Add version 2.7.1

* Fix debug conan build target

* Add viewcoverage and fix coverage generation

* Add more tests to cover new integrity checks

* Add tests to accomplish 100%

* Fix conan-create makefile target

.gitignore

@@ -40,3 +40,4 @@ build_release
 cmake-*
 **/CMakeFiles
 **/gcovr-report
+CMakeUserPresets.json

CHANGELOG.md

@@ -5,44 +5,53 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [Unreleased]
+## [2.1.0] - 2025-06-28
 
 ### Added
+
 - Conan dependency manager support
 - Technical analysis report
 
 ### Changed
+
 - Updated README.md
 - Refactored library version and installation system
 - Updated config variable names
 
 ### Fixed
+
 - Removed unneeded semicolon
 
 ## [2.0.1] - 2024-07-22
 
 ### Added
+
 - CMake install target and make install command
 - Flag to control sample building in Makefile
 
 ### Changed
+
 - Library name changed to `fimdlp`
 - Updated version numbers across test files
 
 ### Fixed
+
 - Version number consistency in tests
 
 ## [2.0.0] - 2024-07-04
 
 ### Added
+
 - Makefile with build & test actions for easier development
 - PyTorch (libtorch) integration for tensor operations
 
 ### Changed
+
 - Major refactoring of build system
 - Updated build workflows and CI configuration
 
 ### Fixed
+
 - BinDisc quantile calculation errors (#9)
 - Error in percentile method calculation
 - Integer type issues in calculations
@@ -51,19 +60,23 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [1.2.1] - 2024-06-08
 
 ### Added
+
 - PyTorch tensor methods for discretization
 - Improved library build system
 
 ### Changed
+
 - Refactored sample build process
 
 ### Fixed
+
 - Library creation and linking issues
 - Multiple GitHub Actions workflow fixes
 
 ## [1.2.0] - 2024-06-05
 
 ### Added
+
 - **Discretizer** - Abstract base class for all discretization algorithms (#8)
 - **BinDisc** - K-bins discretization with quantile and uniform strategies (#7)
 - Transform method to discretize values using existing cut points
@@ -71,11 +84,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Docker development container configuration
 
 ### Changed
+
 - Refactored system types throughout the library
 - Improved sample program with better dataset handling
 - Enhanced build system with debug options
 
 ### Fixed
+
 - Transform method initialization issues
 - ARFF file attribute name extraction
 - Sample program library binary separation
@@ -83,17 +98,20 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [1.1.3] - 2024-06-05
 
 ### Added
+
 - `max_cutpoints` hyperparameter for controlling algorithm complexity
 - `max_depth` and `min_length` as configurable hyperparameters
 - Enhanced sample program with hyperparameter support
 - Additional datasets for testing
 
 ### Changed
+
 - Improved constructor design and parameter handling
 - Enhanced test coverage and reporting
 - Refactored build system configuration
 
 ### Fixed
+
 - Depth initialization in fit method
 - Code quality improvements and smell fixes
 - Exception handling in value cut point calculations
@@ -101,29 +119,35 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [1.1.2] - 2023-04-01
 
 ### Added
+
 - Comprehensive test suite with GitHub Actions CI
 - SonarCloud integration for code quality analysis
 - Enhanced build system with automated testing
 
 ### Changed
+
 - Improved GitHub Actions workflow configuration
 - Updated project structure for better maintainability
 
 ### Fixed
+
 - Build system configuration issues
 - Test execution and coverage reporting
 
 ## [1.1.1] - 2023-02-22
 
 ### Added
+
 - Limits header for proper compilation
 - Enhanced build system support
 
 ### Changed
+
 - Updated version numbering system
 - Improved SonarCloud configuration
 
 ### Fixed
+
 - ValueCutPoint exception handling (removed unnecessary exception)
 - Build system compatibility issues
 - GitHub Actions token configuration
@@ -131,17 +155,20 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [1.1.0] - 2023-02-21
 
 ### Added
+
 - Classic algorithm implementation for performance comparison
 - Enhanced ValueCutPoint logic with same_values detection
 - Glass dataset support in sample program
 - Debug configuration for development
 
 ### Changed
+
 - Refactored ValueCutPoint algorithm for better accuracy
 - Improved candidate selection logic
 - Enhanced sample program with multiple datasets
 
 ### Fixed
+
 - Sign error in valueCutPoint calculation
 - Final cut value computation
 - Duplicate dataset handling in sample
@@ -149,6 +176,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [1.0.0.0] - 2022-12-21
 
 ### Added
+
 - Initial release of MDLP (Minimum Description Length Principle) discretization library
 - Core CPPFImdlp algorithm implementation based on Fayyad & Irani's paper
 - Entropy and information gain calculation methods
@@ -158,6 +186,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - ARFF file format support for datasets
 
 ### Features
+
 - Recursive discretization using entropy-based criteria
 - Stable sorting with tie-breaking for identical values
 - Configurable algorithm parameters
@@ -168,15 +197,18 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## Release Notes
 
 ### Version 2.x
+
 - **Breaking Changes**: Library renamed to `fimdlp`
 - **Major Enhancement**: PyTorch integration for improved performance
 - **New Features**: Comprehensive discretization framework with multiple algorithms
 
 ### Version 1.x
+
 - **Core Algorithm**: MDLP discretization implementation
 - **Extensibility**: Hyperparameter support and algorithm variants
 - **Quality**: Comprehensive testing and CI/CD pipeline
 
 ### Version 1.0.x
+
 - **Foundation**: Initial stable implementation
 - **Algorithm**: Core MDLP discretization functionality

CMakeLists.txt

@@ -10,12 +10,11 @@ set(CMAKE_CXX_STANDARD 17)
 cmake_policy(SET CMP0135 NEW)
 
 # Find dependencies
-find_package(Torch REQUIRED)
+find_package(Torch CONFIG REQUIRED)
 
 # Options
 # -------
 option(ENABLE_TESTING OFF)
-option(ENABLE_SAMPLE  OFF)
 option(COVERAGE       OFF)
 
 add_subdirectory(config)
@@ -26,21 +25,24 @@ if (NOT ${CMAKE_SYSTEM_NAME} MATCHES "Darwin")
     set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fno-default-inline")
 endif()
 
