1
0
mirror of https://github.com/searx/searx synced 2024-11-29 21:50:11 +01:00
searx/manage
Pierre Chevalier 3a0f896b68 [enh] Add Springer Nature engine
Springer Nature is a global publisher dedicated to providing service to research
community [1] with official API [2].

To test this PR, first get your API key following this page:

   https://dev.springernature.com/signup

In searx/engines/springer.py at line 24, add this API key.  I left my own key,
commented out in the line aboce.  Feel free to use it, if needed.

[1] https://www.springernature.com/
[2] https://dev.springernature.com/
2021-04-29 22:43:52 +02:00

504 lines
15 KiB
Bash
Executable File

#!/usr/bin/env bash
# -*- coding: utf-8; mode: sh indent-tabs-mode: nil -*-
# SPDX-License-Identifier: AGPL-3.0-or-later
# shellcheck disable=SC2031
# shellcheck source=utils/lib.sh
source "$(dirname "${BASH_SOURCE[0]}")/utils/lib.sh"
# shellcheck source=utils/brand.env
source "${REPO_ROOT}/utils/brand.env"
source_dot_config
# config
PY_SETUP_EXTRAS='[test]'
NPM_PACKAGES="less@2.7 less-plugin-clean-css grunt-cli"
GECKODRIVER_VERSION="v0.28.0"
# SPHINXOPTS=
# These py files are linted by test.pylint(), all other files are linted by
# test.pep8()
PYLINT_FILES=(
searx/preferences.py
searx/testing.py
searx/engines/gigablast.py
searx/engines/deviantart.py
searx/engines/digg.py
searx/engines/google.py
searx/engines/google_news.py
searx/engines/google_videos.py
searx/engines/google_images.py
searx/engines/mediathekviewweb.py
searx/engines/meilisearch.py
searx/engines/solidtorrents.py
searx/engines/solr.py
searx/engines/springer.py
searx/engines/google_scholar.py
searx/engines/yahoo_news.py
searx/engines/apkmirror.py
searx/engines/core.py
searx_extra/update/update_external_bangs.py
)
PYLINT_SEARX_DISABLE_OPTION="\
I,C,R,\
W0105,W0212,W0511,W0603,W0613,W0621,W0702,W0703,W1401,\
E1136"
PYLINT_ADDITIONAL_BUILTINS_FOR_ENGINES="supported_languages,language_aliases"
PYLINT_OPTIONS="-m pylint -j 0 --rcfile .pylintrc"
help() {
cat <<EOF
buildenv
rebuild ./utils/brand.env
babel.compile
pybabel compile ./searx/translations
data.*
all : update searx/languages.py and ./data/*
languages : update searx/data/engines_languages.json & searx/languages.py
useragents: update searx/data/useragents.json with the most recent versions of Firefox.
docs.*
html : build HTML documentation
live : autobuild HTML documentation while editing
gh-pages : deploy on gh-pages branch
prebuild : build reST include files (./${DOCS_BUILD}/includes)
clean : clean documentation build
docker
build : build docker image
push : build and push docker image
gecko.driver
download & install geckodriver if not already installed (required for
robot_tests)
node.*
env : download & install npm dependencies locally
clean : drop npm installations
py.*
build : Build python packages at ./${PYDIST}
clean : delete virtualenv and intermediate py files
pyenv.* :
install : developer install of searx into virtualenv
uninstall : uninstall developer installation
cmd ... : run command ... in virtualenv
OK : test if virtualenv is OK
pypi.upload:
Upload python packages to PyPi (to test use pypi.upload.test)
test.* :
pylint : lint PYLINT_FILES, searx/engines, searx & tests
pep8 : pycodestyle (pep8) for all files except PYLINT_FILES
unit : run unit tests
coverage : run unit tests with coverage
robot : run robot test
clean : clean intermediate test stuff
themes.* :
all : build all themes
oscar : build oscar theme
simple : build simple theme
EOF
}
if [ "$VERBOSE" = "1" ]; then
SPHINX_VERBOSE="-v"
PYLINT_VERBOSE="-v"
fi
# needed by sphinx-docs
export DOCS_BUILD
buildenv() {
SEARX_DEBUG=1 pyenv.cmd python utils/build_env.py 2>&1 \
| prefix_stdout "${_Blue}BUILDENV${_creset} "
return "${PIPESTATUS[0]}"
}
babel.compile() {
build_msg BABEL compile
pyenv.cmd pybabel compile -d "${REPO_ROOT}/searx/translations"
dump_return $?
