Clean metadata from Jupyter notebooks

nbclean

#!/bin/bash

function usage {
    echo "Usage: $(basename "$0") [OPTION]... FILE" 2>&1
    echo "   -i     modify the file in place"
    echo "   -o     don't strip the outputs"
    echo "   -h     display this help and exit"
    exit 0
}

INPLACE=false
TMP="$(mktemp)"
# shellcheck disable=SC2064
trap "rm -rf $TMP" EXIT

# shellcheck disable=SC2220
while getopts ":hio" flag; do
    case "$flag" in
        i) INPLACE=true;;
        o) DONTSTRIP="yes";;
        h) usage;;
    esac
done

for (( i=0; i<OPTIND-1; i++ )); do
    shift
done
FILE="$1"

if [[ ! -f "$FILE" ]]; then
    usage
fi

jq --indent 1 --arg dontstrip "$DONTSTRIP" \
'.
| del(.cells[] | select(.source == []))
| if $dontstrip == "yes" then
    .
  else
    (.cells[] | select(has("outputs")) | .outputs) = []
  end
| (.cells[] | select(has("execution_count")) | .execution_count) = null
| .cells[].metadata = {}
| .metadata = {}
' \
"$FILE" > "$TMP"

if $INPLACE; then
    mv -f "$TMP" "$FILE"
else
    cat "$TMP"
fi