-if (ENABLE_TESTING)
+if (CMAKE_BUILD_TYPE STREQUAL "Debug")
-    message("Debug mode")
+    message(STATUS "Debug mode")
+else()
+    message(STATUS "Release mode")
+endif()
 
+if (ENABLE_TESTING)
+    message(STATUS "Testing is enabled")
     enable_testing()
     set(CODE_COVERAGE ON)
     set(GCC_COVERAGE_LINK_FLAGS "${GCC_COVERAGE_LINK_FLAGS} -lgcov --coverage")
     add_subdirectory(tests)
 else()
-    message("Release mode")
+    message(STATUS "Testing is disabled")
 endif()
 
-if (ENABLE_SAMPLE)
+message(STATUS "Building sample")
-    message("Building sample")
+add_subdirectory(sample)
-    add_subdirectory(sample)
-endif()
 
 include_directories(
     ${fimdlp_SOURCE_DIR}/src
@@ -62,11 +64,10 @@ write_basic_package_version_file(
 install(TARGETS fimdlp
         EXPORT fimdlpTargets
         ARCHIVE DESTINATION lib
-        LIBRARY DESTINATION lib
+        LIBRARY DESTINATION lib)
-        CONFIGURATIONS Release)
 
-install(DIRECTORY src/ DESTINATION include/fimdlp FILES_MATCHING CONFIGURATIONS Release PATTERN "*.h")
+install(DIRECTORY src/ DESTINATION include/fimdlp FILES_MATCHING PATTERN "*.h")
-install(FILES ${CMAKE_BINARY_DIR}/configured_files/include/config.h DESTINATION include/fimdlp CONFIGURATIONS Release)
+install(FILES ${CMAKE_BINARY_DIR}/configured_files/include/config.h DESTINATION include/fimdlp)
 
 install(EXPORT fimdlpTargets
         FILE fimdlpTargets.cmake

CMakeUserPresets.json

@@ -1,10 +0,0 @@
-{
-    "version": 4,
-    "vendor": {
-        "conan": {}
-    },
-    "include": [
-        "build_release/build/Release/generators/CMakePresets.json",
-        "build_debug/build/Debug/generators/CMakePresets.json"
-    ]
-}

Makefile

@@ -1,35 +1,43 @@
 SHELL := /bin/bash
-.DEFAULT_GOAL := build
+.DEFAULT_GOAL := release
-.PHONY: build install test
+.PHONY: debug release install test conan-create viewcoverage
 lcov := lcov
 
 f_debug = build_debug
 f_release = build_release
+genhtml = genhtml
+docscdir = docs
 
-build: ## Build the project for Release
+define build_target
-	@echo ">>> Building the project for Release..."
+	@echo ">>> Building the project for $(1)..."
-	@if [ -d $(f_release) ]; then rm -fr $(f_release); fi
+	@if [ -d $(2) ]; then rm -fr $(2); fi
-	@conan install . --build=missing -of $(f_release) -s build_type=Release --profile:build=default --profile:host=default
+	@conan install . --build=missing -of $(2) -s build_type=$(1)
-	cmake -S . -B $(f_release) -DCMAKE_TOOLCHAIN_FILE=$(f_release)/build/Release/generators/conan_toolchain.cmake -DCMAKE_BUILD_TYPE=Release -DENABLE_TESTING=OFF -DENABLE_SAMPLE=ON
+	@cmake -S . -B $(2) -DCMAKE_TOOLCHAIN_FILE=$(2)/build/$(1)/generators/conan_toolchain.cmake -DCMAKE_BUILD_TYPE=$(1) -D$(3)
-	@cmake --build $(f_release) -j 8
+	@cmake --build $(2) --config $(1) -j 8
+endef
+
+debug: ## Build Debug version of the library
+	@$(call build_target,"Debug","$(f_debug)", "ENABLE_TESTING=ON")
+
+release: ## Build Release version of the library
+	@$(call build_target,"Release","$(f_release)", "ENABLE_TESTING=OFF")
 
 install: ## Install the project
 	@echo ">>> Installing the project..."
-	@cmake --build build_release --target install -j 8		
+	@cmake --build $(f_release) --target install -j 8		
 
