scripts/install.sh

#!/usr/bin/env bash
#
# Installs dereko2vec and everything it needs, in one go:
#
#   1. the packages of the distribution (dnf, apt-get or brew)
#   2. rocksdb, but only where the distribution has none or none with the
#      static library (scripts/build-rocksdb.sh)
#   3. collocatordb
#   4. dereko2vec, with the fast static link of rocksdb and collocatordb
#
# Works on Fedora, Rocky Linux, RHEL, Debian, Ubuntu and MacOS.
#
# Options:
#   --shared               link rocksdb and collocatordb shared (default:
#                          static, which is about 22% faster when indexing)
#   --prefix DIR           install there instead of /usr/local
#   --rocksdb-prefix DIR   build rocksdb into DIR instead of $HOME/rocksdb
#   --rocksdb-version V    build rocksdb version V instead of the one of the
#                          distribution
#   --skip-deps            do not install the packages of the distribution
#   --skip-tests           do not run ctest
#   --build-only           build everything but install nothing: no packages of
#                          the distribution, no root needed, the binaries stay
#                          in the build directory
#   --static-binary        build dereko2vec as a completely static binary
#                          (needs static compression libraries, which Debian
#                          and Ubuntu ship and Fedora, Rocky and RHEL do not)
#   -j N                   build with N jobs (default: all cores)
#
set -euo pipefail

DEREKO2VEC_SRC=$(cd "$(dirname "$0")/.." && pwd)
PREFIX=/usr/local
ROCKSDB_PREFIX=$HOME/rocksdb
ROCKSDB_VERSION=
STATIC=1
STATIC_BINARY=0
SKIP_DEPS=0
SKIP_TESTS=0
BUILD_ONLY=0
JOBS=

usage() {
    sed -n '2,26p' "$0"
    exit "${1:-0}"
}

while [ $# -gt 0 ]; do
    case $1 in
        --shared)           STATIC=0; shift ;;
        --static-binary)    STATIC_BINARY=1; shift ;;
        --prefix)           PREFIX=$2; shift 2 ;;
        --rocksdb-prefix)   ROCKSDB_PREFIX=$2; shift 2 ;;
        --rocksdb-version)  ROCKSDB_VERSION=$2; shift 2 ;;
        --skip-deps)        SKIP_DEPS=1; shift ;;
        --skip-tests)       SKIP_TESTS=1; shift ;;
        --build-only)       BUILD_ONLY=1; shift ;;
        -j)                 JOBS=$2; shift 2 ;;
        -h|--help)          usage 0 ;;
        *) echo "Unknown option: $1" >&2; usage 1 ;;
    esac
done

# Nothing is installed, so nothing of the distribution is needed either
if [ $BUILD_ONLY = 1 ]; then
    SKIP_DEPS=1
fi

if [ -z "$JOBS" ]; then
    JOBS=$(getconf _NPROCESSORS_ONLN 2> /dev/null || sysctl -n hw.ncpu)
fi
export CMAKE_BUILD_PARALLEL_LEVEL=$JOBS

step() { printf '\n=== %s ===\n' "$*"; }

# Root is needed only where the target is not writable
NEED_ROOT=
if ! [ -w "$PREFIX" ] && ! { [ ! -e "$PREFIX" ] && [ -w "$(dirname "$PREFIX")" ]; }; then
    NEED_ROOT=1
fi

# Runs a command as root. Prefer sudo, fall back to su, which prompts for the
# root password - a server may well have only that.
root() {
    if [ -z "$NEED_ROOT" ] || [ "$(id -u)" = 0 ]; then
        "$@"
    elif command -v sudo > /dev/null; then
        sudo "$@" || su -c "$*"
    elif command -v su > /dev/null; then
        su -c "$*"
    else
        echo "Installing to $PREFIX needs root, but there is neither sudo nor su." >&2
        echo "Use --prefix with a directory you can write to." >&2
        return 1
    fi
}

# The rocksdb and the compression libraries of homebrew are below
# /opt/homebrew on Apple Silicon, which is not searched by default
BREW_PREFIX=
if command -v brew > /dev/null; then
    BREW_PREFIX=$(brew --prefix)
fi

