[MISC] Run formatting

seqan · Sep 26, 2024 · 7f4e563 · 7f4e563
1 parent 4e2e201
commit 7f4e563
Show file tree

Hide file tree

Showing 26 changed files with 1,179 additions and 883 deletions.
diff --git a/.clang-format b/.clang-format
@@ -2,8 +2,8 @@
 # SPDX-FileCopyrightText: 2016-2024, Knut Reinert & MPI für molekulare Genetik
 # SPDX-License-Identifier: CC0-1.0
 
-# Format all files in include/test folder, including std module, excluding contrib module
-# find . \( -iname "*.cpp" -or -iname "*.hpp" \) -and -not -path "./lib/*" -and -not -path "./build/*" | xargs clang-format-18 --style=file -i
+# Format all files in include/test folder
+# find . \( -iname "*.cpp" -or -iname "*.hpp" \) -and -not -path "./build/*" | xargs clang-format-18 --style=file -i
 # Staged files: git diff --name-only HEAD --diff-filter=ACMRT | grep -E "(\.cpp|\.hpp)$" | xargs clang-format-18 --style=file -i
 ---
 Language:        Cpp

diff --git a/.cmake-format.yaml b/.cmake-format.yaml
@@ -2,18 +2,20 @@
 # SPDX-FileCopyrightText: 2016-2024, Knut Reinert & MPI für molekulare Genetik
 # SPDX-License-Identifier: CC0-1.0
 
-# find . \( -iname CMakeLists.txt -o -iname *.cmake \) -a -not -path "./lib/*" -a -not -path "./build/*" | xargs cmake-format -c .cmake-format.yaml -i
+# find . \( -iname CMakeLists.txt -o -iname *.cmake \) -a -not -path "./build/*" | xargs cmake-format -c .cmake-format.yaml -i
 _help_parse: Options affecting listfile parsing
 parse:
   _help_additional_commands:
   - Specify structure for custom cmake functions
   additional_commands:
-    declare_internal_datasource:
+    declare_datasource:
+      pargs:
+        nargs: '*'
+        flags: []
       kwargs:
-        FILE: '*'
+        FILE: 1
         URL: '*'
-        URL_HASH: '*'
-        CONFIGURE: '*'
+        URL_HASH: 1
     cpmgetpackage:
       pargs: 1
       spelling: CPMGetPackage

diff --git a/cmake/CPM.cmake b/cmake/CPM.cmake
@@ -2,23 +2,22 @@
 #
 # SPDX-FileCopyrightText: Copyright (c) 2019-2023 Lars Melchior and contributors
 
-set(CPM_DOWNLOAD_VERSION 0.40.2)
-set(CPM_HASH_SUM "c8cdc32c03816538ce22781ed72964dc864b2a34a310d3b7104812a5ca2d835d")
+set (CPM_DOWNLOAD_VERSION 0.40.2)
+set (CPM_HASH_SUM "c8cdc32c03816538ce22781ed72964dc864b2a34a310d3b7104812a5ca2d835d")
 
-if(CPM_SOURCE_CACHE)
-  set(CPM_DOWNLOAD_LOCATION "${CPM_SOURCE_CACHE}/cpm/CPM_${CPM_DOWNLOAD_VERSION}.cmake")
-elseif(DEFINED ENV{CPM_SOURCE_CACHE})
-  set(CPM_DOWNLOAD_LOCATION "$ENV{CPM_SOURCE_CACHE}/cpm/CPM_${CPM_DOWNLOAD_VERSION}.cmake")
-else()
-  set(CPM_DOWNLOAD_LOCATION "${CMAKE_BINARY_DIR}/cmake/CPM_${CPM_DOWNLOAD_VERSION}.cmake")
-endif()
+if (CPM_SOURCE_CACHE)
+    set (CPM_DOWNLOAD_LOCATION "${CPM_SOURCE_CACHE}/cpm/CPM_${CPM_DOWNLOAD_VERSION}.cmake")
+elseif (DEFINED ENV{CPM_SOURCE_CACHE})
+    set (CPM_DOWNLOAD_LOCATION "$ENV{CPM_SOURCE_CACHE}/cpm/CPM_${CPM_DOWNLOAD_VERSION}.cmake")
+else ()
+    set (CPM_DOWNLOAD_LOCATION "${CMAKE_BINARY_DIR}/cmake/CPM_${CPM_DOWNLOAD_VERSION}.cmake")
+endif ()
 
 # Expand relative path. This is important if the provided path contains a tilde (~)
-get_filename_component(CPM_DOWNLOAD_LOCATION ${CPM_DOWNLOAD_LOCATION} ABSOLUTE)
+get_filename_component (CPM_DOWNLOAD_LOCATION ${CPM_DOWNLOAD_LOCATION} ABSOLUTE)
 
