#! /bin/bash # "scan2pdf.sh" # A script to scan one or multiple pages and convert them to pdf # ----------------------------------------------------------------- # Copyright (c) 2004-2023 Joerg Hau # # This program is free software; you can redistribute it and/or # modify it under the terms of version 2 of the GNU General Public # License as published by the Free Software Foundation. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # ----------------------------------------------------------------- # Revisions: # 2004-xx-xx, first operational version (JHa) # 2004-12-10, cmd line parsing and A4/Letter options (JHa) # 2005-03-02, Zip compression (instead of "Fax") (JHa) # 2005-03-18, fix in checkfor() error message (JHa) # 2005-03-30, added resolution on cmd line (JHa) # 2005-10-04, added option for color and grayscale (JHa) # 2005-11-30, added option for multi-page scanning (JHa) # 2006-11-28, grayscale now default (JHa) # 2006-12-28, workaround for broken "page" instruction # in some versions of Imagemagick (JHa) # 2009-10-31, re-introduced "page" instruction, this fixes # also the page size of the final PDF (JHa) # 2009-11-07, added rotation (JHa) # 2010-09-05, added pdf2ps && ps2pdf compression step (JHa) # 2012-10-31, added possibility to wait (for editing slides) # 2015-03-03, some changes to accomodate for EPSON GT-1500 (JHa) # 2017-01-30, check for existing output file; improved contrast; # use of ADF; changes for Fujitsu SP-1120 (JHa) # 2017-02-08, select scanner type via cmd line; added OCR (JHa) # 2017-02-12, SP-1120 now defaults to multi-page scan (JHa) # 2017-02-13, added -sPAPERSIZE=a4 -dPDFFitPage (JHa) # 2017-09-02, bugfix in button-wait for standard flatbed scanners (JHa) # 2017-09-03, parallel processing works (JHa) # 2018-06-19, chgd GT1500 switch from -E to -G, added Epson 1660 Photo, # adressing two different scanners OK (JHa) # 2018-07-15, fixed the tesseract page size issue (JHa) # 2018-08-31, added scanning over the network (JHa) # 2018-09-15, not using pdfopt anymore due to some unidentified bug (JHa) # 2018-10-01, minor fixes caused by the removal of pdfopt (JHa) # 2019-02-23, switch for contrast option (JHa) # 2019-03-28, switch for tesseract options (JHa) # 2020-01-30, corrected last message, added explicit device (JHa) # 2020-02-05, added export OMP_THREAD_LIMIT=1 to limit CPU ressource # conflicts on processors that support AVX2 # 2020-03-16, added LC_ALL to get rid of error messages in Debian 10 KDE # 2020-04-05, refined checkfor() error message # 2021-11-05, modified defaults # 2022-02-09, now uses temp files (and deletes these upon error) # 2022-02-16, option to launch PDF viewer after scanning # 2022-07-08, added "set -e" in case some command does not work as intended # 2022-08-09, all echo commands now go to stderr # 2022-08-25, revised settings to allow 1-page scanning on Fujitsu (explicitly # specify "-p 1"). Added some time measurement for OCR process. # 2023-04-15, added info for 1-page scanning on Fujitsu # 2024-01-15, changed order of commands to avoid "scanimage: rounded value # of br-y from 380 to 297" errors for Custom page size # 2024-01-17, corrected some errors I instzriduced with the last change ;-) # ----------------------------------------------------------------- # ----------------------------------------------------------------- # Idea: options could be queried with scanimage -h but this is slow # ----------------------------------------------------------------- # Debian 10+11 have problems with LC... so we set them explicitly # export LC_ALL=C # default paper size and resolution. # Note that some scanners support only a fixed set of resolutions # SIZE="A4" RES="225" # Scan modes. The exact syntax depends a