 test: ## Build Debug version and run tests
 	@echo ">>> Building Debug version and running tests..."
-	@if [ -d $(f_debug) ]; then rm -fr $(f_debug); fi
+	@$(MAKE) debug;
-	@conan install . --build=missing -of $(f_debug) -s build_type=Debug
+	@cp -r tests/datasets $(f_debug)/tests/datasets
-	@cmake -B $(f_debug) -S . -DCMAKE_BUILD_TYPE=Debug -DCMAKE_TOOLCHAIN_FILE=$(f_debug)/build/Debug/generators/conan_toolchain.cmake -DENABLE_TESTING=ON -DENABLE_SAMPLE=ON
-	@cmake --build $(f_debug) -j 8
 	@cd $(f_debug)/tests && ctest --output-on-failure -j 8
 	@cd $(f_debug)/tests && $(lcov) --capture --directory ../ --demangle-cpp --ignore-errors source,source --ignore-errors mismatch --output-file coverage.info >/dev/null 2>&1; \
 	$(lcov) --remove coverage.info '/usr/*' --output-file coverage.info >/dev/null 2>&1; \
 	$(lcov) --remove coverage.info 'lib/*' --output-file coverage.info >/dev/null 2>&1; \
 	$(lcov) --remove coverage.info 'libtorch/*' --output-file coverage.info >/dev/null 2>&1; \
 	$(lcov) --remove coverage.info 'tests/*' --output-file coverage.info >/dev/null 2>&1; \
-	$(lcov) --remove coverage.info 'gtest/*' --output-file coverage.info >/dev/null 2>&1;
+	$(lcov) --remove coverage.info 'gtest/*' --output-file coverage.info >/dev/null 2>&1; \
+	$(lcov) --remove coverage.info '*/.conan2/*' --ignore-errors unused --output-file coverage.info >/dev/null 2>&1;
 	@genhtml $(f_debug)/tests/coverage.info --demangle-cpp --output-directory $(f_debug)/tests/coverage --title "Discretizer mdlp Coverage Report" -s -k -f --legend
 	@echo "* Coverage report is generated at $(f_debug)/tests/coverage/index.html"
 	@which python || (echo ">>> Please install python"; exit 1)
@@ -39,3 +47,24 @@ test: ## Build Debug version and run tests
 	fi
 	@echo ">>> Updating coverage badge..."
 	@env python update_coverage.py $(f_debug)/tests
+	@echo ">>> Done"
+
+viewcoverage: ## View the html coverage report
+	@which $(genhtml) >/dev/null || (echo ">>> Please install lcov (genhtml not found)"; exit 1)
+	@if [ ! -d $(docscdir)/coverage ]; then mkdir -p $(docscdir)/coverage; fi
+	@if [ ! -f $(f_debug)/tests/coverage.info ]; then \
+		echo ">>> No coverage.info file found. Run make coverage first!"; \
+		exit 1; \
+	fi
+	@$(genhtml) $(f_debug)/tests/coverage.info --demangle-cpp --output-directory $(docscdir)/coverage --title "FImdlp Coverage Report" -s -k -f --legend >/dev/null 2>&1;
+	@xdg-open $(docscdir)/coverage/index.html || open $(docscdir)/coverage/index.html 2>/dev/null
+	@echo ">>> Done";
+
+conan-create: ## Create the conan package
+	@echo ">>> Creating the conan package..."
+	conan create . --build=missing -tf "" -s:a build_type=Release 
+	conan create . --build=missing -tf "" -s:a build_type=Debug -o "&:enable_testing=False"
+	@echo ">>> Done"
+
+
+

build_conan/CMakeCache.txt

@@ -1,101 +0,0 @@
-# This is the CMakeCache file.
-# For build in directory: /home/rmontanana/Code/mdlp/build_conan
-# It was generated by CMake: /usr/bin/cmake
-# You can edit this file to change values found and used by cmake.
-# If you do not want to change any of the values, simply exit the editor.
-# If you do want to change a value, simply edit, save, and exit the editor.
-# The syntax for the file is as follows:
-# KEY:TYPE=VALUE
-# KEY is the name of a variable in the cache.
-# TYPE is a hint to GUIs for the type of VALUE, DO NOT EDIT TYPE!.
-# VALUE is the current value for the KEY.
-
-########################
-# EXTERNAL cache entries
-########################
-
-//No help, variable specified on the command line.
-CMAKE_BUILD_TYPE:UNINITIALIZED=Release
-
-//Value Computed by CMake.
-CMAKE_FIND_PACKAGE_REDIRECTS_DIR:STATIC=/home/rmontanana/Code/mdlp/build_conan/CMakeFiles/pkgRedirects
-
-//Value Computed by CMake
-CMAKE_PROJECT_DESCRIPTION:STATIC=Discretization algorithm based on the paper by Fayyad & Irani Multi-Interval Discretization of Continuous-Valued Attributes for Classification Learning.
-
-//Value Computed by CMake
-CMAKE_PROJECT_HOMEPAGE_URL:STATIC=https://github.com/rmontanana/mdlp
-
-//Value Computed by CMake
-CMAKE_PROJECT_NAME:STATIC=fimdlp
-
-//Value Computed by CMake
-CMAKE_PROJECT_VERSION:STATIC=2.1.0
-
-//Value Computed by CMake
-CMAKE_PROJECT_VERSION_MAJOR:STATIC=2
-
-//Value Computed by CMake
-CMAKE_PROJECT_VERSION_MINOR:STATIC=1
-
-//Value Computed by CMake
-CMAKE_PROJECT_VERSION_PATCH:STATIC=0
-
-//Value Computed by CMake
-CMAKE_PROJECT_VERSION_TWEAK:STATIC=
-
-//No help, variable specified on the command line.
-CMAKE_TOOLCHAIN_FILE:UNINITIALIZED=conan_toolchain.cmake
-
-//Value Computed by CMake
-fimdlp_BINARY_DIR:STATIC=/home/rmontanana/Code/mdlp/build_conan
-
-//Value Computed by CMake
-fimdlp_IS_TOP_LEVEL:STATIC=ON
-
-//Value Computed by CMake
-fimdlp_SOURCE_DIR:STATIC=/home/rmontanana/Code/mdlp
-
-
-########################
-# INTERNAL cache entries
-########################
-
-//This is the directory where this CMakeCache.txt was created
-CMAKE_CACHEFILE_DIR:INTERNAL=/home/rmontanana/Code/mdlp/build_conan
-//Major version of cmake used to create the current loaded cache
-CMAKE_CACHE_MAJOR_VERSION:INTERNAL=3
-//Minor version of cmake used to create the current loaded cache
-CMAKE_CACHE_MINOR_VERSION:INTERNAL=30
-//Patch version of cmake used to create the current loaded cache
-CMAKE_CACHE_PATCH_VERSION:INTERNAL=8
-//Path to CMake executable.
-CMAKE_COMMAND:INTERNAL=/usr/bin/cmake
-//Path to cpack program executable.
-CMAKE_CPACK_COMMAND:INTERNAL=/usr/bin/cpack
-//Path to ctest program executable.
-CMAKE_CTEST_COMMAND:INTERNAL=/usr/bin/ctest
-//Path to cache edit program executable.
-CMAKE_EDIT_COMMAND:INTERNAL=/usr/bin/ccmake
-//Name of external makefile project generator.
-CMAKE_EXTRA_GENERATOR:INTERNAL=
-//Name of generator.
-CMAKE_GENERATOR:INTERNAL=Unix Makefiles
-//Generator instance identifier.
-CMAKE_GENERATOR_INSTANCE:INTERNAL=
-//Name of generator platform.
-CMAKE_GENERATOR_PLATFORM:INTERNAL=
-//Name of generator toolset.
-CMAKE_GENERATOR_TOOLSET:INTERNAL=
-//Source directory with the top level CMakeLists.txt file for this
-// project
-CMAKE_HOME_DIRECTORY:INTERNAL=/home/rmontanana/Code/mdlp
-//number of local generators
-CMAKE_NUMBER_OF_MAKEFILES:INTERNAL=1
-//Platform information initialized
-CMAKE_PLATFORM_INFO_INITIALIZED:INTERNAL=1
-//Path to CMake installation.
-CMAKE_ROOT:INTERNAL=/usr/share/cmake
-//uname command
-CMAKE_UNAME:INTERNAL=/usr/bin/uname
-

