rmontanana/BayesNet
1
0
Fork 0
Code Issues 2 Pull Requests 1 Actions Packages Projects Releases 6 Wiki Activity

Fix Conan debug build

Browse Source 
Fix smell issues in markdown and python
This commit is contained in:
Ricardo Montañana Gómez
2025-07-02 00:44:24 +02:00
parent 28be43db02 dfa74056f5
commit 189d314990
5 changed files with 49 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
CLAUDE.md
View File

@@ -9,14 +9,17 @@ BayesNet is a C++ library implementing Bayesian Network Classifiers. It provides
## Build System & Dependencies
### Dependency Management
The project supports **two package managers**:
#### vcpkg (Default)
- Uses vcpkg with private registry at https://github.com/rmontanana/vcpkg-stash
- Uses vcpkg with private registry at <https://github.com/rmontanana/vcpkg-stash>
- Core dependencies: libtorch, nlohmann-json, folding, fimdlp, arff-files, catch2
- All dependencies defined in `vcpkg.json` with version overrides
#### Conan (Alternative)
- Modern C++ package manager with better dependency resolution
- Configured via `conanfile.py` for packaging and distribution
- Supports subset of dependencies (libtorch, nlohmann-json, catch2)
@@ -25,6 +28,7 @@ The project supports **two package managers**:
### Build Commands
#### Using vcpkg (Default)
```bash
# Initialize dependencies
make init
@@ -49,6 +53,7 @@ make clean
```
#### Using Conan
```bash
# Install Conan first: pip install conan
@@ -74,6 +79,7 @@ make conan-clean
```
### CMake Configuration
- Uses CMake 3.27+ with C++17 standard
- Debug builds automatically enable testing and coverage
- Release builds optimize with `-Ofast`
@@ -89,6 +95,7 @@ make conan-clean
- Coverage reporting with lcov/genhtml
### Test Categories
- A2DE, BoostA2DE, BoostAODE, XSPODE, XSPnDE, XBAODE, XBA2DE
- Classifier, Ensemble, FeatureSelection, Metrics, Models
- Network, Node, MST, Modules
@@ -96,6 +103,7 @@ make conan-clean
## Code Architecture
### Core Structure
```
bayesnet/
├── BaseClassifier.h # Abstract base for all classifiers
@@ -107,12 +115,14 @@ bayesnet/
```
### Key Design Patterns
- **BaseClassifier** abstract interface for all algorithms
- Template-based design with both std::vector and torch::Tensor support
- Network/Node abstraction for Bayesian network representation
- Feature selection as separate, composable modules
### Data Handling
- Supports both discrete integer data and continuous data with discretization
- ARFF file format support through arff-files library
- Tensor operations via PyTorch C++ (libtorch)
@@ -128,6 +138,7 @@ bayesnet/
## Sample Applications
Sample code in `sample/` directory demonstrates library usage:
```bash
make sample fname=tests/data/iris.arff model=TANLd
```
@@ -135,6 +146,7 @@ make sample fname=tests/data/iris.arff model=TANLd
## Package Distribution
### Creating Conan Packages
```bash
# Create package locally
make conan-create
@@ -148,7 +160,9 @@ make conan-upload remote=myremote profile=myprofile
```
### Using the Library
With Conan:
```python
# conanfile.txt or conanfile.py
[requires]
@@ -159,6 +173,7 @@ cmake
```
With vcpkg:
```json
{
"dependencies": ["bayesnet"]
@@ -170,7 +185,7 @@ With vcpkg:
- **Add new classifier**: Extend BaseClassifier, implement in appropriate subdirectory
- **Add new test**: Update `tests/CMakeLists.txt` and create test in `tests/`
- **Modify build**: Edit main `CMakeLists.txt` or use Makefile targets
- **Update dependencies**:
- **Update dependencies**:
- vcpkg: Modify `vcpkg.json` and run `make init`
- Conan: Modify `conanfile.py` and run `make conan-init`
- **Package for distribution**: Use `make conan-create` for Conan packaging
- **Package for distribution**: Use `make conan-create` for Conan packaging

17
CMakeLists.txt
View File

@@ -18,8 +18,7 @@ set(CMAKE_CXX_EXTENSIONS OFF)
set(CMAKE_EXPORT_COMPILE_COMMANDS ON)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${TORCH_CXX_FLAGS}")
SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -pthread")