bit on the scanner used, # e.g. you may need to replace "binary" (Epson) by "lineart" (HP) # ... just run 'scanimage --h' to find out ;-) # MODE_BW="Binary" MODE_GRAY="Gray" MODE_COL="Color" # default number of pages # PAGES="1" # page number passed via command line? Default no = empty # PAGES_CMD="" # command for button press. Again, the exact syntax depends on the # scanner used, check also the "--batch-prompt" ;-) # default is "no" for 1 page # BUTTON_CMD="--wait-for-button=no" # pause for editing? default no # PAUSE="" # default mode is color # MODE=$MODE_COL # Default language for OCR # use the codes from tesseract, i.e. "eng", "fra", "deu" # empty = no OCR is performed # OCRLANG="fra" # compression algorithm; PDF supports embedded zip compression # COMPR="-compress Zip" # by default, scan in portrait = do not rotate # ROT="" # by default, scan from flatbed (otherwise "--source ADF" or the like) # SRC="" # contrast. You may play with 15%,85%; for faint text leave this out (option "-C") # CONTRAST="-level 3%,97%" # any other tesseract options that may be useful # -c debug_file=/dev/null invokes quiet mode: https://stackoverflow.com/questions/31806648/tesseract-quiet-mode # TESSOPT="-c debug_file=/dev/null" # PDF viewer # VIEWER="atril" # -------------------------------------------------------------- # subroutine to check for some required executables # argument: (list of) programs to test for # will exit with rc=1 if any program was not found # -------------------------------------------------------------- function checkfor() { for i in $*; do PROG=`which $i` if [ $? != 0 ] ; then echo "Problem: '$i' command not found, please install it." >&2 exit 1 fi done } checkfor scanimage convert pdftk pdf2ps ps2pdf rm parallel mktemp ${VIEWER} # ------------------------------------------------------------ # subroutine to print usage mode # ------------------------------------------------------------ function usage() { cat >&2 << eof ${0##*/} - scan to PDF A script to scan one or more pages and convert them to PDF. Copyright (c) 2004...2023 Joerg Hau . This program is free software; you can redistribute it and/or modify it under the terms of version 2 of the GNU General Public License as published by the Free Software Foundation. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. Usage: ${0##*/} [options] path/to/outfile[.pdf] Options: -4, --A4 use A4 paper size. -l, --letter use Letter paper size. -s, --other use 'other' paper size (as specified inside this script) -p, --pages x scan x pages and combine them (with ADF, use big numbers like "199"). -r resolution use this resolution. -b, --bw scan in black/white. -g, --gray scan in grayscale -c, --color scan in color (default). -C, --contrast do NOT use contrast adjustment (use this for faint documents) -L, --Left rotate 90 degrees. -R, --Right rotate 270 degrees. -e, --edit pause to edit pages before compressing -o, --ocr lang run OCR with language 'lang' (deu, fra, eng / default = fra) -E, --Epson use Epson 1660 Photo scanner -X, --ENet use Epson 1660 Photo scanner over network -H, --HP use HP all-in-one scanner -G, --GT1500 use ADF of Epson GT1500 scanner -F, --SP1120 use ADF of Fujitsu SP1120 document scanner -n, --nomulti do NOT use SP1120 multifeed detection (implies -F) -v, --view launch PDF viewer after scanning -? | -h | --help this help ;-) Invoking "${0##*/}" with no options will - initiate a single-page scan in $MODE at $RES dpi in $SIZE size, - perform OCR in $OCRLANG, - and convert the resulting file to compressed PDF. eof } # -------------------------------------------------------------- # parse & handle command line parameters (flags) # -------------------------------------------------------------- while