessilib

src/test/split.cpp

@@ -20,6 +20,7 @@
 #include <cassert>
 #include <deque>
 #include <list>
+#include <random>
 #include "test.hpp"
 
 using namespace jessilib;
@@ -54,12 +55,24 @@ public:
 };
 TYPED_TEST_SUITE(SplitViewStringTest, char_types);
 
+template<typename T>
+class SplitOnceTest : public ::testing::Test {
+public:
+};
+TYPED_TEST_SUITE(SplitOnceTest, char_types);
+
+template<typename T>
+class SplitNTest : public ::testing::Test {
+public:
+};
+TYPED_TEST_SUITE(SplitNTest, char_types);
+
 template<typename T>
 std::basic_string<T> make_word(size_t length = 8, T delim = static_cast<T>(0)) {
 	std::basic_string<T> result;
 
 	if (length == 0) {
-		throw std::runtime_error{ "length == 0" };
+		return {};
 	}
 
 	result.push_back(delim + 1);
@@ -86,15 +99,58 @@ std::basic_string_view<T> make_word_view(size_t length = 8, T delim = static_cas
 	return s_result;
 }
 
+template<typename T>
+struct RandomTestData {
+	RandomTestData(T in_delim)
+		: m_delim{ in_delim } {
+		operator()();
+	}
+
+	void operator()() {
+		m_tokens.clear();
+		m_str.clear();
+
+		std::mt19937 randgen(static_cast<unsigned int>(std::chrono::system_clock::now().time_since_epoch().count()));
+		std::uniform_int_distribution<uint32_t> word_count_distribution(5, 64);
+		std::uniform_int_distribution<uint32_t> word_length_distribution(0, 16);
+
+		auto random_words = word_count_distribution(randgen);
+		while (m_tokens.size() < random_words) {
+			m_tokens.push_back(make_word<T>(word_length_distribution(randgen)));
+			m_str += m_tokens.back();
+			if (m_tokens.size() < random_words) {
+				m_str += m_delim;
+			}
+		}
+	}
+
+	std::basic_string<T> get_remainder(size_t in_times_split) {
+		std::basic_string<T> result;
+		while (in_times_split < m_tokens.size()) {
+			result += m_tokens[in_times_split];
+			++in_times_split;
+			if (in_times_split < m_tokens.size()) {
+				result += m_delim;
+			}
+		}
+
+		return result;
+	}
+
+	T m_delim;
+	std::basic_string<T> m_str;
+	std::vector<std::basic_string<T>> m_tokens;
+};
+
 TYPED_TEST(SplitSVTest, empty) {
 	std::basic_string_view<TypeParam> empty;
-	std::vector<decltype(empty)> split_result = split(empty, static_cast<TypeParam>(0));
+	std::vector<std::basic_string<TypeParam>> split_result = split(empty, static_cast<TypeParam>(0));
 	EXPECT_TRUE(split_result.empty());
 }
 
 TYPED_TEST(SplitSVTest, single_word) {
 	std::basic_string_view<TypeParam> single_word = make_word_view<TypeParam>();
-	std::vector<decltype(single_word)> split_result = split(single_word, static_cast<TypeParam>(0));
+	std::vector<std::basic_string<TypeParam>> split_result = split(single_word, static_cast<TypeParam>(0));
 	EXPECT_EQ(split_result.size(), 1);
 	EXPECT_EQ(split_result[0].size(), 8);
 }
@@ -103,7 +159,7 @@ TYPED_TEST(SplitSVTest, single_word_trailing_delim) {
 	auto word = make_word<TypeParam>();
 	word += static_cast<TypeParam>(0);
 	std::basic_string_view<TypeParam> single_word = word;
-	std::vector<decltype(single_word)> split_result = split(single_word, static_cast<TypeParam>(0));
+	std::vector<std::basic_string<TypeParam>> split_result = split(single_word, static_cast<TypeParam>(0));
 	EXPECT_EQ(split_result.size(), 2);
 	EXPECT_EQ(split_result[0].size(), 8);
 	EXPECT_EQ(split_result[1].size(), 0);
@@ -114,7 +170,7 @@ TYPED_TEST(SplitSVTest, single_word_prefix_delim) {
 	word += static_cast<TypeParam>(0);
 	word += make_word<TypeParam>();
 	std::basic_string_view<TypeParam> single_word = word;
-	std::vector<decltype(single_word)> split_result = split(single_word, static_cast<TypeParam>(0));
+	std::vector<std::basic_string<TypeParam>> split_result = split(single_word, static_cast<TypeParam>(0));
 	EXPECT_EQ(split_result.size(), 2);
 	EXPECT_EQ(split_result[0].size(), 0);
 	EXPECT_EQ(split_result[1].size(), 8);
@@ -126,7 +182,7 @@ TYPED_TEST(SplitSVTest, single_word_surround_delim) {
 	word += make_word<TypeParam>();
 	word += static_cast<TypeParam>(0);
 	std::basic_string_view<TypeParam> single_word = word;
-	std::vector<decltype(single_word)> split_result = split(single_word, static_cast<TypeParam>(0));
+	std::vector<std::basic_string<TypeParam>> split_result = split(single_word, static_cast<TypeParam>(0));
 	EXPECT_EQ(split_result.size(), 3);
 	EXPECT_EQ(split_result[0].size(), 0);
 	EXPECT_EQ(split_result[1].size(), 8);
@@ -138,7 +194,7 @@ TYPED_TEST(SplitSVTest, two_words) {
 	word += static_cast<TypeParam>(0);
 	word += make_word<TypeParam>();
 	std::basic_string_view<TypeParam> words = word;
-	std::vector<decltype(words)> split_result = split(words, static_cast<TypeParam>(0));
+	std::vector<std::basic_string<TypeParam>> split_result = split(words, static_cast<TypeParam>(0));
 	EXPECT_EQ(split_result.size(), 2);
 	EXPECT_EQ(split_result[0].size(), 8);
 	EXPECT_EQ(split_result[1].size(), 8);
@@ -151,14 +207,62 @@ TYPED_TEST(SplitSVTest, three_words) {
 	word += static_cast<TypeParam>(0);
 	word += make_word<TypeParam>(9);
 	std::basic_string_view<TypeParam> words = word;
-	std::vector<decltype(words)> split_result = split(words, static_cast<TypeParam>(0));
+	std::vector<std::basic_string<TypeParam>> split_result = split(words, static_cast<TypeParam>(0));
 	EXPECT_EQ(split_result.size(), 3);
 	EXPECT_EQ(split_result[0].size(), 3);
 	EXPECT_EQ(split_result[1].size(), 5);
 	EXPECT_EQ(split_result[2].size(), 9);
 }
 