-file(DOWNLOAD
-     https://github.com/cpm-cmake/CPM.cmake/releases/download/v${CPM_DOWNLOAD_VERSION}/CPM.cmake
-     ${CPM_DOWNLOAD_LOCATION} EXPECTED_HASH SHA256=${CPM_HASH_SUM}
+file (DOWNLOAD https://github.com/cpm-cmake/CPM.cmake/releases/download/v${CPM_DOWNLOAD_VERSION}/CPM.cmake
+      ${CPM_DOWNLOAD_LOCATION} EXPECTED_HASH SHA256=${CPM_HASH_SUM}
 )
 
-include(${CPM_DOWNLOAD_LOCATION})
+include (${CPM_DOWNLOAD_LOCATION})
diff --git a/cmake/test/config.cmake b/cmake/test/config.cmake
@@ -19,7 +19,6 @@ if (NOT TARGET ${PROJECT_NAME}_test)
     add_library (${PROJECT_NAME}_test INTERFACE)
     target_compile_options (${PROJECT_NAME}_lib PUBLIC "-pedantic" "-Wall" "-Wextra" "-Werror")
 
-
     if ("${CMAKE_CXX_COMPILER_ID}" STREQUAL "GNU")
         # Disable bogus warnings in GCC12.
         if (CMAKE_CXX_COMPILER_VERSION VERSION_GREATER_EQUAL 12 AND CMAKE_CXX_COMPILER_VERSION VERSION_LESS 13)
@@ -36,8 +35,8 @@ if (NOT TARGET ${PROJECT_NAME}_test)
 
     # !Workaround: Get seqan3 test include dir from seqan3 target
     find_path (SEQAN3_TEST_INCLUDE_DIR
-           NAMES seqan3/test/tmp_directory.hpp
-           HINTS "${seqan3_SOURCE_DIR}/test/include"
+               NAMES seqan3/test/tmp_directory.hpp
+               HINTS "${seqan3_SOURCE_DIR}/test/include"
     )
     target_include_directories (${PROJECT_NAME}_test SYSTEM INTERFACE "${SEQAN3_TEST_INCLUDE_DIR}")
 

diff --git a/doc/CMakeLists.txt b/doc/CMakeLists.txt
@@ -19,7 +19,8 @@ if (${DOXYGEN_FOUND})
                        COMMAND ${DOXYGEN_EXECUTABLE}
                        WORKING_DIRECTORY ${APP_TEMPLATE_DOXYGEN_OUTPUT_DIR}
                        COMMENT "Generating (developer) API documentation with Doxygen."
-                       VERBATIM)
+                       VERBATIM
+    )
     message (STATUS "${FontBold}You can run `make doc` to build api documentation.${FontReset}")
 else ()
     message (STATUS "Doxygen not found.")

diff --git a/include/ibf.hpp b/include/ibf.hpp
@@ -7,37 +7,40 @@
 
 #pragma once
 
+#include <filesystem>
 #include <iostream>
 #include <math.h>
 #include <numeric>
 #include <string>
 
 #include <seqan3/alphabet/container/concatenated_sequences.hpp>
 #include <seqan3/alphabet/nucleotide/dna4.hpp>
-#include <filesystem>
 
 #include "shared.hpp"
 
 struct minimiser_arguments
 {
-    std::filesystem::path include_file; // Needs to be defined when only minimisers appearing in this file should be stored
-    std::filesystem::path exclude_file; // Needs to be defined when minimisers appearing in this file should NOT be stored
-    std::vector<int> samples{}; // Can be used to indicate that sequence files belong to the same experiment
-    bool paired = false; // If true, than experiments are seen as paired-end experiments
+    std::filesystem::path
+        include_file; // Needs to be defined when only minimisers appearing in this file should be stored
+    std::filesystem::path
+        exclude_file;              // Needs to be defined when minimisers appearing in this file should NOT be stored
+    std::vector<int> samples{};    // Can be used to indicate that sequence files belong to the same experiment
+    bool paired = false;           // If true, than experiments are seen as paired-end experiments
     bool experiment_names = false; // Flag, if names of experiment should be stored in a txt file
     bool ram_friendly = false;
 };
 
 //!\brief Generates a random integer not greater than a given maximum
-struct RandomGenerator {
-	int maxi;
-	RandomGenerator(int max) :
-			maxi(max) {
-	}
-
-	int operator()() {
-		return rand() % maxi;
-	}
+struct RandomGenerator
+{
+    int maxi;
+    RandomGenerator(int max) : maxi(max)
+    {}
+
+    int operator()()
+    {
+        return rand() % maxi;
+    }
 };
 
