On device offline dictionary for reMarkable

updateDefinitionsFromHighlights.sh

#!/bin/bash
#
# opkg requirements :
# opkg install 7z git jq
#
# git clone https://github.com/wordset/wordset-dictionary/
#
# assumes ~/xochitl-data links to /home/root/.local/share/remarkable/xochitl/
#
# warning ~/epubTemplate/ is required. See instead https://gist.github.com/Utopiah/119f5c96fae048609a2091f7f3d81f53
# It's the most basic ePub one can use so for now use yours.

addEpubWithMetadata(){
    cp "$1" ~/xochitl-data && echo "{'parent':'','type':'DocumentType','visibleName':'$1'}" | sed s/\'/\"/g > ~/xochitl-data/`echo $1 | sed "s/.epub//"`.metadata;
    echo '{ "fileType": "epub" }' > ~/xochitl-data/`echo $1 | sed "s/.epub//"`.content;
}

getLatestRemarkableHighlights(){
        cd ~/xochitl-data
        grep text THE_FUTURE_OF_TEXT_2020.highlights/* | sed 's/.*: \"\(.*\)\"/\1/' # hardcoded for testing        
        #grep text $(ls -rt | grep -v mynewwords | tail -1 | sed "s/\..*//").highlights/* | sed 's/.*: \"\(.*\)\"/\1/'
}

getDefinitions(){
    [ -f /tmp/newdefinitions ] && rm /tmp/newdefinitions
        cd
        for W in $(getLatestRemarkableHighlights); do
        [[ $(echo $W | wc -w) -ge 2 ]] && continue
                AW=$(echo $W|sed "s/[^[:alpha:]]//g")
        # could remove stopwords e.g a an the ... with a for sed s///i but making sure there is a trailing space for articles
        grep $AW: /tmp/definitionscache && continue
        DEF=$(jq -r .$AW.meanings[0].def /home/root/wordset-dictionary-master/data/$(echo $AW|sed "s/^\(.\).*/\1/").json)
        echo "<p><b>$AW:</b> $DEF</p>" >> /tmp/newdefinitions
    done
    [ -f /tmp/newdefinitions ] && cat /tmp/newdefinitions >> /tmp/definitionscache
}

generateNewEpub(){
    cd
    rm -rf newWords
    cp -r epubTemplate/ newWords
    head -6 newWords/OEBPS/content.xhtml > newWords/OEBPS/newcontent.xhtml
    cat /tmp/definitionscache >> newWords/OEBPS/newcontent.xhtml
    # these lookups can take up to 2s per word so better do some caching or try in RAM.
    tail -2 newWords/OEBPS/content.xhtml >> newWords/OEBPS/newcontent.xhtml
    mv newWords/OEBPS/newcontent.xhtml newWords/OEBPS/content.xhtml
    cd newWords/
    7z a -r my.epub mimetype META-INF/ OEBPS/
    mv my.epub ~/mynewwords.epub

    if [ -f /home/root/.local/share/remarkable/xochitl/mynewwords.epub ]; then
        rm /home/root/.local/share/remarkable/xochitl/mynewwords.pdf #cache busting
        cp ~/mynewwords.epub /home/root/.local/share/remarkable/xochitl/mynewwords.epub
    else
        cd
        addEpubWithMetadata mynewwords.epub
        systemctl restart xochitl
    fi
}

# if there is no .lock file then xochitl is not active
# one can then check if there are new highlights since the last epub generation
# if not then just give up.

getDefinitions
[ -f /tmp/newdefinitions ] && generateNewEpub

Hi,
Sound amazing, thanks.
do you have a tutor of how to install?
Best

Hi, thanks for the interest. It's an old, more than a year old, prototype so I can't say for sure if it still works. In fact due to reMarkable OS v3 not yet being supported by Toltec (details on toltec-dev.org ) some of the dependencies might be tricky to get. Anyway it runs via ssh on the device so the first step would be connecting this way (information on the device itself for root password) then running the script assuming everything is setup correctly (directories, dependencies, etc) as briefly mentioned in the comments at the beginning of the file. Let me know if that's sufficient.