# --- 1. the packages of the distribution -------------------------------------
if [ $SKIP_DEPS = 0 ]; then
    step "Installing the packages of the distribution"
    if command -v dnf > /dev/null; then
        root dnf install -y cmake gcc-c++ git make pkgconf-pkg-config \
             snappy-devel zlib-devel bzip2-devel lz4-devel libzstd-devel
        # Rocky Linux and RHEL have no rocksdb-devel everywhere, that is what
        # step 2 is for
        root dnf install -y rocksdb-devel || true
    elif command -v apt-get > /dev/null; then
        root apt-get update
        root apt-get install -y cmake g++ git make pkg-config \
             librocksdb-dev libsnappy-dev zlib1g-dev libbz2-dev liblz4-dev libzstd-dev
    elif command -v brew > /dev/null; then
        brew install cmake rocksdb snappy zlib bzip2 lz4 zstd
    else
        echo "No dnf, apt-get or brew - install cmake, a C++ compiler, git, rocksdb" >&2
        echo "and the devel packages of snappy, zlib, bzip2, lz4 and zstd by hand." >&2
    fi
fi

# --- 2. rocksdb --------------------------------------------------------------
static_rocksdb_of_the_distribution() {
    local f
    for f in /usr/lib64/librocksdb.a /usr/lib/librocksdb.a /usr/lib/*/librocksdb.a; do
        [ -e "$f" ] && return 0
    done
    if command -v brew > /dev/null && brew list --versions rocksdb > /dev/null 2>&1; then
        [ -e "$(brew --prefix rocksdb)/lib/librocksdb.a" ] && return 0
    fi
    return 1
}

DISTRO_ROCKSDB=
if { command -v rpm > /dev/null && rpm -q rocksdb > /dev/null 2>&1; } ||
   { command -v dpkg-query > /dev/null && dpkg-query -W librocksdb-dev > /dev/null 2>&1; } ||
   { command -v brew > /dev/null && brew list --versions rocksdb > /dev/null 2>&1; }; then
    DISTRO_ROCKSDB=1
fi

ROCKSDB_PREFIX_USED=
ROCKSDB_SHARED_BUILT=0
if [ -n "$DISTRO_ROCKSDB" ] && [ -z "$ROCKSDB_VERSION" ]; then
    if [ $STATIC = 1 ] && ! static_rocksdb_of_the_distribution; then
        step "rocksdb: building the static library, same version as the distribution"
        "$DEREKO2VEC_SRC/scripts/build-rocksdb.sh" --static-only -p "$ROCKSDB_PREFIX"
        ROCKSDB_PREFIX_USED=$ROCKSDB_PREFIX
    else
        step "rocksdb: the one of the distribution is used"
    fi
else
    if [ -z "$DISTRO_ROCKSDB" ]; then
        step "rocksdb: no rocksdb package - building the shared and the static library"
    else
        step "rocksdb: building version $ROCKSDB_VERSION as asked"
    fi
    "$DEREKO2VEC_SRC/scripts/build-rocksdb.sh" -p "$ROCKSDB_PREFIX" ${ROCKSDB_VERSION:+-v "$ROCKSDB_VERSION"}
    ROCKSDB_PREFIX_USED=$ROCKSDB_PREFIX
    ROCKSDB_SHARED_BUILT=1
fi

# Programs that link the shared rocksdb of the prefix have to find it. Only a
# system-wide install can register the directory; for one below $HOME the
# programs carry the rpath of the libraries they were linked against.
if [ $BUILD_ONLY = 0 ] && [ $ROCKSDB_SHARED_BUILT = 1 ] && [ "$(uname)" = Linux ] && [ -n "$NEED_ROOT" ]; then
    RDB_LIBDIR=$(dirname "$(find "$ROCKSDB_PREFIX_USED" -name 'librocksdb.so' | head -1)")
    echo "$RDB_LIBDIR" | root tee /etc/ld.so.conf.d/rocksdb-prefix.conf > /dev/null
    root ldconfig
fi