}
data.all() {
data.languages
data.useragents
build_msg DATA "update searx/data/ahmia_blacklist.txt"
pyenv.cmd python searx_extra/update/update_ahmia_blacklist.py
build_msg DATA "update searx/data/wikidata_units.json"
pyenv.cmd python searx_extra/update/update_wikidata_units.py
build_msg DATA "update searx/data/currencies.json"
pyenv.cmd python searx_extra/update/update_currencies.py
}
data.languages() {
( set -e
build_msg ENGINES "fetch languages .."
pyenv.cmd python searx_extra/update/update_languages.py
build_msg ENGINES "update update searx/languages.py"
build_msg DATA "update searx/data/engines_languages.json"
)
dump_return $?
}
data.useragents() {
build_msg DATA "update searx/data/useragents.json"
pyenv.cmd python searx_extra/update/update_firefox_version.py
dump_return $?
}
docs.prebuild() {
build_msg DOCS "build ${DOCS_BUILD}/includes"
(
set -e
[ "$VERBOSE" = "1" ] && set -x
mkdir -p "${DOCS_BUILD}/includes"
./utils/searx.sh doc | cat > "${DOCS_BUILD}/includes/searx.rst"
./utils/filtron.sh doc | cat > "${DOCS_BUILD}/includes/filtron.rst"
./utils/morty.sh doc | cat > "${DOCS_BUILD}/includes/morty.rst"
)
dump_return $?
}
docker.push() {
docker.build push
}
# shellcheck disable=SC2119
docker.build() {
pyenv.install
build_msg DOCKER build
# run installation in a subprocess and activate pyenv
# See https://www.shellcheck.net/wiki/SC1001 and others ..
# shellcheck disable=SC2031,SC2230,SC2002,SC2236,SC2143,SC1001
( set -e
# shellcheck source=/dev/null
source "${PY_ENV_BIN}/activate"
# Check if it is a git repository
if [ ! -d .git ]; then
die 1 "This is not Git repository"
fi
if [ ! -x "$(which git)" ]; then
die 1 "git is not installed"
fi
if ! git remote get-url origin 2> /dev/null; then
die 1 "there is no remote origin"
fi
# This is a git repository
# "git describe" to get the Docker version (for example : v0.15.0-89-g0585788e)
# awk to remove the "v" and the "g"
SEARX_GIT_VERSION=$(git describe --match "v[0-9]*\.[0-9]*\.[0-9]*" HEAD 2>/dev/null | awk -F'-' '{OFS="-"; $1=substr($1, 2); if ($3) { $3=substr($3, 2); } print}')
# add the suffix "-dirty" if the repository has uncommited change
# /!\ HACK for searx/searx: ignore utils/brand.env
git update-index -q --refresh
if [ ! -z "$(git diff-index --name-only HEAD -- | grep -v 'utils/brand.env')" ]; then
SEARX_GIT_VERSION="${SEARX_GIT_VERSION}-dirty"
fi
# Get the last git commit id, will be added to the Searx version (see Dockerfile)
VERSION_GITCOMMIT=$(echo "$SEARX_GIT_VERSION" | cut -d- -f2-4)
build_msg DOCKER "Last commit : $VERSION_GITCOMMIT"
# Check consistency between the git tag and the searx/version.py file
# /! HACK : parse Python file with bash /!
# otherwise it is not possible build the docker image without all Python
# dependencies ( version.py loads __init__.py )
# SEARX_PYTHON_VERSION=$(python3 -c "import six; import searx.version; six.print_(searx.version.VERSION_STRING)")
SEARX_PYTHON_VERSION=$(cat searx/version.py | grep "\(VERSION_MAJOR\|VERSION_MINOR\|VERSION_BUILD\) =" | cut -d\= -f2 | sed -e 's/^[[:space:]]*//' | paste -sd "." -)
if [ "$(echo "$SEARX_GIT_VERSION" | cut -d- -f1)" != "$SEARX_PYTHON_VERSION" ]; then
err_msg "git tag: $SEARX_GIT_VERSION"
err_msg "searx/version.py: $SEARX_PYTHON_VERSION"
die 1 "Inconsistency between the last git tag and the searx/version.py file"
fi
# define the docker image name
GITHUB_USER=$(echo "${GIT_URL}" | sed 's/.*github\.com\/\([^\/]*\).*/\1/')
SEARX_IMAGE_NAME="${SEARX_IMAGE_NAME:-${GITHUB_USER:-searx}/searx}"
# build Docker image
build_msg DOCKER "Building image ${SEARX_IMAGE_NAME}:${SEARX_GIT_VERSION}"
sudo docker build \
--build-arg GIT_URL="${GIT_URL}" \
--build-arg SEARX_GIT_VERSION="${SEARX_GIT_VERSION}" \
--build-arg VERSION_GITCOMMIT="${VERSION_GITCOMMIT}" \
--build-arg LABEL_DATE="$(date -u +"%Y-%m-%dT%H:%M:%SZ")" \
--build-arg LABEL_VCS_REF="$(git rev-parse HEAD)" \
--build-arg LABEL_VCS_URL="${GIT_URL}" \
--build-arg TIMESTAMP_SETTINGS="$(git log -1 --format="%cd" --date=unix -- searx/settings.yml)" \
--build-arg TIMESTAMP_UWSGI="$(git log -1 --format="%cd" --date=unix -- dockerfiles/uwsgi.ini)" \
-t "${SEARX_IMAGE_NAME}:latest" -t "${SEARX_IMAGE_NAME}:${SEARX_GIT_VERSION}" .
