Files
docscan/run.sh
Thomas Ruoff 85162c511b inital commit
2021-01-27 01:09:21 +01:00

47 lines
902 B
Bash
Executable File

#!/usr/bin/env bash
set -eu
shopt -s globstar nullglob
OCRMYPDF_CMD='docker run --rm -i ocrmypdf'
SRC_DIR=./test
DST_DIR=./dst
FAILED_DIR=$DST_DIR/failed
LOG_DIR=$DST_DIR/logs
ARCHIVE_DIR=$DST_DIR/archive
mkdir -p "$DST_DIR"
mkdir -p "$ARCHIVE_DIR"
function ocrFile() {
SRC=$1
DST_DIR=$2
FILE_NAME=$(basename -- "$SRC")
DST_PDF=$DST_DIR/${FILE_NAME%.*}.pdf
DST_LOG=$LOG_DIR/${FILE_NAME%.*}.log
$OCRMYPDF_CMD \
-l deu \
--clean \
--rotate-pages \
--deskew \
- - \
< "$FILE" \
> "$DST_PDF" 2> "$DST_LOG"
}
for FILE in "$SRC_DIR"/*.pdf
do
echo "processing $FILE..."
ocrFile "$FILE" "$DST_DIR"
# TODO: move to FAILED_DIR and error out
echo processed "$FILE"
mv "$FILE" "$ARCHIVE_DIR"
echo moved "$FILE" to archive
done