Merge commit '1b6617188f37d94d127222bbd1364265d54d25ee'

lib/mmseqs/CMakeLists.txt

@@ -16,6 +16,7 @@ set(HAVE_POWER9 0 CACHE BOOL "Have POWER9 CPU")
 set(HAVE_POWER8 0 CACHE BOOL "Have POWER8 CPU")
 set(HAVE_ARM8 0 CACHE BOOL "Have ARMv8 CPU")
 set(NATIVE_ARCH 1 CACHE BOOL "Assume native architecture for SIMD. Use one of the HAVE_* options or set CMAKE_CXX_FLAGS to the appropriate flags if you disable this.")
+set(USE_SYSTEM_ZSTD 0 CACHE BOOL "Use zstd provided by system instead of bundled version")
 
 if (HAVE_SANITIZER)
     include(FindUBSan)
@@ -81,6 +82,8 @@ if (NATIVE_ARCH AND (MMSEQS_ARCH STREQUAL ""))
         set(X64 1)
     elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "x86|X86")
         set(X86 1)
+    elseif(CMAKE_SYSTEM_PROCESSOR MATCHES "^sparc")
+        set(SPARC 1)
     else ()
         message(WARNING "CPU without native SIMD instructions. Performance will be bad.")
     endif ()
@@ -151,7 +154,7 @@ if ((CMAKE_CXX_COMPILER_ID STREQUAL "Clang" AND CMAKE_CXX_COMPILER_VERSION VERSI
     set(DISABLE_IPS4O 1)
 endif ()
 
-if (PPC64)
+if (PPC64 OR SPARC)
     # FIXME: investigate why on ppc the regression seems to fail randomly
     set(DISABLE_IPS4O 1)
 endif ()
@@ -165,23 +168,31 @@ if (CMAKE_COMPILER_IS_CLANG AND (NOT EMSCRIPTEN))
     set(MMSEQS_CXX_FLAGS "${MMSEQS_CXX_FLAGS} -stdlib=libc++")
 endif ()
 
-
-# zstd
-# We use ZSTD_findDecompressedSize which is only available with ZSTD_STATIC_LINKING_ONLY
-# Thus we cannot use a system provided libzstd
-set(ZSTD_SOURCE_DIR "${CMAKE_CURRENT_SOURCE_DIR}/lib/zstd")
-set(CMAKE_INSTALL_LIBDIR bin)
-set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_CURRENT_SOURCE_DIR}/lib/zstd/build/cmake/CMakeModules")
-option(ZSTD_LEGACY_SUPPORT "LEGACY SUPPORT" OFF)
-option(ZSTD_BUILD_STATIC "BUILD STATIC LIBRARIES" ON)
-option(ZSTD_BUILD_SHARED "BUILD SHARED LIBRARIES" OFF)
-option(ZSTD_MULTITHREAD_SUPPORT "MULTITHREADING SUPPORT" OFF)
-option(ZSTD_BUILD_PROGRAMS "BUILD PROGRAMS" OFF)
-option(ZSTD_BUILD_CONTRIB "BUILD CONTRIB" OFF)
-option(ZSTD_BUILD_TESTS "BUILD TESTS" OFF)
-include_directories(lib/zstd/lib)
-add_subdirectory(lib/zstd/build/cmake/lib EXCLUDE_FROM_ALL)
-set_target_properties(libzstd_static PROPERTIES COMPILE_FLAGS "${MMSEQS_C_FLAGS}" LINK_FLAGS "${MMSEQS_C_FLAGS}")
+if (USE_SYSTEM_ZSTD)
+    include(FindPackageHandleStandardArgs)
+    find_path(ZSTD_INCLUDE_DIRS NAMES zstd.h REQUIRED)
+    # We use ZSTD_findDecompressedSize which is only available with ZSTD_STATIC_LINKING_ONLY
+    find_library(ZSTD_LIBRARIES NAMES libzstd.a libzstd_static REQUIRED)
+    find_package_handle_standard_args(ZSTD DEFAULT_MSG ZSTD_LIBRARIES ZSTD_INCLUDE_DIRS)
+    mark_as_advanced(ZSTD_LIBRARIES ZSTD_INCLUDE_DIRS)
+    include_directories(${ZSTD_INCLUDE_DIRS})
+else ()
+    # We use ZSTD_findDecompressedSize which is only available with ZSTD_STATIC_LINKING_ONLY
+    set(ZSTD_SOURCE_DIR "${CMAKE_CURRENT_SOURCE_DIR}/lib/zstd")
+    set(CMAKE_INSTALL_LIBDIR bin)
+    set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_CURRENT_SOURCE_DIR}/lib/zstd/build/cmake/CMakeModules")
+    option(ZSTD_LEGACY_SUPPORT "LEGACY SUPPORT" OFF)
+    option(ZSTD_BUILD_STATIC "BUILD STATIC LIBRARIES" ON)
+    option(ZSTD_BUILD_SHARED "BUILD SHARED LIBRARIES" OFF)
+    option(ZSTD_MULTITHREAD_SUPPORT "MULTITHREADING SUPPORT" OFF)
+    option(ZSTD_BUILD_PROGRAMS "BUILD PROGRAMS" OFF)
+    option(ZSTD_BUILD_CONTRIB "BUILD CONTRIB" OFF)
+    option(ZSTD_BUILD_TESTS "BUILD TESTS" OFF)
+    include_directories(lib/zstd/lib)
+    add_subdirectory(lib/zstd/build/cmake/lib EXCLUDE_FROM_ALL)
+    set_target_properties(libzstd_static PROPERTIES COMPILE_FLAGS "${MMSEQS_C_FLAGS}" LINK_FLAGS "${MMSEQS_C_FLAGS}")
+    set(ZSTD_LIBRARIES libzstd_static)
+endif()
 
