refactor: Refactored all of randomElement related functions and itera…

…tors to use ranges instead (cieslarmichal#843) * Added compiler_commands.json to be generated in CMake and removed unnecessary headers + range error * Refactored randomElement to use ranges and optimized algo further * Refactored sortedSizeRandomElement and randomElement's uses Everything now uses ranges instead of iterators, much cleaner and easier * Further replaced extra overloads and changed to range concept for sortedSizeRandomElement
00thirdeye00 · Aug 3, 2024 · 723e362 · 723e362
1 parent 99dd6ec
commit 723e362
Show file tree

Hide file tree

Showing 11 changed files with 104 additions and 179 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -13,6 +13,8 @@ option(BUILD_EXAMPLES "Build examples" OFF)
 option(BUILD_TESTING "Build tests" ON)
 option(CODE_COVERAGE "Build faker-cxx with coverage support" OFF)
 
+SET(CMAKE_EXPORT_COMPILE_COMMANDS ON)
+
 if (MSVC)
     set(CMAKE_REQUIRED_FLAGS /std:c++20)
 else()

diff --git a/include/faker-cxx/helper.h b/include/faker-cxx/helper.h
@@ -1,176 +1,81 @@
 #pragma once
 
-#include <initializer_list>
 #include <numeric>
-#include <span>
+#include <random>
 #include <vector>
-#include <iostream>
 
 #include "number.h"
 
 namespace faker::helper
 {
-/**
- * @brief Get a random element from an STL container.
- *
- * @tparam T an element type of the container.
- *
- * @param data The container.
- *
- * @return T a random element from the container.
- *
- * @code
- * faker::helper::randomElement<char>(std::string{"abcd"}) // "b"
- * faker::helper::randomElement<std::string>(std::vector<std::string>{{"hello"}, {"world"}}) // "hello"
- * @endcode
- */
-template <class T>
-T randomElement(std::span<const T> data)
-{
-    if (data.empty())
-    {
-        throw std::invalid_argument{"Data is empty."};
-    }
-
-    const auto index = number::integer<size_t>(data.size() - 1);
 
-    return data[index];
-}
+template <typename T>
+concept input_range_with_faster_size_compute_than_linear_rng =
+    std::ranges::input_range<T>            // must still be an input range no matter what, but additionally
+    && (std::ranges::sized_range<T>        // either knows its size in constant time, or
+        || std::ranges::forward_range<T>); // can multipass to compute the size
 
-template <typename T, std::size_t N>
-T randomElement(const std::array<T, N>& data)
+template <input_range_with_faster_size_compute_than_linear_rng Range>
+decltype(auto) randomElement(Range&& range)
 {
-    if (data.empty())
-    {
-        throw std::invalid_argument{"Data is empty."};
-    }
-
-    const auto index = number::integer<size_t>(data.size() - 1);
-
-    return data[index];
-}
-
-template <std::random_access_iterator It>
-auto randomElement(It start, It end)
-{
-    if (start == end)
+    if (std::ranges::empty(range))
     {
         throw std::invalid_argument{"Range [start, end) is empty."};
     }
 
-    auto size = end - start;
+    auto size = std::ranges::distance(range);
+
     const auto index = number::integer(size - 1);
 
-    return start[index];
+    return (*std::ranges::next(range.begin(), index));
 }
 
-
-template <std::forward_iterator It>
-auto& randomElement(It start, It end)
+template <std::ranges::input_range Range>
+auto randomElement(Range&& range)
 {
-    if (start == end)
+    auto const end = range.end();
+    auto itr = range.begin();
+
+    // Note: std::ranges::empty in general may need to grab begin/end
+    // then drop the iterator/sentinel, which can invalidate being, so
+    // it's not always usable with input_range's that aren't forward_range's
+    // we're going to "consume" the iterators ourselves so we can manually
+    // emptiness check by taking the iterator/sentinel pair and not dropping
+    // them
+    if (itr == end)
     {
         throw std::invalid_argument{"Range [start, end) is empty."};
     }
 
     std::random_device rd;
     std::mt19937 gen(rd());
 