 /*!\brief Get the concrete expression values (= median of all counts of one transcript) for given experiments.
@@ -48,8 +51,11 @@ struct RandomGenerator {
 * \param genome_file        A "*.genome" file constructed with the command genome.
 * \param paired             Flag to indicate if input data is paired or not.
 */
-void count(min_arguments const & args, std::vector<std::filesystem::path> sequence_files, std::filesystem::path include_file,
-           std::filesystem::path genome_file, bool paired);
+void count(min_arguments const & args,
+           std::vector<std::filesystem::path> sequence_files,
+           std::filesystem::path include_file,
+           std::filesystem::path genome_file,
+           bool paired);
 
 /*!\brief Creates a set of minimizers to ignore, which should be used as an input to count.
 * \param args               The minimiser arguments to use (seed, shape, window size).
@@ -71,7 +77,10 @@ void read_binary(std::filesystem::path filename, robin_hood::unordered_node_map<
 * \param num_of_minimisers  Variable, where to number of minimisers should be stored.
 * \param cutoff             cutoff value.
 */
-void read_binary_start(min_arguments & args, std::filesystem::path filename, uint64_t & num_of_minimisers, uint8_t & cutoff);
+void read_binary_start(min_arguments & args,
+                       std::filesystem::path filename,
+                       uint64_t & num_of_minimisers,
+                       uint8_t & cutoff);
 
 /*! \brief Creates IBFs.
  * \param sequence_files  A vector of sequence file paths.
@@ -85,8 +94,11 @@ void read_binary_start(min_arguments & args, std::filesystem::path filename, uin
  * \param num_hash        The number of hash functions to use.
  *  \returns The expression thresholds per experiment.
  */
-std::vector<uint16_t> ibf(std::vector<std::filesystem::path> const & sequence_files, estimate_ibf_arguments & ibf_args,
-                          minimiser_arguments & minimiser_args, std::vector<double> & fpr, std::vector<uint8_t> & cutoffs,
+std::vector<uint16_t> ibf(std::vector<std::filesystem::path> const & sequence_files,
+                          estimate_ibf_arguments & ibf_args,
+                          minimiser_arguments & minimiser_args,
+                          std::vector<double> & fpr,
+                          std::vector<uint8_t> & cutoffs,
                           std::filesystem::path const expression_by_genome_file = "",
                           size_t num_hash = 1);
 
@@ -101,7 +113,8 @@ std::vector<uint16_t> ibf(std::vector<std::filesystem::path> const & sequence_fi
  *  \returns The expression thresholds per experiment.
  */
 std::vector<uint16_t> ibf(std::vector<std::filesystem::path> const & minimiser_files,
-                          estimate_ibf_arguments & ibf_args, std::vector<double> & fpr,
+                          estimate_ibf_arguments & ibf_args,
+                          std::vector<double> & fpr,
                           std::filesystem::path const expression_by_genome_file = "",
                           size_t num_hash = 1);
 
@@ -111,8 +124,10 @@ std::vector<uint16_t> ibf(std::vector<std::filesystem::path> const & minimiser_f
 * \param minimiser_args  The minimiser specific arguments to use.
 * \param cutoffs         List of cutoffs.
 */
-void minimiser(std::vector<std::filesystem::path> const & sequence_files, min_arguments const & args,
-               minimiser_arguments & minimiser_args, std::vector<uint8_t> & cutoffs);
+void minimiser(std::vector<std::filesystem::path> const & sequence_files,
+               min_arguments const & args,
+               minimiser_arguments & minimiser_args,
+               std::vector<uint8_t> & cutoffs);
 
 /*! \brief Insert into IBFs.
 * \param sequence_files  A vector of sequence file paths.
@@ -127,9 +142,12 @@ void minimiser(std::vector<std::filesystem::path> const & sequence_files, min_ar
 *  \returns The expression thresholds per experiment.
 */
 std::vector<uint16_t> insert(std::vector<std::filesystem::path> const & sequence_files,
-                             estimate_ibf_arguments & ibf_args, minimiser_arguments & minimiser_args,
+                             estimate_ibf_arguments & ibf_args,
+                             minimiser_arguments & minimiser_args,
                              std::vector<uint8_t> & cutoffs,
-                             std::filesystem::path const expression_by_genome_file, std::filesystem::path path_in, bool samplewise);
+                             std::filesystem::path const expression_by_genome_file,
+                             std::filesystem::path path_in,
+                             bool samplewise);
 