conanfile.py

@@ -1,7 +1,8 @@
+import os
+import re
 from conan import ConanFile
 from conan.tools.cmake import CMakeToolchain, CMake, cmake_layout, CMakeDeps
-from conan.tools.files import copy
+from conan.tools.files import load, copy
-import os
 
 
 class FimdlpConan(ConanFile):
@@ -32,14 +33,13 @@ class FimdlpConan(ConanFile):
     exports_sources = "CMakeLists.txt", "src/*", "sample/*", "tests/*", "config/*", "fimdlpConfig.cmake.in"
 
     def set_version(self):
-        # Read the CMakeLists.txt file to get the version
+        content = load(self, "CMakeLists.txt")
-        try:
+        version_pattern = re.compile(r'project\s*\([^\)]*VERSION\s+([0-9]+\.[0-9]+\.[0-9]+)', re.IGNORECASE | re.DOTALL)
-            content = load(self, "CMakeLists.txt")
+        match = version_pattern.search(content)
-            match = re.search(r"VERSION\s+(\d+\.\d+\.\d+)", content)
+        if match:
-            if match:
+            self.version = match.group(1)
-                self.version = match.group(1)
+        else:
-        except Exception:
+            raise Exception("Version not found in CMakeLists.txt")
-            self.version = "0.0.1"  # fallback version
     
     def config_options(self):
         if self.settings.os == "Windows":
@@ -51,7 +51,7 @@ class FimdlpConan(ConanFile):
     
     def requirements(self):
         # PyTorch dependency for tensor operations
-        self.requires("libtorch/2.7.0")
+        self.requires("libtorch/2.7.1")
         
     def build_requirements(self):
         self.requires("arff-files/1.2.0") # for tests and sample

sample/CMakeLists.txt

@@ -1,14 +1,10 @@
 set(CMAKE_CXX_STANDARD 17)
 
-set(CMAKE_BUILD_TYPE Debug)
-
 find_package(arff-files REQUIRED)
 
 include_directories(
     ${fimdlp_SOURCE_DIR}/src
-    ${fimdlp_SOURCE_DIR}/tests/lib/Files
     ${CMAKE_BINARY_DIR}/configured_files/include
-    ${libtorch_INCLUDE_DIRS_RELEASE}
     ${arff-files_INCLUDE_DIRS}
 )
 

src/BinDisc.cpp

@@ -41,19 +41,15 @@ namespace mdlp {
     }
     void BinDisc::fit(samples_t& X, labels_t& y)
     {
-        // Input validation for supervised interface
+        if (X.empty()) {
-        if (X.size() != y.size()) {
+            throw std::invalid_argument("X cannot be empty");
-            throw std::invalid_argument("X and y must have the same size");
-        }
-        if (X.empty() || y.empty()) {
-            throw std::invalid_argument("X and y cannot be empty");
         }
 
         // BinDisc is inherently unsupervised, but we validate inputs for consistency
         // Note: y parameter is validated but not used in binning strategy
         fit(X);
     }
-    std::vector<precision_t> linspace(precision_t start, precision_t end, int num)
+    std::vector<precision_t> BinDisc::linspace(precision_t start, precision_t end, int num)
     {
         // Input validation
         if (num < 2) {
@@ -81,7 +77,7 @@ namespace mdlp {
     {
         return std::max(lower, std::min(n, upper));
     }
-    std::vector<precision_t> percentile(samples_t& data, const std::vector<precision_t>& percentiles)
+    std::vector<precision_t> BinDisc::percentile(samples_t& data, const std::vector<precision_t>& percentiles)
     {
         // Input validation
         if (data.empty()) {

src/BinDisc.h

@@ -23,6 +23,9 @@ namespace mdlp {
         // y is included for compatibility with the Discretizer interface
         void fit(samples_t& X_, labels_t& y) override;
         void fit(samples_t& X);
+    protected:
+        std::vector<precision_t> linspace(precision_t start, precision_t end, int num);
+        std::vector<precision_t> percentile(samples_t& data, const std::vector<precision_t>& percentiles);
     private:
         void fit_uniform(const samples_t&);
         void fit_quantile(const samples_t&);

src/CPPFImdlp.cpp

@@ -39,7 +39,7 @@ namespace mdlp {
         if (proposed_cuts == 0) {
             return numeric_limits<size_t>::max();
         }
-        if (proposed_cuts < 0 || proposed_cuts > static_cast<precision_t>(X.size())) {
+        if (proposed_cuts > static_cast<precision_t>(X.size())) {
             throw invalid_argument("wrong proposed num_cuts value");
         }
         if (proposed_cuts < 1)
@@ -56,7 +56,7 @@ namespace mdlp {
         discretizedData.clear();
         cutPoints.clear();
         if (X.size() != y.size()) {
-            throw invalid_argument("X and y must have the same size");
+            throw std::invalid_argument("X and y must have the same size: " + std::to_string(X.size()) + " != " + std::to_string(y.size()));
         }
         if (X.empty() || y.empty()) {
             throw invalid_argument("X and y must have at least one element");
@@ -105,9 +105,10 @@ namespace mdlp {
         // # of duplicates before cutpoint
         n = safe_subtract(safe_subtract(cut, 1), idxPrev);
         // # of duplicates after cutpoint
-        m = safe_subtract(safe_subtract(idxNext, cut), 1);
+        m = idxNext - cut - 1;
         // Decide which values to use
         if (backWall) {
+            m = int(idxNext - cut - 1) < 0 ? 0 : m; // Ensure m right
             cut = cut + m + 1;
         } else {
             cut = safe_subtract(cut, n);

