cryptonas/scripts/userdocexport.sh

116 lines
4.6 KiB
Bash
Executable File

#!/bin/sh
#
# export wiki pages to the cryptobox development tree
# this creates static and integrated pages
#
set -ue
# root directory of the cryptobox development environment
ROOT_DIR="$(dirname $(dirname $0))"
# retrieve these pages from the wiki
PAGES="CryptoBox CryptoBoxIntro CryptoBoxDev CryptoBoxDevPreparation CryptoBoxDevCustomBuild CryptoBoxDevWorkFlow CryptoBoxDevValidation CryptoBoxDevCustomConfigure CryptoBoxDevBackground CryptoBoxDevKnownProblems CryptoBoxUser CryptoBoxUserGettingStarted CryptoBoxUserConfiguration CryptoBoxUserDailyUse"
# base URL
WIKI_HOST="https://systemausfall.org"
# the trailing slash is important
WIKI_URL=/toolforge/
CBOX_CGI="/cryptobox?action=doc\&page="
LANGUAGES="en de"
DEST_DIR="$ROOT_DIR/cbox-tree.d/usr/share/doc/cryptobox/html"
OFFLINE_DIR="$ROOT_DIR/cbox-tree.d/_offline/doc"
IMAGE_DIR="$ROOT_DIR/cbox-tree.d/var/www/cryptobox-img"
TMP_DIR=/tmp/$(basename $0)-$$.d
HEADER_FILE=doc_header.inc
FOOTER_FILE=doc_footer.inc
[ ! -e "$DEST_DIR" ] && echo "$DEST_DIR does not exist" && exit 1
for LANG in $LANGUAGES; do
for PAGE in $PAGES; do
PAGE_SRC="$WIKI_HOST$WIKI_URL$PAGE/$LANG"
echo "Importing $PAGE/$LANG:"
# replace sub-page-style '/' like moin does it (by '_2f')
TMP_FILE=$TMP_DIR/${PAGE}.html
mkdir -p "$TMP_DIR"
echo " downloading the page ..."
wget --quiet --output-document="$TMP_FILE" "$PAGE_SRC" || { echo "Downloading ($PAGE_SRC) failed!"; exit 1; }
# check if this moin page exists (by looking for the template selection)
if grep -q "$WIKI_URL$PAGE/$LANG.action=edit&template=" "$TMP_FILE"
then rm "$TMP_FILE"
PAGE_SRC=$(dirname $PAGE_SRC)
echo " trying to download default language page instead"
wget --quiet --output-document="$TMP_FILE" "$PAGE_SRC" || { echo "Downloading ($PAGE_SRC) failed!" >&2; exit 1; }
# check, if there is even no default page
grep -q "$WIKI_URL$PAGE/$LANG.action=edit&template=" "$TMP_FILE" && echo "This page ($PAGE_SRC) was not found!" >&2 && exit 1
fi
echo " removing header and footer ..."
# break lines before start of content
sed -i 's#<div [^>]* id="content" [^>]*>#_END_OF_HEADER_\n#' "$TMP_FILE"
sed -i 's#<a id="bottom">#\n_START_OF_FOOTER_#' "$TMP_FILE"
# remove all lines before and after "body"
sed -i '1,/_END_OF_HEADER_/d; /_START_OF_FOOTER_/,$d' "$TMP_FILE"
echo " removing link images (moin specific) ..."
# remove inter-wiki images
sed -i 's#<[^<]*moin-inter.png[^>]*>##g' "$TMP_FILE"
# remove moin-www images
sed -i 's#<[^<]*moin-www.png[^>]*> ##g' "$TMP_FILE"
echo " downloading requisites ..."
wget --quiet --ignore-tags=a --no-clobber --page-requisites --convert-links --no-parent --no-directories --base="$WIKI_HOST$WIKI_URL" --directory-prefix="$TMP_DIR" --html-extension --force-html --input-file="$TMP_FILE" || { echo "Downloading requisites for ($PAGE_SRC) failed!"; exit 1; }
echo " adjusting links for requisites ..."
find "$TMP_DIR" -type f | grep -v '\.html$' | while read a
do fname=$(basename "$a")
sed -i "s#=\"[^\"]*/$fname\"#=\"/cryptobox-img/$fname\"#g" "$TMP_FILE"
done
echo " adjusting wiki links ..."
# redirect wiki links to cryptobox cgi
sed -i "s#=\"$WIKI_URL\([^\.]*\)\"#=\"$CBOX_CGI\1\"#g" "$TMP_FILE"
# do it twice - somehow, the "g" flag does not work (it should replace multiple occurrences on a line)
sed -i "s#=\"$WIKI_URL\([^\.]*\)\"#=\"$CBOX_CGI\1\"#g" "$TMP_FILE"
# remove language specific part of moin link
for TLANG in $LANGUAGES
do sed -i "s#=\"$CBOX_CGI\(.*\)/$TLANG\"#=\"$CBOX_CGI\1\"#g" "$TMP_FILE"
done
# build the static pages
echo " building static doc page"
offline_file=$OFFLINE_DIR/$LANG/$(basename $TMP_FILE)
mkdir -p "$OFFLINE_DIR/$LANG"
cat "$OFFLINE_DIR/$HEADER_FILE" "$OFFLINE_DIR/$LANG/$HEADER_FILE" "$TMP_FILE" "$OFFLINE_DIR/$LANG/$FOOTER_FILE" "$OFFLINE_DIR/$FOOTER_FILE" >"$offline_file"
sed -i "s#=\"$CBOX_CGI\([^\"]*\)\"#=\"\1.html\"#g" "$offline_file"
# do it twice - this should not be necessary
sed -i "s#=\"$CBOX_CGI\([^\"]*\)\"#=\"\1.html\"#g" "$offline_file"
sed -i "s#=\"/cryptobox-img#=\"../../../var/www/cryptobox-img#g" "$offline_file"
# split language specific part of moin link and replace it by current language
for TLANG in $LANGUAGES
do sed -i "s#=\"\([^/]*\)/${TLANG}.html\"#=\"\1.html\"#g" "$offline_file"
done
# some last changes to the dynamic pages (must be done _after_ the static pages)
# add weblang for current language to query string
sed -i "s#=\"$CBOX_CGI\(.*\)\"#=\"$CBOX_CGI\1\&weblang=$LANG\"#g" "$TMP_FILE"
# move cgi-doc
mv "$TMP_FILE" "$DEST_DIR/$LANG"
echo " finished!"
done
done
mv "$TMP_DIR"/* "$IMAGE_DIR"
rmdir "$TMP_DIR"