-    std::reference_wrapper<typename std::iterator_traits<It>::value_type> result = *start;
-    ++start;
-    size_t count = 1;
-
-    while (start != end) {
-        std::uniform_int_distribution<size_t> distrib(0, count - 1);
-        if (distrib(gen) == 0) {
-            result = *start;
-        }
-        ++start;
-        ++count;
-    }
-
-    return result.get();
-}
-
-template <std::input_iterator It>
-auto randomElement(It start, It end)
-{
-    if (start == end)
+    using RangeValue = std::ranges::range_value_t<decltype(range)>;
+    auto consume_itr = [&itr]() -> decltype(auto)
+    {
+        using reference_type = std::ranges::range_reference_t<decltype(range)>;
+        if constexpr (std::is_reference_v<reference_type>)
+            return std::move(*itr);
+        else
+            return *itr;
+    };
+
+    RangeValue result = consume_itr();
+    ++itr;
+    std::size_t count = 1;
+
+    for (; itr != end; ++itr, ++count)
     {
-        throw std::invalid_argument{"Range [start, end) is empty."};
-    }
-
-    std::random_device rd;
-    std::mt19937 gen(rd());
-
-    auto result = std::move(*start);
-    ++start;
-    size_t count = 1;
-
-    while (start != end) {
         std::uniform_int_distribution<size_t> distrib(0, count);
-        if (distrib(gen) == 0) {
-            result = std::move(*start);
+        if (distrib(gen) == 0)
+        {
+            result = consume_itr();
         }
-        ++start;
-        ++count;
     }
 
     return result;
 }
 
-/**
- * @brief Get a random element from a vector.
- *
- * @tparam T an element type of the vector.
- *
- * @param data vector of elements.
- *
- * @return T a random element from the vector.
- *
- * @code
- * faker::helper::randomElement<std::string>(std::vector<std::string>{{"hello"}, {"world"}}) // "hello"
- * @endcode
- */
-template <class T>
-T randomElement(const std::vector<T>& data)
-{
-    if (data.empty())
-    {
-        throw std::invalid_argument{"Data is empty."};
-    }
-
-    const auto index = number::integer<size_t>(data.size() - 1);
-
-    return data[index];
-}
-
-/**
- * @brief Get a random element from an initializer list.
- *
- * @tparam T an element type of the initializer list.
- *
- * @param data initializer list of elements.
- *
- * @return T a random element from the initializer list.
- *
- * @code
- * faker::helper::randomElement<std::string>(std::initializer_list<std::string>{{"hello"}, {"world"}}) // "hello"
- * @endcode
- */
-template <class T>
-T randomElement(const std::initializer_list<T>& data)
-{
-    if (data.size() == 0)
-    {
-        throw std::invalid_argument{"Data is empty."};
-    }
-
-    const auto index = number::integer<size_t>(data.size() - 1);
-
-    return *(data.begin() + index);
-}
-
 /**
  * @brief Get a random element by weight from a vector.
  *

diff --git a/include/faker-cxx/word.h b/include/faker-cxx/word.h
@@ -144,22 +144,37 @@ FAKER_CXX_EXPORT std::string_view preposition(std::optional<unsigned> length = s
  */
 FAKER_CXX_EXPORT std::string_view verb(std::optional<unsigned> length = std::nullopt);
 
-template <std::input_iterator It>
-auto sortedSizeRandomElement(std::optional<unsigned int> length, It start, It end) ->
-    typename std::iterator_traits<It>::value_type
+/**
+ * @brief Returns random element of length
+ *
+ * @param length The length of the elements to be picked from
+ *
+ * @ range The range of elements
+ *
+ * @returns An element of the range value type
+ *
+ * @code
+ * faker::word::sortedSizeRandomElement(3, {"hi, "hello", "hey"}) // "hey" - Since "hey" is the only element of length 3
+ * @endcode
+ */
+template <std::ranges::range Range>
+auto sortedSizeRandomElement(std::optional<unsigned int> length, Range&& range) -> decltype(auto)
 {
     if (!length)
     {
-        return helper::randomElement(start, end);
+        return helper::randomElement(range);
     }
 
     size_t length_64 = *length;
+    auto start = range.begin();
+    auto end = range.end();
+
     auto lower_it = ::std::lower_bound(start, end, length_64,
                                        [](const auto& lhs, const auto& value) { return lhs.size() < value; });
 
     if (lower_it == end)
     {
-        return helper::randomElement(start, end);
+        return helper::randomElement(range);
     }
 
     if (lower_it->size() != length)
@@ -177,6 +192,6 @@ auto sortedSizeRandomElement(std::optional<unsigned int> length, It start, It en
         }
     }
 
-    return helper::randomElement(lower_it, upper_it);
+    return helper::randomElement(std::ranges::subrange(lower_it, upper_it));
 }
 }
diff --git a/src/common/algo_helper.h b/src/common/algo_helper.h
@@ -2,11 +2,11 @@
 
 #include <algorithm>
 #include <functional>
+#include <iterator>
 #include <random>
 #include <set>
 #include <stdexcept>
 #include <string>
-#include <iterator>
 