 # tinyexpr
 include_directories(lib/tinyexpr)

lib/mmseqs/README.md

@@ -9,6 +9,8 @@ MMseqs2 (Many-against-Many sequence searching) is a software suite to search and
 
 [Mirdita M, Steinegger M and Soeding J. MMseqs2 desktop and local web server app for fast, interactive sequence searches. Bioinformatics, doi: 10.1093/bioinformatics/bty1057 (2019)](https://academic.oup.com/bioinformatics/article/35/16/2856/5280135).
 
+[Mirdita M, Steinegger M, Breitwieser F, Soding J, Levy Karin E: Fast and sensitive taxonomic assignment to metagenomic contigs. bioRxiv, doi: 10.1101/2020.11.27.401018 (2020)](https://www.biorxiv.org/content/10.1101/2020.11.27.401018v1).
+
 [![BioConda Install](https://img.shields.io/conda/dn/bioconda/mmseqs2.svg?style=flag&label=BioConda%20install)](https://anaconda.org/bioconda/mmseqs2)
 [![Github All Releases](https://img.shields.io/github/downloads/soedinglab/mmseqs2/total.svg)](https://github.com/soedinglab/mmseqs2/releases/latest)
 [![Biocontainer Pulls](https://img.shields.io/endpoint?url=https%3A%2F%2Fmmseqs.com%2Fbiocontainer.php%3Fcontainer%3Dmmseqs2)](https://biocontainers.pro/#/tools/mmseqs2)

lib/mmseqs/azure-pipelines.yml

@@ -7,19 +7,29 @@ variables:
     regression: 1
 
 jobs:
-  - job: build_ubuntu_1804_userguide
-    displayName: Ubuntu 1804 Userguide
+  - job: build_ubuntu_2004_userguide
+    displayName: Ubuntu 2004 Userguide
     pool:
-      vmImage: 'Ubuntu-18.04'
+      vmImage: 'Ubuntu-20.04'
     steps:
       - checkout: "none"
+      - task: Cache@2
+        inputs:
+          key: '"tectonic" | "$(Agent.OS)"'
+          restoreKeys: |
+            "tectonic" | "$(Agent.OS)"
+            "tectonic"
+          path: $(Pipeline.Workspace)/tectonic-cache/
+        displayName: Cache Tectonic
       - script: |
-          sudo apt-get update
-          sudo apt-get -y install pandoc texlive-latex-recommended texlive-fonts-extra
+          wget -qO- https://github.com/tectonic-typesetting/tectonic/releases/download/tectonic%400.4.1/tectonic-0.4.1-x86_64-unknown-linux-gnu.tar.gz | tar xzvf - tectonic
+          wget -qO- https://github.com/jgm/pandoc/releases/download/2.11.3.2/pandoc-2.11.3.2-linux-amd64.tar.gz | tar --strip-components=2 -xzvf - pandoc-2.11.3.2/bin/pandoc
+          sudo mv -f pandoc tectonic /usr/local/bin
         displayName: Install Dependencies
       - script: |
           cd ${SYSTEM_DEFAULTWORKINGDIRECTORY}
           git clone https://github.com/soedinglab/MMseqs2.wiki.git .
+          export XDG_CACHE_HOME=${PIPELINE_WORKSPACE}/tectonic-cache/
           .pandoc/make-pdf.sh
         displayName: Build Userguide
       - task: PublishPipelineArtifact@0
@@ -242,7 +252,7 @@ jobs:
     pool:
       vmImage: 'Ubuntu-18.04'
     dependsOn:
-      - build_ubuntu_1804_userguide
+      - build_ubuntu_2004_userguide
       - build_macos_1015
       - build_ubuntu_1804
       - build_ubuntu_cross_2004
@@ -251,7 +261,7 @@ jobs:
       - script: |
           cd "${BUILD_SOURCESDIRECTORY}"
           mkdir mmseqs
-          cp -f README.md LICENCE.md mmseqs
+          cp -f README.md LICENSE.md mmseqs
           cp -r examples mmseqs
           mkdir mmseqs/matrices
           cp -f data/*.out mmseqs/matrices