src/CPPFImdlp.h

@@ -39,8 +39,8 @@ namespace mdlp {
         size_t getCandidate(size_t, size_t);
         size_t compute_max_num_cut_points() const;
         pair<precision_t, size_t> valueCutPoint(size_t, size_t, size_t);
-    private:
+        inline precision_t safe_X_access(size_t idx) const
-        inline precision_t safe_X_access(size_t idx) const {
+        {
             if (idx >= indices.size()) {
                 throw std::out_of_range("Index out of bounds for indices array");
             }
@@ -50,7 +50,8 @@ namespace mdlp {
             }
             return X[real_idx];
         }
-        inline label_t safe_y_access(size_t idx) const {
+        inline label_t safe_y_access(size_t idx) const
+        {
             if (idx >= indices.size()) {
                 throw std::out_of_range("Index out of bounds for indices array");
             }
@@ -60,7 +61,8 @@ namespace mdlp {
             }
             return y[real_idx];
         }
-        inline size_t safe_subtract(size_t a, size_t b) const {
+        inline size_t safe_subtract(size_t a, size_t b) const
+        {
             if (b > a) {
                 throw std::underflow_error("Subtraction would cause underflow");
             }

src/Discretizer.cpp

@@ -40,9 +40,6 @@ namespace mdlp {
     void Discretizer::fit_t(const torch::Tensor& X_, const torch::Tensor& y_)
     {
         // Validate tensor properties for security
-        if (!X_.is_contiguous() || !y_.is_contiguous()) {
-            throw std::invalid_argument("Tensors must be contiguous");
-        }
         if (X_.sizes().size() != 1 || y_.sizes().size() != 1) {
             throw std::invalid_argument("Only 1D tensors supported");
         }
@@ -67,9 +64,6 @@ namespace mdlp {
     torch::Tensor Discretizer::transform_t(const torch::Tensor& X_)
     {
         // Validate tensor properties for security
-        if (!X_.is_contiguous()) {
-            throw std::invalid_argument("Tensor must be contiguous");
-        }
         if (X_.sizes().size() != 1) {
             throw std::invalid_argument("Only 1D tensors supported");
         }
@@ -88,9 +82,6 @@ namespace mdlp {
     torch::Tensor Discretizer::fit_transform_t(const torch::Tensor& X_, const torch::Tensor& y_)
     {
         // Validate tensor properties for security
-        if (!X_.is_contiguous() || !y_.is_contiguous()) {
-            throw std::invalid_argument("Tensors must be contiguous");
-        }
         if (X_.sizes().size() != 1 || y_.sizes().size() != 1) {
             throw std::invalid_argument("Only 1D tensors supported");
         }

test_package/CMakeLists.txt

@@ -2,7 +2,8 @@ cmake_minimum_required(VERSION 3.20)
 project(test_fimdlp)
 
 find_package(fimdlp REQUIRED)
+find_package(Torch REQUIRED)
 
 add_executable(test_fimdlp src/test_fimdlp.cpp)
-target_link_libraries(test_fimdlp fimdlp::fimdlp)
+target_link_libraries(test_fimdlp fimdlp::fimdlp torch::torch)
 target_compile_features(test_fimdlp PRIVATE cxx_std_17)

test_package/CMakeUserPresets.json

@@ -0,0 +1,10 @@
+{
+    "version": 4,
+    "vendor": {
+        "conan": {}
+    },
+    "include": [
+        "build/gcc-14-x86_64-gnu17-release/generators/CMakePresets.json",
+        "build/gcc-14-x86_64-gnu17-debug/generators/CMakePresets.json"
+    ]
+}

tests/BinDisc_unittest.cpp

@@ -11,6 +11,16 @@
 #include <ArffFiles.hpp>
 #include "BinDisc.h"
 #include "Experiments.hpp"
+#include <cmath>
+
+#define EXPECT_THROW_WITH_MESSAGE(stmt, etype, whatstring) EXPECT_THROW( \
+try { \
+stmt; \
+} catch (const etype& ex) { \
+EXPECT_EQ(whatstring, std::string(ex.what())); \
+throw; \
+} \
+, etype)
 
