Merge branch 'rapidfuzz:main' into bvandercar/reduce-len-calls

rapidfuzz · Nov 14, 2024 · 0323bc8 · 0323bc8
2 parents ba243a8 + 268a382
commit 0323bc8
Show file tree

Hide file tree

Showing 10 changed files with 31 additions and 21 deletions.
diff --git a/.github/workflows/branchbuild.yml b/.github/workflows/branchbuild.yml
@@ -60,7 +60,7 @@ jobs:
         run: |
           git clone https://github.com/rapidfuzz/rapidfuzz-cpp.git
           cd rapidfuzz-cpp
-          git checkout v3.0.5
+          git checkout v3.1.1
           mkdir build && cd build
           cmake .. -DCMAKE_BUILD_TYPE=Release
           cmake --build .
@@ -104,7 +104,7 @@ jobs:
         run: |
           git clone https://github.com/rapidfuzz/rapidfuzz-cpp.git
           cd rapidfuzz-cpp
-          git checkout v3.0.5
+          git checkout v3.1.1
           mkdir build && cd build
           cmake .. -DCMAKE_BUILD_TYPE=Release
           cmake --build .

diff --git a/CHANGELOG.rst b/CHANGELOG.rst
@@ -1,6 +1,17 @@
 Changelog
 ---------
 
+[3.10.1] - 2024-10-24
+^^^^^^^^^^^^^^^^^^^^^
+Fixed
+~~~~~
+- fix compilation on clang-19
+- fix incorrect results in simd optimized implementation of Levenshtein and OSA on 32bit targets
+
+Added
+~~~~~
+* added support for taskflow 3.8.0
+
 [3.10.0] - 2024-09-21
 ^^^^^^^^^^^^^^^^^^^^^
 Fixed

diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -77,7 +77,10 @@ check_cpu_arch_x86(RAPIDFUZZ_ARCH_X86)
 
 set(RF_BASE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/src)
 
-find_package(Taskflow 3.7.0 QUIET)
+find_package(Taskflow 3.8.0 QUIET)
+if(NOT Taskflow_FOUND)
+  find_package(Taskflow 3.7.0 QUIET)
+endif()
 if(NOT Taskflow_FOUND)
   find_package(Taskflow 3.6.0 QUIET)
 endif()
@@ -107,7 +110,7 @@ else()
   add_library(Taskflow::Taskflow ALIAS Taskflow)
 endif()
 
-find_package(rapidfuzz 3.0.5 QUIET)
+find_package(rapidfuzz 3.1.1 QUIET)
 if(rapidfuzz_FOUND)
   message(STATUS "Using system supplied version of rapidfuzz-cpp")
 else()

diff --git a/extern/rapidfuzz-cpp b/extern/rapidfuzz-cpp
diff --git a/extern/taskflow b/extern/taskflow
diff --git a/src/rapidfuzz/__init__.py b/src/rapidfuzz/__init__.py
@@ -6,7 +6,7 @@
 
 __author__: str = "Max Bachmann"
 __license__: str = "MIT"
-__version__: str = "3.10.0"
+__version__: str = "3.10.1"
 
 from rapidfuzz import distance, fuzz, process, utils
 

diff --git a/src/rapidfuzz/cpp_common.hpp b/src/rapidfuzz/cpp_common.hpp
@@ -708,15 +708,15 @@ static inline bool multi_normalized_distance_init(RF_ScorerFunc* self, int64_t s
 static inline PyObject* opcodes_apply(const rf::Opcodes& ops, const RF_String& str1, const RF_String& str2)
 {
     return visitor(str1, str2, [&](auto s1, auto s2) {
-        auto proc_str = rf::opcodes_apply<uint32_t>(ops, s1, s2);
+        auto proc_str = rf::opcodes_apply_vec<uint32_t>(ops, s1, s2);
         return PyUnicode_FromKindAndData(PyUnicode_4BYTE_KIND, proc_str.data(), (Py_ssize_t)proc_str.size());
     });
 }
 
 static inline PyObject* editops_apply(const rf::Editops& ops, const RF_String& str1, const RF_String& str2)
 {
     return visitor(str1, str2, [&](auto s1, auto s2) {
-        auto proc_str = rf::editops_apply<uint32_t>(ops, s1, s2);
+        auto proc_str = rf::editops_apply_vec<uint32_t>(ops, s1, s2);
         return PyUnicode_FromKindAndData(PyUnicode_4BYTE_KIND, proc_str.data(), (Py_ssize_t)proc_str.size());
     });
 }
diff --git a/src/rapidfuzz/process_cpp.hpp b/src/rapidfuzz/process_cpp.hpp
@@ -1,8 +1,8 @@
 #pragma once
 #include "cpp_common.hpp"
 #include "rapidfuzz.h"
-#include "taskflow/algorithm/for_each.hpp"
 #include "taskflow/taskflow.hpp"
+#include "taskflow/algorithm/for_each.hpp"
 #include <atomic>
 #include <exception>
 #include <stdexcept>

diff --git a/src/rapidfuzz/utils.hpp b/src/rapidfuzz/utils.hpp
@@ -10,7 +10,7 @@
 #include <cwctype>
 #include <limits>
 #include <stdint.h>
-#include <string>
+#include <vector>
 
 uint32_t UnicodeDefaultProcess(uint32_t ch);
 
@@ -80,11 +80,10 @@ int64_t default_process(CharT* str, int64_t len)
 }
 
 template <typename CharT>
-std::basic_string<CharT> default_process(std::basic_string<CharT> s)
+std::vector<CharT> default_process_copy(CharT* str_, int64_t len_)
 {
-    std::basic_string<CharT> str(s);
-
-    int64_t len = default_process(&str[0], str.size());
+    std::vector<CharT> str(str_, str_ + len_);
+    int64_t len = default_process(str.data(), str.size());
     str.resize(len);
     return str;
 }
diff --git a/src/rapidfuzz/utils_cpp.hpp b/src/rapidfuzz/utils_cpp.hpp
@@ -9,20 +9,17 @@ static inline PyObject* default_process_impl(PyObject* sentence)
     switch (c_sentence.kind) {
     case RF_UINT8:
     {
-        auto proc_str = default_process(
-            std::basic_string<uint8_t>(static_cast<uint8_t*>(c_sentence.data), c_sentence.length));
+        auto proc_str = default_process_copy(static_cast<uint8_t*>(c_sentence.data), c_sentence.length);
         return PyUnicode_FromKindAndData(PyUnicode_1BYTE_KIND, proc_str.data(), (Py_ssize_t)proc_str.size());
     }
     case RF_UINT16:
     {
-        auto proc_str = default_process(
-            std::basic_string<uint16_t>(static_cast<uint16_t*>(c_sentence.data), c_sentence.length));
+        auto proc_str = default_process_copy(static_cast<uint16_t*>(c_sentence.data), c_sentence.length);
         return PyUnicode_FromKindAndData(PyUnicode_2BYTE_KIND, proc_str.data(), (Py_ssize_t)proc_str.size());
     }
     case RF_UINT32:
     {
-        auto proc_str = default_process(
-            std::basic_string<uint32_t>(static_cast<uint32_t*>(c_sentence.data), c_sentence.length));
+        auto proc_str = default_process_copy(static_cast<uint32_t*>(c_sentence.data), c_sentence.length);
         return PyUnicode_FromKindAndData(PyUnicode_4BYTE_KIND, proc_str.data(), (Py_ssize_t)proc_str.size());
     }
     // ToDo: for now do not process these elements should be done in some way in the future