 #include "faker-cxx/datatype.h"
 #include "faker-cxx/export.h"
@@ -44,9 +44,10 @@ static T::key_type objectKey(const T& object)
     std::vector<typename T::key_type> keys;
     keys.reserve(object.size());
 
-    std::transform(object.begin(), object.end(), std::back_inserter(keys), [](const auto& entry) { return entry.first; });
+    std::transform(object.begin(), object.end(), std::back_inserter(keys),
+                   [](const auto& entry) { return entry.first; });
 
-    return randomElement<typename T::key_type>(keys);
+    return randomElement(keys);
 }
 
 template <typename TResult>

diff --git a/src/modules/airline.cpp b/src/modules/airline.cpp
@@ -33,7 +33,7 @@ Airport airport()
 std::string seat(AircraftType aircraftType)
 {
     return std::to_string(number::integer(1, aircraftTypeMaxRows.at(aircraftType))) +
-           helper::randomElement<char>(aircraftTypeSeatLetters.at(aircraftType));
+           helper::randomElement(aircraftTypeSeatLetters.at(aircraftType));
 }
 
 std::string recordLocator(bool allowNumerics)

diff --git a/src/modules/internet.cpp b/src/modules/internet.cpp
@@ -100,8 +100,7 @@ std::string username(std::optional<std::string> firstName, std::optional<std::st
         break;
     case 2:
         username = common::format("{}{}{}", lastNameLower,
-                                  helper::randomElement(std::vector<std::string>{".", "_", ""}),
-                                  firstNameLower);
+                                  helper::randomElement(std::vector<std::string>{".", "_", ""}), firstNameLower);
         break;
     }
 
@@ -149,7 +148,7 @@ std::string password(int length, const PasswordOptions& options)
 
     for (int i = 0; i < length; ++i)
     {
-        password += helper::randomElement<char>(characters);
+        password += helper::randomElement(characters);
     }
 
     return password;

diff --git a/src/modules/science.cpp b/src/modules/science.cpp
@@ -20,7 +20,7 @@ Unit unit()
     units.insert(units.end(), currentUnits.begin(), currentUnits.end());
     units.insert(units.end(), temperatureUnits.begin(), temperatureUnits.end());
 
-    return helper::randomElement<Unit>(units);
+    return helper::randomElement(units);
 }
 
 Unit distanceUnit()

diff --git a/src/modules/string.cpp b/src/modules/string.cpp
@@ -155,7 +155,7 @@ std::string fromCharacters(const std::string& characters, unsigned int length)
 
     for (unsigned i = 0; i < length; i++)
     {
-        result += helper::randomElement<char>(characters);
+        result += helper::randomElement(characters);
     }
 
     return result;
@@ -196,7 +196,7 @@ std::string alpha(unsigned length, StringCasing casing, const std::string& exclu
 
     for (unsigned i = 0; i < length; i++)
     {
-        alpha += helper::randomElement<char>(targetCharacters);
+        alpha += helper::randomElement(targetCharacters);
     }
 
     return alpha;
@@ -232,7 +232,7 @@ std::string alphanumeric(unsigned int length, StringCasing casing, const std::st
 
     for (unsigned i = 0; i < length; i++)
     {
-        alphanumeric += helper::randomElement<char>(targetCharacters);
+        alphanumeric += helper::randomElement(targetCharacters);
     }
 
     return alphanumeric;
@@ -263,11 +263,11 @@ std::string numeric(unsigned int length, bool allowLeadingZeros)
     {
         if (i == 0 && allowLeadingZeros)
         {
-            alphanumericStr.push_back(helper::randomElement<char>(numericCharacters));
+            alphanumericStr.push_back(helper::randomElement(numericCharacters));
         }
         else
         {
-            alphanumericStr.push_back(helper::randomElement<char>(numericCharactersWithoutZero));
+            alphanumericStr.push_back(helper::randomElement(numericCharactersWithoutZero));
         }
     }
 
@@ -321,7 +321,7 @@ std::string hexadecimal(unsigned int length, HexCasing casing, HexPrefix prefix)
 
     for (unsigned i = 0; i < length; i++)
     {
-        hexadecimal += helper::randomElement<char>(hexadecimalCharacters);
+        hexadecimal += helper::randomElement(hexadecimalCharacters);
     }
 
     return hexadecimal;

diff --git a/src/modules/system.cpp b/src/modules/system.cpp
@@ -267,6 +267,6 @@ std::string cron(const CronOptions& options)
                                                        "@reboot",   "@weekly", "@yearly"};
 
     return (!includeNonStandard || datatype::boolean(0)) ? standardExpression :
-                                                           helper::randomElement<std::string>(nonStandardExpressions);
+                                                           helper::randomElement(nonStandardExpressions);
 }
 }