# --- 3. collocatordb ---------------------------------------------------------
step "collocatordb"
WORK=$(mktemp -d)
trap 'rc=$?; if [ $rc = 0 ]; then rm -rf "$WORK"; else echo "Keeping $WORK for a look at what went wrong" >&2; fi' EXIT
git clone "https://korap.ids-mannheim.de/gerrit/ids-kl/collocatordb" "$WORK/collocatordb"
CDB_PREFIXES=$ROCKSDB_PREFIX_USED
if [ -n "$BREW_PREFIX" ]; then
    CDB_PREFIXES=${CDB_PREFIXES:+$CDB_PREFIXES;}$BREW_PREFIX
fi
cmake -S "$WORK/collocatordb" -B "$WORK/collocatordb/build" \
      -DCMAKE_INSTALL_PREFIX="$PREFIX" \
      ${CDB_PREFIXES:+-DCMAKE_PREFIX_PATH=$CDB_PREFIXES}
cmake --build "$WORK/collocatordb/build" -j "$JOBS"
CDB_STAGE=
if [ $BUILD_ONLY = 1 ]; then
    # install into a staging prefix only, where the dereko2vec build below
    # finds the static library and the header
    CDB_STAGE=$WORK/stage
    cmake --install "$WORK/collocatordb/build" --prefix "$CDB_STAGE"
else
    root cmake --install "$WORK/collocatordb/build"
    if [ "$(uname)" = Linux ] && [ -n "$NEED_ROOT" ]; then
        root ldconfig
    fi
fi
if [ $SKIP_TESTS = 0 ]; then
    ctest --test-dir "$WORK/collocatordb/build" --output-on-failure
fi

# --- 4. dereko2vec -----------------------------------------------------------
step "dereko2vec"
# A stale build directory would silently keep the collocatordb and rocksdb it
# was once configured with, and an install manifest of a root install breaks a
# later install without sudo. Start over, the build takes a minute.
rm -rf "$DEREKO2VEC_SRC/build"
CMAKE_PREFIXES=$PREFIX
if [ -n "$CDB_STAGE" ]; then
    CMAKE_PREFIXES="$CDB_STAGE;$CMAKE_PREFIXES"
fi
if [ -n "$ROCKSDB_PREFIX_USED" ]; then
    CMAKE_PREFIXES="$CMAKE_PREFIXES;$ROCKSDB_PREFIX_USED"
fi
if [ -n "$BREW_PREFIX" ]; then
    CMAKE_PREFIXES="$CMAKE_PREFIXES;$BREW_PREFIX"
fi
STATIC_FLAG=
if [ $STATIC_BINARY = 1 ]; then
    STATIC_FLAG=-DSTATIC_DEREKO2VEC=ON
elif [ $STATIC = 1 ]; then
    STATIC_FLAG=-DSTATIC_ROCKSDB=ON
fi
cmake -S "$DEREKO2VEC_SRC" -B "$DEREKO2VEC_SRC/build" \
      -DCMAKE_INSTALL_PREFIX="$PREFIX" \
      -DCMAKE_PREFIX_PATH="$CMAKE_PREFIXES" \
      $STATIC_FLAG
cmake --build "$DEREKO2VEC_SRC/build" -j "$JOBS"
if [ $SKIP_TESTS = 0 ]; then
    ctest --test-dir "$DEREKO2VEC_SRC/build" --output-on-failure
fi
if [ $BUILD_ONLY = 0 ]; then
    root cmake --install "$DEREKO2VEC_SRC/build"
fi

step "Done"
if [ $BUILD_ONLY = 1 ]; then
    echo "dereko2vec and vecs2mmap are in $DEREKO2VEC_SRC/build"
    echo "Nothing was installed. Install them with:"
    echo "  sudo cmake --install \"$DEREKO2VEC_SRC/build\""
    if [ $STATIC = 0 ] && [ $ROCKSDB_SHARED_BUILT = 1 ]; then
        RDB_LIBDIR=$(dirname "$(find "$ROCKSDB_PREFIX_USED" -name 'librocksdb.so' | head -1)")
        echo "With --shared the programs need the rocksdb library at runtime, e.g."
        echo "  LD_LIBRARY_PATH=$RDB_LIBDIR $DEREKO2VEC_SRC/build/dereko2vec"
    fi
else
    echo "dereko2vec and vecs2mmap are in $PREFIX/bin"
fi