if [ "$1" = "push" ]; then
sudo docker push "${SEARX_IMAGE_NAME}:latest"
sudo docker push "${SEARX_IMAGE_NAME}:${SEARX_GIT_VERSION}"
fi
)
dump_return $?
}
# shellcheck disable=SC2119
gecko.driver() {
pyenv.install
build_msg INSTALL "gecko.driver"
# run installation in a subprocess and activate pyenv
( set -e
# shellcheck source=/dev/null
source "${PY_ENV_BIN}/activate"
# TODO : check the current geckodriver version
geckodriver -V > /dev/null 2>&1 || NOTFOUND=1
set +e
if [ -z "$NOTFOUND" ]; then
build_msg INSTALL "geckodriver already installed"
return
fi
PLATFORM="$(python3 -c 'import platform; print(platform.system().lower(), platform.architecture()[0])')"
case "$PLATFORM" in
"linux 32bit" | "linux2 32bit") ARCH="linux32";;
"linux 64bit" | "linux2 64bit") ARCH="linux64";;
"windows 32 bit") ARCH="win32";;
"windows 64 bit") ARCH="win64";;
"mac 64bit") ARCH="macos";;
esac
GECKODRIVER_URL="https://github.com/mozilla/geckodriver/releases/download/$GECKODRIVER_VERSION/geckodriver-$GECKODRIVER_VERSION-$ARCH.tar.gz";
build_msg GECKO "Installing ${PY_ENV_BIN}/geckodriver from $GECKODRIVER_URL"
FILE="$(mktemp)"
wget -qO "$FILE" -- "$GECKODRIVER_URL" && tar xz -C "${PY_ENV_BIN}" -f "$FILE" geckodriver
rm -- "$FILE"
chmod 755 -- "${PY_ENV_BIN}/geckodriver"
)
dump_return $?
}
node.env() {
local err=0
pushd "${REPO_ROOT}" &> /dev/null
# shellcheck disable=SC2230
which npm &> /dev/null || die 1 'node.env - npm is not found!'
( set -e
# shellcheck disable=SC2030
PATH="$(npm bin):$PATH"
export PATH
build_msg INSTALL "npm install $NPM_PACKAGES"
# shellcheck disable=SC2086
npm install $NPM_PACKAGES
cd "${REPO_ROOT}/searx/static/themes/oscar"
build_msg INSTALL "($(pwd)) npm install"
npm install
build_msg INSTALL "($(pwd)) npm install"
cd "${REPO_ROOT}/searx/static/themes/simple"
npm install
)
err=$?
popd &> /dev/null
dump_return "$err"
}
node.clean() {
build_msg CLEAN "locally installed npm dependencies"
rm -rf \
./node_modules \
./package-lock.json \
./searx/static/themes/oscar/package-lock.json \
./searx/static/themes/oscar/node_modules \
./searx/static/themes/simple/package-lock.json \
./searx/static/themes/simple/node_modules
dump_return $?