 /*! \brief Insert into IBFs based on the minimiser files
 * \param minimiser_files A vector of minimiser file paths.
@@ -143,7 +161,9 @@ std::vector<uint16_t> insert(std::vector<std::filesystem::path> const & sequence
 */
 std::vector<uint16_t> insert(std::vector<std::filesystem::path> const & minimiser_files,
                              estimate_ibf_arguments & ibf_args,
-                             std::filesystem::path const expression_by_genome_file, std::filesystem::path path_in, bool samplewise);
+                             std::filesystem::path const expression_by_genome_file,
+                             std::filesystem::path path_in,
+                             bool samplewise);
 
 /*! \brief Delete bins from ibfs
 * \param delete_files    A vector of integers specifiying the bins to delete.
@@ -152,4 +172,7 @@ std::vector<uint16_t> insert(std::vector<std::filesystem::path> const & minimise
 * \param path_in         Input directory.
 * \param samplewise      True, if expression levels were set beforehand.
 */
-void delete_bin(std::vector<uint64_t> const & delete_files, estimate_ibf_arguments & ibf_args, std::filesystem::path path_in, bool samplewise);
+void delete_bin(std::vector<uint64_t> const & delete_files,
+                estimate_ibf_arguments & ibf_args,
+                std::filesystem::path path_in,
+                bool samplewise);
diff --git a/include/shared.hpp b/include/shared.hpp
@@ -15,7 +15,8 @@
 #include <seqan3/search/kmer_index/shape.hpp>
 #include <seqan3/search/views/minimiser_hash.hpp>
 
-inline constexpr static uint64_t adjust_seed(uint8_t const kmer_size, uint64_t const seed = 0x8F3F73B5CF1C9ADEULL) noexcept
+static inline constexpr uint64_t adjust_seed(uint8_t const kmer_size,
+                                             uint64_t const seed = 0x8F'3F'73'B5'CF'1C'9A'DEULL) noexcept
 {
     return seed >> (64u - 2u * kmer_size);
 }
@@ -31,7 +32,7 @@ struct all_arguments
 struct min_arguments : all_arguments
 {
     uint8_t k{20};
-    seqan3::seed s{0x8F3F73B5CF1C9ADEULL};
+    seqan3::seed s{0x8F'3F'73'B5'CF'1C'9A'DEULL};
     seqan3::shape shape = seqan3::ungapped{k};
     seqan3::window_size w_size{60};
 };
@@ -41,10 +42,10 @@ struct estimate_ibf_arguments : min_arguments
 {
     bool compressed = false;
     std::vector<uint16_t> expression_thresholds{}; // Expression levels which should be created
-    uint8_t number_expression_thresholds{}; // If set, the expression levels are determined by the program.
+    uint8_t number_expression_thresholds{};        // If set, the expression levels are determined by the program.
     bool samplewise{false};
 
-    template<class Archive>
+    template <class Archive>
     void save(Archive & archive) const
     {
         archive(k);
@@ -57,7 +58,7 @@ struct estimate_ibf_arguments : min_arguments
         archive(samplewise);
     }
 
-    template<class Archive>
+    template <class Archive>
     void load(Archive & archive)
     {
         archive(k);
@@ -119,8 +120,7 @@ void load_ibf(IBFType & ibf, std::filesystem::path ipath)
  *  \param opath Path, where the IBF should be stored.
  */
 template <class IBFType>
-void store_ibf(IBFType const & ibf,
-               std::filesystem::path opath)
+void store_ibf(IBFType const & ibf, std::filesystem::path opath)
 {
     std::ofstream os{opath, std::ios::binary};
     cereal::BinaryOutputArchive oarchive{os};

diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
@@ -1,10 +1,10 @@
 cmake_minimum_required (VERSION 3.25)
 
-find_package(OpenMP REQUIRED COMPONENTS CXX)
-add_library ("${PROJECT_NAME}_lib" STATIC ibf.cpp estimate.cpp)
+find_package (OpenMP REQUIRED COMPONENTS CXX)
+add_library ("${PROJECT_NAME}_lib" STATIC estimate.cpp ibf.cpp)
 target_link_libraries ("${PROJECT_NAME}_lib" PUBLIC seqan3::seqan3)
 target_link_libraries ("${PROJECT_NAME}_lib" PUBLIC robin_hood::robin_hood)
-target_link_libraries("${PROJECT_NAME}_lib" PUBLIC OpenMP::OpenMP_CXX)
+target_link_libraries ("${PROJECT_NAME}_lib" PUBLIC OpenMP::OpenMP_CXX)
 target_include_directories ("${PROJECT_NAME}_lib" PUBLIC ../include)
 
 add_executable ("${PROJECT_NAME}" main.cpp)