bad attempt at language huristics

.idea/vcs.xml

@@ -3,5 +3,6 @@
   <component name="VcsDirectoryMappings">
     <mapping directory="$PROJECT_DIR$" vcs="Git" />
     <mapping directory="$PROJECT_DIR$/lib/blt" vcs="Git" />
+    <mapping directory="$PROJECT_DIR$/lib/blt/libraries/parallel-hashmap" vcs="Git" />
   </component>
 </project>

CMakeLists.txt

@@ -1,5 +1,5 @@
 cmake_minimum_required(VERSION 3.25)
-project(floc VERSION 0.0.8)
+project(floc VERSION 0.0.9)
 
 include(FetchContent)
 
@@ -19,6 +19,10 @@ FetchContent_Declare(ftxui
         GIT_REPOSITORY https://github.com/ArthurSonzogni/ftxui
         GIT_TAG v5.0.0
 )
+FetchContent_Declare(json
+        GIT_REPOSITORY https://github.com/nlohmann/json.git
+        GIT_TAG v3.11.3
+)
 
 FetchContent_GetProperties(ftxui)
 if(NOT ftxui_POPULATED)
@@ -26,11 +30,12 @@ if(NOT ftxui_POPULATED)
     add_subdirectory(${ftxui_SOURCE_DIR} ${ftxui_BINARY_DIR} EXCLUDE_FROM_ALL)
 endif()
 
-target_link_libraries(floc ${BLT_TARGET})
+target_link_libraries(floc PRIVATE BLT)
 target_link_libraries(floc
         PRIVATE ftxui::screen
         PRIVATE ftxui::dom
         PRIVATE ftxui::component
+        PRIVATE json
 )
 
 target_compile_options(floc PRIVATE -Wall -Werror -Wpedantic -Wno-comment)

include/blt/floc/analyzer.h

@@ -19,6 +19,13 @@
 #ifndef BLT_FLOC_ANALYZER_H
 #define BLT_FLOC_ANALYZER_H
 
+#include <string>
+#include <string_view>
+#include <functional>
+#include <vector>
+#include <blt/std/hashmap.h>
+#include <blt/std/types.h>
+
 /*
  * This file provides an analyzer class for counting lines of code in a file along with a recursive parser for finding files
  * it also provides a utility class to build a structure which mimics the file system directories alongside relevant code
@@ -26,4 +33,72 @@
  * maybe split this into a few header files?
  */
 
+namespace blt::floc
+{
+    static inline const constexpr auto EMPTY_FUNCTION = [](std::string_view contents) {
+        return true;
+    };
+    
+    struct associated_file_t
+    {
+        std::string file_extension;
+        std::string file_name;
+        
+        associated_file_t() = default;
+        
+        associated_file_t(std::string_view file_extension, std::string_view file_name): file_extension(file_extension), file_name(file_name)
+        {}
+    };
+    
+    class language_t
+    {
+        private:
+            // extension -> file name
+            blt::hashmap_t<std::string, std::string> language_files;
+            std::function<bool(std::string_view)> matches_func;
+        public:
+            language_t(): matches_func(EMPTY_FUNCTION)
+            {}
+            
+            explicit language_t(std::function<bool(std::string_view)> matches_func): matches_func(std::move(matches_func))
+            {}
+            
+            inline bool matches(std::string_view file_extension, std::string_view file_contents)
+            {
+                return language_files.find(std::string(file_extension)) != language_files.end() && matches_func(file_contents);
+            }
+            
+            inline language_t& add(const associated_file_t& file)
+            {
+                language_files[file.file_extension] = file.file_name;
+                return *this;
+            }
+            
+            inline language_t& add(std::string_view extension, std::string_view name)
+            {
+                return add(associated_file_t{extension, name});
+            }
+            
+            inline language_t& add(const std::vector<std::string>& extensions, std::string_view name)
+            {
+                for (const auto& v : extensions)
+                    add(v, name);
+                return *this;
+            }
+    };
+    
+    static inline const std::vector<language_t> BASE_LANGUAGES = {
+            language_t().add("c", "C Source").add("h", "C/C++ Header"),
+            language_t().add({"cpp", "cxx", "cc"}, "C++ Source").add("h", "C/C++ Header").add("hpp", "C++ Header"),
+            language_t().add("java", "Java Source"),
+            language_t().add("rs", "Rust Source"),
+            language_t().add({"py", "pyo"}, "Python Source"),
+            language_t().add("rb", "Ruby Source"),
+            language_t().add("pl", "Perl Source"),
+            
+    };
+    
+    
+}
+
 #endif // BLT_FLOC_ANALYZER_H