}
py.build() {
build_msg BUILD "[pylint] python package ${PYDIST}"
pyenv.cmd python setup.py \
sdist -d "${PYDIST}" \
bdist_wheel --bdist-dir "${PYBUILD}" -d "${PYDIST}"
}
py.clean() {
build_msg CLEAN pyenv
( set -e
pyenv.drop
[ "$VERBOSE" = "1" ] && set -x
rm -rf "${PYDIST}" "${PYBUILD}" "${PY_ENV}" ./.tox ./*.egg-info
find . -name '*.pyc' -exec rm -f {} +
find . -name '*.pyo' -exec rm -f {} +
find . -name __pycache__ -exec rm -rf {} +
)
}
pyenv.check() {
cat <<EOF
import yaml
print('import yaml --> OK')
EOF
}
pyenv.install() {
if ! pyenv.OK; then
py.clean > /dev/null
fi
if pyenv.install.OK > /dev/null; then
return 0
fi
( set -e
pyenv
build_msg PYENV "[install] pip install -e 'searx${PY_SETUP_EXTRAS}'"
"${PY_ENV_BIN}/python" -m pip install -e ".${PY_SETUP_EXTRAS}"
buildenv
) || die 42 "error while build & install pyenv (${PY_ENV_BIN})"
}
pyenv.uninstall() {
build_msg PYENV "[pyenv.uninstall] uninstall packages: ${PYOBJECTS}"
pyenv.cmd python setup.py develop --uninstall 2>&1 \
| prefix_stdout "${_Blue}PYENV ${_creset}[pyenv.uninstall] "
}
pypi.upload() {
py.clean
py.build
# https://github.com/pypa/twine
pyenv.cmd twine upload "${PYDIST}"/*
}
pypi.upload.test() {
py.clean
py.build
pyenv.cmd twine upload -r testpypi "${PYDIST}"/*
}
test.pylint() {
# shellcheck disable=SC2086
( set -e
build_msg TEST "[pylint] \$PYLINT_FILES"
pyenv.cmd python ${PYLINT_OPTIONS} ${PYLINT_VERBOSE} \
"${PYLINT_FILES[@]}"
build_msg TEST "[pylint] searx/engines"
pyenv.cmd python ${PYLINT_OPTIONS} ${PYLINT_VERBOSE} \
--disable="${PYLINT_SEARX_DISABLE_OPTION}" \
--additional-builtins="${PYLINT_ADDITIONAL_BUILTINS_FOR_ENGINES}" \
searx/engines
build_msg TEST "[pylint] searx tests"
pyenv.cmd python ${PYLINT_OPTIONS} ${PYLINT_VERBOSE} \
--disable="${PYLINT_SEARX_DISABLE_OPTION}" \
--ignore=searx/engines \
searx tests
)
dump_return $?
}
test.pep8() {
build_msg TEST 'pycodestyle (formerly pep8)'
local _exclude=""
printf -v _exclude '%s, ' "${PYLINT_FILES[@]}"
pyenv.cmd pycodestyle \
--exclude="searx/static, searx/languages.py, $_exclude " \
--max-line-length=120 \
--ignore "E117,E252,E402,E722,E741,W503,W504,W605" \
searx tests
dump_return $?
}
test.unit() {
build_msg TEST 'tests/unit'
pyenv.cmd python -m nose2 -s tests/unit
dump_return $?
}
test.coverage() {
build_msg TEST 'unit test coverage'
( set -e
pyenv.cmd python -m nose2 -C --log-capture --with-coverage --coverage searx -s tests/unit
pyenv.cmd coverage report
pyenv.cmd coverage html
)
dump_return $?
}
test.robot() {
build_msg TEST 'robot'
gecko.driver
PYTHONPATH=. pyenv.cmd python searx/testing.py robot
dump_return $?
}
test.clean() {
build_msg CLEAN "test stuff"
rm -rf geckodriver.log .coverage coverage/
dump_return $?
}
themes.all() {
( set -e
node.env
themes.oscar
themes.simple
)
dump_return $?
}
themes.oscar() {
local gruntfile=searx/static/themes/oscar/gruntfile.js
build_msg GRUNT "${gruntfile}"
PATH="$(npm bin):$PATH" grunt --gruntfile "${gruntfile}"
dump_return $?
}
themes.simple() {
local gruntfile=searx/static/themes/simple/gruntfile.js
build_msg GRUNT "${gruntfile}"
PATH="$(npm bin):$PATH" grunt --gruntfile "${gruntfile}"
dump_return $?
}
# shellcheck disable=SC2119
main() {
local _type
local cmd="$1"; shift
if [ "$cmd" == "" ]; then
help
err_msg "missing command"
return 42
fi
case "$cmd" in
--getenv) var="$1"; echo "${!var}";;
--help) help;;
--*)
help
err_msg "unknown option $cmd"
return 42
;;
*)
_type="$(type -t "$cmd")"
if [ "$_type" != 'function' ]; then
err_msg "unknown command $1 / use --help"
return 42
else
"$cmd" "$@"
fi
;;
esac
}
main "$@"