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fprofile-arcs -ftest-coverage -fno-elide-constructors")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O3")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -Ofast")
if (NOT ${CMAKE_SYSTEM_NAME} MATCHES "Darwin")
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fno-default-inline")
endif()
@@ -68,18 +67,19 @@ target_link_libraries(bayesnet
# -------
if (CMAKE_BUILD_TYPE STREQUAL "Debug")
MESSAGE("Debug mode")
set(ENABLE_TESTING ON)
set(CODE_COVERAGE ON)
else(CMAKE_BUILD_TYPE STREQUAL "Debug")
MESSAGE("Release mode")
endif (CMAKE_BUILD_TYPE STREQUAL "Debug")
if (ENABLE_TESTING)
MESSAGE(STATUS "Testing enabled")
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fprofile-arcs -ftest-coverage -fno-elide-constructors")
find_package(Catch2 CONFIG REQUIRED)
find_package(arff-files CONFIG REQUIRED)
enable_testing()
include(CTest)
add_subdirectory(tests)
else(ENABLE_TESTING)
message("Release mode")
endif (ENABLE_TESTING)
# Installation
@@ -99,17 +99,14 @@ configure_package_config_file(
install(TARGETS bayesnet
EXPORT bayesnetTargets
ARCHIVE DESTINATION lib
LIBRARY DESTINATION lib
CONFIGURATIONS Release)
LIBRARY DESTINATION lib)
install(DIRECTORY bayesnet/
DESTINATION include/bayesnet
FILES_MATCHING
CONFIGURATIONS Release
PATTERN "*.h")
install(FILES ${CMAKE_BINARY_DIR}/configured_files/include/bayesnet/config.h
DESTINATION include/bayesnet
CONFIGURATIONS Release)
DESTINATION include/bayesnet)
install(EXPORT bayesnetTargets
FILE bayesnetTargets.cmake

39
CONAN_README.md
View File

@@ -16,29 +16,28 @@ conan profile new default --detect
1. Create a `conanfile.txt` in your project:
```ini
[requires]
libtorch/2.7.0
bayesnet/1.2.0
```ini
[requires]
libtorch/2.7.0
bayesnet/1.2.0
[generators]
CMakeDeps
CMakeToolchain
[generators]
CMakeDeps
CMakeToolchain
```
```
1. Install dependencies:
2. Install dependencies:
```bash
conan install . --build=missing
```
```bash
conan install . --build=missing
```
1. In your CMakeLists.txt:
3. In your CMakeLists.txt:
```cmake
find_package(bayesnet REQUIRED)
target_link_libraries(your_target bayesnet::bayesnet)
```
```cmake
find_package(bayesnet REQUIRED)
target_link_libraries(your_target bayesnet::bayesnet)
```
### Building BayesNet with Conan
@@ -69,8 +68,8 @@ make conan-create
For the custom dependencies, you'll need to create Conan recipes:
1. **folding**: Cross-validation library
2. **fimdlp**: Discretization library
3. **arff-files**: ARFF file format parser
1. **fimdlp**: Discretization library
1. **arff-files**: ARFF file format parser
Contact the maintainer or create custom recipes for these packages.

4
Makefile
View File

@@ -218,8 +218,8 @@ release: ## Build release version using Conan
conan-create: ## Create Conan package
@echo ">>> Creating Conan package..."
@conan create . --build=missing -tf "" --profile=release -tf ""
@conan create . --build=missing -tf "" --profile=debug
@conan create . --build=missing -tf "" --profile=release
@conan create . --build=missing -tf "" --profile=debug -o "&:enable_coverage=False" -o "&:enable_testing=False"
@echo ">>> Done"
profile ?= release

6
conanfile.py
View File

@@ -18,7 +18,7 @@ class BayesNetConan(ConanFile):
"enable_testing": False,
"enable_coverage": False
}
# Sources are located in the same place as this recipe, copy them to the recipe
exports_sources = "CMakeLists.txt", "bayesnet/*", "config/*", "cmake/*", "docs/*", "tests/*", "bayesnetConfig.cmake.in"
@@ -35,7 +35,7 @@ class BayesNetConan(ConanFile):
self.version = match.group(1)
else:
raise Exception("Version not found in CMakeLists.txt")
self.version = match.group(1)
self.version = match.group(1)
def config_options(self):
if self.settings.os == "Windows":
@@ -50,7 +50,7 @@ class BayesNetConan(ConanFile):
self.requires("libtorch/2.7.0")
self.requires("nlohmann_json/3.11.3")
self.requires("folding/1.1.1") # Custom package
self.requires("fimdlp/2.1.0") # Custom package
self.requires("fimdlp/2.1.0") # Custom package
def build_requirements(self):
self.build_requires("cmake/[>=3.27]")