 namespace mdlp {
     const float margin = 1e-4;
@@ -153,20 +163,12 @@ namespace mdlp {
     TEST_F(TestBinDisc3U, EmptyUniform)
     {
         samples_t X = {};
-        fit(X);
+        EXPECT_THROW(fit(X), std::invalid_argument);
-        auto cuts = getCutPoints();
-        ASSERT_EQ(2, cuts.size());
-        EXPECT_NEAR(0, cuts.at(0), margin);
-        EXPECT_NEAR(0, cuts.at(1), margin);
     }
     TEST_F(TestBinDisc3Q, EmptyQuantile)
     {
         samples_t X = {};
-        fit(X);
+        EXPECT_THROW(fit(X), std::invalid_argument);
-        auto cuts = getCutPoints();
-        ASSERT_EQ(2, cuts.size());
-        EXPECT_NEAR(0, cuts.at(0), margin);
-        EXPECT_NEAR(0, cuts.at(1), margin);
     }
     TEST(TestBinDisc3, ExceptionNumberBins)
     {
@@ -406,6 +408,66 @@ namespace mdlp {
                 EXPECT_NEAR(exp.cutpoints_.at(i), cuts.at(i), margin);
             }
         }
-        std::cout << "* Number of experiments tested: " << num << std::endl;
+        // std::cout << "* Number of experiments tested: " << num << std::endl;
+    }
+
+    TEST_F(TestBinDisc3U, FitDataSizeTooSmall)
+    {
+        // Test when data size is smaller than n_bins
+        samples_t X = { 1.0, 2.0 }; // Only 2 elements for 3 bins
+        EXPECT_THROW_WITH_MESSAGE(fit(X), std::invalid_argument, "Input data size must be at least equal to n_bins");
+    }
+
+    TEST_F(TestBinDisc3Q, FitDataSizeTooSmall)
+    {
+        // Test when data size is smaller than n_bins
+        samples_t X = { 1.0, 2.0 }; // Only 2 elements for 3 bins
+        EXPECT_THROW_WITH_MESSAGE(fit(X), std::invalid_argument, "Input data size must be at least equal to n_bins");
+    }
+
+    TEST_F(TestBinDisc3U, FitWithYEmptyX)
+    {
+        // Test fit(X, y) with empty X
+        samples_t X = {};
+        labels_t y = { 1, 2, 3 };
+        EXPECT_THROW_WITH_MESSAGE(fit(X, y), std::invalid_argument, "X cannot be empty");
+    }
+
+    TEST_F(TestBinDisc3U, LinspaceInvalidNumPoints)
+    {
+        // Test linspace with num < 2
+        EXPECT_THROW_WITH_MESSAGE(linspace(0.0f, 1.0f, 1), std::invalid_argument, "Number of points must be at least 2 for linspace");
+    }
+
+    TEST_F(TestBinDisc3U, LinspaceNaNValues)
+    {
+        // Test linspace with NaN values
+        float nan_val = std::numeric_limits<float>::quiet_NaN();
+        EXPECT_THROW_WITH_MESSAGE(linspace(nan_val, 1.0f, 3), std::invalid_argument, "Start and end values cannot be NaN");
+        EXPECT_THROW_WITH_MESSAGE(linspace(0.0f, nan_val, 3), std::invalid_argument, "Start and end values cannot be NaN");
+    }
+
+    TEST_F(TestBinDisc3U, LinspaceInfiniteValues)
+    {
+        // Test linspace with infinite values
+        float inf_val = std::numeric_limits<float>::infinity();
+        EXPECT_THROW_WITH_MESSAGE(linspace(inf_val, 1.0f, 3), std::invalid_argument, "Start and end values cannot be infinite");
+        EXPECT_THROW_WITH_MESSAGE(linspace(0.0f, inf_val, 3), std::invalid_argument, "Start and end values cannot be infinite");
+    }
+
+    TEST_F(TestBinDisc3U, PercentileEmptyData)
+    {
+        // Test percentile with empty data
+        samples_t empty_data = {};
+        std::vector<precision_t> percentiles = { 25.0f, 50.0f, 75.0f };
+        EXPECT_THROW_WITH_MESSAGE(percentile(empty_data, percentiles), std::invalid_argument, "Data cannot be empty for percentile calculation");
+    }
+
+    TEST_F(TestBinDisc3U, PercentileEmptyPercentiles)
+    {
+        // Test percentile with empty percentiles
+        samples_t data = { 1.0f, 2.0f, 3.0f };
+        std::vector<precision_t> empty_percentiles = {};
+        EXPECT_THROW_WITH_MESSAGE(percentile(data, empty_percentiles), std::invalid_argument, "Percentiles cannot be empty");
     }
 }

tests/CMakeLists.txt

@@ -1,6 +1,7 @@
 
 find_package(arff-files REQUIRED)
 find_package(GTest REQUIRED)
+find_package(Torch CONFIG REQUIRED)
 
 include_directories(
         ${libtorch_INCLUDE_DIRS_DEBUG}

tests/Discretizer_unittest.cpp

@@ -13,17 +13,26 @@
 #include "BinDisc.h"
 #include "CPPFImdlp.h"
 