-/** std::string split test, really just testing compilation */
+/** SplitOnceTest */
+
+TYPED_TEST(SplitOnceTest, random) {
+	RandomTestData<TypeParam> data{ static_cast<TypeParam>(0) };
+	std::pair<std::basic_string<TypeParam>, std::basic_string<TypeParam>> split_result = split_once(data.m_str, data.m_delim);
+
+	EXPECT_EQ(split_result.first, data.m_tokens[0]);
+	EXPECT_EQ(split_result.second, data.get_remainder(1));
+}
+
+TYPED_TEST(SplitOnceTest, random_view) {
+	RandomTestData<TypeParam> data{ static_cast<TypeParam>(0) };
+	std::pair<std::basic_string_view<TypeParam>, std::basic_string_view<TypeParam>> split_result = split_once_view(data.m_str, data.m_delim);
+
+	EXPECT_EQ(split_result.first, data.m_tokens[0]);
+	EXPECT_EQ(split_result.second, data.get_remainder(1));
+}
+
+/** SplitNTest */
+
+TYPED_TEST(SplitNTest, random) {
+	RandomTestData<TypeParam> data{ static_cast<TypeParam>(0) };
+	constexpr size_t n = 4;
+	std::vector<std::basic_string<TypeParam>> split_result = split_n(data.m_str, data.m_delim, n);
+
+	// Tokens shall be same up until last one (n + 1)
+	EXPECT_EQ(split_result.size(), n + 1);
+	for (size_t index = 0; index != n; ++index) {
+		EXPECT_EQ(split_result[index], data.m_tokens[index]);
+	}
+
+	EXPECT_EQ(split_result[n], data.get_remainder(n));
+}
+
+TYPED_TEST(SplitNTest, random_view) {
+	RandomTestData<TypeParam> data{ static_cast<TypeParam>(0) };
+	constexpr size_t n = 4;
+	std::vector<std::basic_string_view<TypeParam>> split_result = split_n_view(data.m_str, data.m_delim, n);
+
+	// Tokens shall be same up until last one (n + 1)
+	EXPECT_EQ(split_result.size(), n + 1);
+	for (size_t index = 0; index != n; ++index) {
+		EXPECT_EQ(split_result[index], data.m_tokens[index]);
+	}
+
+	EXPECT_EQ(split_result[n], data.get_remainder(n));
+}
+
+/** std::string split test, really just testing compilation and returned types */
 
 TYPED_TEST(SplitSringTest, empty) {
 	std::basic_string<TypeParam> empty;
@@ -173,17 +277,23 @@ TYPED_TEST(SplitSringTest, single_word) {
 	EXPECT_EQ(split_result[0].size(), 8);
 }
 
+TYPED_TEST(SplitSringTest, random) {
+	RandomTestData<TypeParam> data{ static_cast<TypeParam>(0) };
+	auto split_result = split(data.m_str, data.m_delim);
+	EXPECT_EQ(split_result, data.m_tokens);
+}
+
 /** Some basic tests for compiling with different containers */
 
 TYPED_TEST(SplitSVTest, empty_deque) {
 	std::basic_string_view<TypeParam> empty;
-	std::deque<decltype(empty)> split_result = split<decltype(empty), std::deque<decltype(empty)>>(empty, static_cast<TypeParam>(0));
+	std::deque<std::basic_string<TypeParam>> split_result = split<std::deque>(empty, static_cast<TypeParam>(0));
 	EXPECT_TRUE(split_result.empty());
 }
 
 TYPED_TEST(SplitSVTest, empty_list) {
 	std::basic_string_view<TypeParam> empty;
-	std::list<decltype(empty)> split_result = split<decltype(empty), std::list<decltype(empty)>>(empty, static_cast<TypeParam>(0));
+	std::list<std::basic_string<TypeParam>> split_result = split<std::list>(empty, static_cast<TypeParam>(0));
 	EXPECT_TRUE(split_result.empty());
 }
 
@@ -212,3 +322,9 @@ TYPED_TEST(SplitViewStringTest, single_word) {
 	std::vector<std::basic_string_view<TypeParam>> split_result = split_view(single_word, static_cast<TypeParam>(0));
 	EXPECT_EQ(split_result.size(), 1);
 }
+
+TYPED_TEST(SplitViewStringTest, random) {
+	RandomTestData<TypeParam> data{ static_cast<TypeParam>(0) };
+	auto split_result = split_view(data.m_str, data.m_delim);
+	EXPECT_EQ(split_result.size(), data.m_tokens.size());
+}