+#define EXPECT_THROW_WITH_MESSAGE(stmt, etype, whatstring) EXPECT_THROW( \
+try { \
+stmt; \
+} catch (const etype& ex) { \
+EXPECT_EQ(whatstring, std::string(ex.what())); \
+throw; \
+} \
+, etype)
+
 namespace mdlp {
     const float margin = 1e-4;
     static std::string set_data_path()
     {
-        std::string path = "datasets/";
+        std::string path = "tests/datasets/";
         std::ifstream file(path + "iris.arff");
         if (file.is_open()) {
             file.close();
             return path;
         }
-        return "tests/datasets/";
+        return "datasets/";
     }
     const std::string data_path = set_data_path();
     const labels_t iris_quantile = { 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 2, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 3, 3, 3, 1, 3, 1, 2, 0, 3, 1, 0, 2, 2, 2, 1, 3, 1, 2, 2, 1, 2, 2, 2, 2, 3, 3, 3, 3, 2, 1, 1, 1, 2, 2, 1, 2, 3, 2, 1, 1, 1, 2, 2, 0, 1, 1, 1, 2, 1, 1, 2, 2, 3, 2, 3, 3, 0, 3, 3, 3, 3, 3, 3, 1, 2, 3, 3, 3, 3, 2, 3, 1, 3, 2, 3, 3, 2, 2, 3, 3, 3, 3, 3, 2, 2, 3, 2, 3, 2, 3, 3, 3, 2, 3, 3, 3, 2, 3, 2, 2 };
@@ -32,7 +41,6 @@ namespace mdlp {
         Discretizer* disc = new BinDisc(4, strategy_t::UNIFORM);
         auto version = disc->version();
         delete disc;
-        std::cout << "Version computed: " << version;
         EXPECT_EQ("2.1.0", version);
     }
     TEST(Discretizer, BinIrisUniform)
@@ -271,4 +279,110 @@ namespace mdlp {
             EXPECT_EQ(computed[i], expected[i]);
         }
     }
+
+    TEST(Discretizer, TransformEmptyData)
+    {
+        Discretizer* disc = new BinDisc(4, strategy_t::UNIFORM);
+        samples_t empty_data = {};
+        EXPECT_THROW_WITH_MESSAGE(disc->transform(empty_data), std::invalid_argument, "Data for transformation cannot be empty");
+        delete disc;
+    }
+
+    TEST(Discretizer, TransformNotFitted)
+    {
+        Discretizer* disc = new BinDisc(4, strategy_t::UNIFORM);
+        samples_t data = { 1.0f, 2.0f, 3.0f };
+        EXPECT_THROW_WITH_MESSAGE(disc->transform(data), std::runtime_error, "Discretizer not fitted yet or no valid cut points found");
+        delete disc;
+    }
+
+    TEST(Discretizer, TensorValidationFit)
+    {
+        Discretizer* disc = new BinDisc(4, strategy_t::UNIFORM);
+
+        auto X = torch::tensor({ 1.0f, 2.0f, 3.0f }, torch::kFloat32);
+        auto y = torch::tensor({ 1, 2, 3 }, torch::kInt32);
+
+        // Test non-1D tensors
+        auto X_2d = torch::tensor({ {1.0f, 2.0f}, {3.0f, 4.0f} }, torch::kFloat32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_t(X_2d, y), std::invalid_argument, "Only 1D tensors supported");
+
+        auto y_2d = torch::tensor({ {1, 2}, {3, 4} }, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_t(X, y_2d), std::invalid_argument, "Only 1D tensors supported");
+
+        // Test wrong tensor types
+        auto X_int = torch::tensor({ 1, 2, 3 }, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_t(X_int, y), std::invalid_argument, "X tensor must be Float32 type");
+
+        auto y_float = torch::tensor({ 1.0f, 2.0f, 3.0f }, torch::kFloat32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_t(X, y_float), std::invalid_argument, "y tensor must be Int32 type");
+
+        // Test mismatched sizes
+        auto y_short = torch::tensor({ 1, 2 }, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_t(X, y_short), std::invalid_argument, "X and y tensors must have same number of elements");
+
+        // Test empty tensors
+        auto X_empty = torch::tensor({}, torch::kFloat32);
+        auto y_empty = torch::tensor({}, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_t(X_empty, y_empty), std::invalid_argument, "Tensors cannot be empty");
+
+        delete disc;
+    }
+
+    TEST(Discretizer, TensorValidationTransform)
+    {
+        Discretizer* disc = new BinDisc(4, strategy_t::UNIFORM);
+
+        // First fit with valid data
+        auto X_fit = torch::tensor({ 1.0f, 2.0f, 3.0f, 4.0f }, torch::kFloat32);
+        auto y_fit = torch::tensor({ 1, 2, 3, 4 }, torch::kInt32);
+        disc->fit_t(X_fit, y_fit);
+
+        // Test non-1D tensor
+        auto X_2d = torch::tensor({ {1.0f, 2.0f}, {3.0f, 4.0f} }, torch::kFloat32);
+        EXPECT_THROW_WITH_MESSAGE(disc->transform_t(X_2d), std::invalid_argument, "Only 1D tensors supported");
+
+        // Test wrong tensor type
+        auto X_int = torch::tensor({ 1, 2, 3 }, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->transform_t(X_int), std::invalid_argument, "X tensor must be Float32 type");
+
+        // Test empty tensor
+        auto X_empty = torch::tensor({}, torch::kFloat32);
+        EXPECT_THROW_WITH_MESSAGE(disc->transform_t(X_empty), std::invalid_argument, "Tensor cannot be empty");
+
+        delete disc;
+    }
+
+    TEST(Discretizer, TensorValidationFitTransform)
+    {
+        Discretizer* disc = new BinDisc(4, strategy_t::UNIFORM);
+
+        auto X = torch::tensor({ 1.0f, 2.0f, 3.0f }, torch::kFloat32);
+        auto y = torch::tensor({ 1, 2, 3 }, torch::kInt32);
+
+        // Test non-1D tensors
+        auto X_2d = torch::tensor({ {1.0f, 2.0f}, {3.0f, 4.0f} }, torch::kFloat32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_transform_t(X_2d, y), std::invalid_argument, "Only 1D tensors supported");
+
+        auto y_2d = torch::tensor({ {1, 2}, {3, 4} }, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_transform_t(X, y_2d), std::invalid_argument, "Only 1D tensors supported");
+
+        // Test wrong tensor types
+        auto X_int = torch::tensor({ 1, 2, 3 }, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_transform_t(X_int, y), std::invalid_argument, "X tensor must be Float32 type");
+
+        auto y_float = torch::tensor({ 1.0f, 2.0f, 3.0f }, torch::kFloat32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_transform_t(X, y_float), std::invalid_argument, "y tensor must be Int32 type");
+
+        // Test mismatched sizes
+        auto y_short = torch::tensor({ 1, 2 }, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_transform_t(X, y_short), std::invalid_argument, "X and y tensors must have same number of elements");
+
+        // Test empty tensors
+        auto X_empty = torch::tensor({}, torch::kFloat32);
+        auto y_empty = torch::tensor({}, torch::kInt32);
+        EXPECT_THROW_WITH_MESSAGE(disc->fit_transform_t(X_empty, y_empty), std::invalid_argument, "Tensors cannot be empty");
+
+        delete disc;
+    }
 }

tests/FImdlp_unittest.cpp

@@ -64,7 +64,7 @@ namespace mdlp {
         {
             EXPECT_EQ(computed.size(), expected.size());
             for (unsigned long i = 0; i < computed.size(); i++) {
-                cout << "(" << computed[i] << ", " << expected[i] << ") ";
+                // cout << "(" << computed[i] << ", " << expected[i] << ") ";
                 EXPECT_NEAR(computed[i], expected[i], precision);
             }
         }
@@ -76,7 +76,7 @@ namespace mdlp {
             X = X_;
             y = y_;
             indices = sortIndices(X, y);
-            cout << "* " << title << endl;
+            // cout << "* " << title << endl;
             result = valueCutPoint(0, cut, 10);
             EXPECT_NEAR(result.first, midPoint, precision);
             EXPECT_EQ(result.second, limit);
@@ -95,9 +95,9 @@ namespace mdlp {
                 test.fit(X[feature], y);
                 EXPECT_EQ(test.get_depth(), depths[feature]);
                 auto computed = test.getCutPoints();
-                cout << "Feature " << feature << ": ";
+                // cout << "Feature " << feature << ": ";
                 checkCutPoints(computed, expected[feature]);
-                cout << endl;
+                // cout << endl;
             }
         }
     };
@@ -113,17 +113,16 @@ namespace mdlp {
     {
         X = { 1, 2, 3 };
         y = { 1, 2 };
-        EXPECT_THROW_WITH_MESSAGE(fit(X, y), invalid_argument, "X and y must have the same size");
+        EXPECT_THROW_WITH_MESSAGE(fit(X, y), invalid_argument, "X and y must have the same size: " + std::to_string(X.size()) + " != " + std::to_string(y.size()));
     }
 
-    TEST_F(TestFImdlp, FitErrorMinLengtMaxDepth)
+    TEST_F(TestFImdlp, FitErrorMinLength)
     {
-        auto testLength = CPPFImdlp(2, 10, 0);
+        EXPECT_THROW_WITH_MESSAGE(CPPFImdlp(2, 10, 0), invalid_argument, "min_length must be greater than 2");
-        auto testDepth = CPPFImdlp(3, 0, 0);
+    }
-        X = { 1, 2, 3 };
+    TEST_F(TestFImdlp, FitErrorMaxDepth)
-        y = { 1, 2, 3 };
+    {
-        EXPECT_THROW_WITH_MESSAGE(testLength.fit(X, y), invalid_argument, "min_length must be greater than 2");
+        EXPECT_THROW_WITH_MESSAGE(CPPFImdlp(3, 0, 0), invalid_argument, "max_depth must be greater than 0");
-        EXPECT_THROW_WITH_MESSAGE(testDepth.fit(X, y), invalid_argument, "max_depth must be greater than 0");
     }
 
     TEST_F(TestFImdlp, JoinFit)
@@ -137,14 +136,16 @@ namespace mdlp {
         checkCutPoints(computed, expected);
     }
 
+    TEST_F(TestFImdlp, FitErrorMinCutPoints)
+    {
+        EXPECT_THROW_WITH_MESSAGE(CPPFImdlp(3, 10, -1), invalid_argument, "proposed_cuts must be non-negative");
+    }
     TEST_F(TestFImdlp, FitErrorMaxCutPoints)
     {
-        auto testmin = CPPFImdlp(2, 10, -1);
+        auto test = CPPFImdlp(3, 1, 8);
-        auto testmax = CPPFImdlp(3, 0, 200);
+        samples_t X_ = { 1, 2, 2, 3, 4, 2, 3 };
-        X = { 1, 2, 3 };
+        labels_t y_ = { 0, 0, 1, 2, 3, 4, 5 };
-        y = { 1, 2, 3 };
+        EXPECT_THROW_WITH_MESSAGE(test.fit(X_, y_), invalid_argument, "wrong proposed num_cuts value");
-        EXPECT_THROW_WITH_MESSAGE(testmin.fit(X, y), invalid_argument, "wrong proposed num_cuts value");
-        EXPECT_THROW_WITH_MESSAGE(testmax.fit(X, y), invalid_argument, "wrong proposed num_cuts value");
     }
 
     TEST_F(TestFImdlp, SortIndices)
@@ -166,6 +167,15 @@ namespace mdlp {
         indices = { 1, 2, 0 };
     }
 
+    TEST_F(TestFImdlp, SortIndicesOutOfBounds)
+    {
+        // Test for out of bounds exception in sortIndices
+        samples_t X_long = { 1.0f, 2.0f, 3.0f };
+        labels_t y_short = { 1, 2 };
+        EXPECT_THROW_WITH_MESSAGE(sortIndices(X_long, y_short), std::out_of_range, "Index out of bounds in sort comparison");
+    }
+
+
     TEST_F(TestFImdlp, TestShortDatasets)
     {
         vector<precision_t> computed;
@@ -363,4 +373,55 @@ namespace mdlp {
             EXPECT_EQ(computed_ft[i], expected[i]);
         }
     }
+    TEST_F(TestFImdlp, SafeXAccessIndexOutOfBounds)
+    {
+        // Test safe_X_access with index out of bounds for indices array
+        X = { 1.0f, 2.0f, 3.0f };
+        y = { 1, 2, 3 };
+        indices = { 0, 1 }; // shorter than expected
+
+        // This should trigger the first exception in safe_X_access (idx >= indices.size())
+        EXPECT_THROW_WITH_MESSAGE(safe_X_access(2), std::out_of_range, "Index out of bounds for indices array");
+    }
+
+    TEST_F(TestFImdlp, SafeXAccessXOutOfBounds)
+    {
+        // Test safe_X_access with real_idx out of bounds for X array
+        X = { 1.0f, 2.0f }; // shorter array
+        y = { 1, 2, 3 };
+        indices = { 0, 1, 5 }; // indices[2] = 5 is out of bounds for X
+
+        // This should trigger the second exception in safe_X_access (real_idx >= X.size())
+        EXPECT_THROW_WITH_MESSAGE(safe_X_access(2), std::out_of_range, "Index out of bounds for X array");
+    }
+
+    TEST_F(TestFImdlp, SafeYAccessIndexOutOfBounds)
+    {
+        // Test safe_y_access with index out of bounds for indices array
+        X = { 1.0f, 2.0f, 3.0f };
+        y = { 1, 2, 3 };
+        indices = { 0, 1 }; // shorter than expected
+
+        // This should trigger the first exception in safe_y_access (idx >= indices.size())
+        EXPECT_THROW_WITH_MESSAGE(safe_y_access(2), std::out_of_range, "Index out of bounds for indices array");
+    }
+
+    TEST_F(TestFImdlp, SafeYAccessYOutOfBounds)
+    {
+        // Test safe_y_access with real_idx out of bounds for y array
+        X = { 1.0f, 2.0f, 3.0f };
+        y = { 1, 2 }; // shorter array
+        indices = { 0, 1, 5 }; // indices[2] = 5 is out of bounds for y
+
+        // This should trigger the second exception in safe_y_access (real_idx >= y.size())
+        EXPECT_THROW_WITH_MESSAGE(safe_y_access(2), std::out_of_range, "Index out of bounds for y array");
+    }
+
+    TEST_F(TestFImdlp, SafeSubtractUnderflow)
+    {
+        // Test safe_subtract with underflow condition (b > a)
+        EXPECT_THROW_WITH_MESSAGE(safe_subtract(3, 5), std::underflow_error, "Subtraction would cause underflow");
+    }
+
+
 }