true; do case $1 in -4 | --A4 ) SIZE="A4" ;; -l | --letter ) SIZE="letter" ;; -s | --sizeOther ) SIZE="other" ;; -p | --pages ) shift PAGES="$1" PAGES_CMD="1" # any non-empty string will do ;; -r | --resolution ) shift RES="$1" ;; -b | --bw ) MODE=$MODE_BW ;; -g | --gray ) MODE=$MODE_GRAY ;; -c | --color) MODE=$MODE_COL ;; -C | --contrast) CONTRAST="" ;; -L | --Left) ROT="-rotate 270" ;; -R | --Right) ROT="-rotate 90" ;; -e | --edit) PAUSE="pause" ;; -o | --ocr ) shift OCRLANG="$1" ;; -E | --Epson ) SCANNER="EPSON" SRC="" MULTIFEED="" ;; -X | --ENet ) SCANNER="EPSON-NET" SRC="" MULTIFEED="" ;; -G | --GT1500 ) SCANNER="GT1500" SRC="ADF" MULTIFEED="" ;; -F | --SP1120 ) SCANNER="SP1120" SRC="ADF" MULTIFEED="Stop" ;; -n | --nomulti ) SCANNER="SP1120" SRC="ADF" MULTIFEED="Do-not-detect" ;; -H | --HP ) SCANNER="HP" SRC="" MULTIFEED="" ;; -v | --view ) DO_VIEW="1" # any non-empty string will do ;; -h | -? | --help ) usage exit 2 ;; *) break ;; esac shift done # -------------------------------------------------------------- # if OCR is requested, check for tesseract and ghostscript # ------------------------------------------------------------- if [ -n $OCRLANG ] ; then checkfor tesseract gs fi # -------------------------------------------------------------- # if view is requested, check for viewer # ------------------------------------------------------------- if [ -n $DO_VIEW ] ; then checkfor $VIEWER fi # -------------------------------------------------------------- # if script is called w/o remaining arguments, give instructions # -------------------------------------------------------------- if [ -z $1 ] ; then usage exit 1 fi # -------------------------------------------------------------- # if the output file exists, do not ask but exit # -------------------------------------------------------------- if [ -e ${1%.*}.pdf ] ; then echo "File ${1%.*}.pdf exists! Exiting." >&2 exit 2 fi # -------------------------------------------------------------- # assign scan size from paper size. # "other"can be anything that you assign here # -------------------------------------------------------------- case $SIZE in "A4" ) X="210" Y="297" ;; "letter" ) X="215" Y="280" ;; "other" ) X="165" Y="240" ;; esac # -------------------------------------------------------------- # Assign other variables depending on scanner and cmd line # The syntax varies between scanners! # Note: we could actually run all scanners over the network ;-) # -------------------------------------------------------------- case $SCANNER in "GT1500" ) RES_CMD="--x-resolution ${RES}dpi --y-resolution ${RES}dpi" ADF_CMD="--source Automatic" MULTIFEED_CMD="" BUTTON_CMD="--wait-for-button=no" ;; "SP1120" ) RES_CMD="--resolution ${RES}dpi --paper-size Custom --paper-size Custom --page-width $X --page-height $Y" # if you want to scan single-side only, use '--source Adf-front' ADF_CMD="-d pfusp --source Adf-duplex --blank-page-skip=yes --blank-page-skip-sensitivity 5 --page-auto=yes --autofeed=yes " MULTIFEED_CMD="--multifeed-detection ${MULTIFEED}" BUTTON_CMD="" # if number of pages was NOT passed on command line, set default to 300 for ADF if [ -z "$PAGES_CMD" ] ; then PAGES="300" ; fi ;; "HP" ) RES_CMD="--resolution ${RES}dpi" ADF_CMD="" MULTIFEED_CMD="" BUTTON_CMD="" if [ "$Y" == "297" ] ; then Y="296.9" ; fi # many HP cannot scan full A4 size ;; "EPSON" ) RES_CMD="--resolution ${RES}dpi" ADF_CMD="-d epkowa " # we're re-using that variable to assign the device MULTIFEED_CMD="" BUTTON_CMD="--wait-for-button=yes" if [ "$PAGES" == "1" ] ; then BUTTON_CMD="" ; fi ;; "EPSON-NET" ) RES_CMD="--resolution ${RES}dpi" ADF_CMD="-d net:192.168.11.99:epkowa " # we're re-using that variable to assign the device MULTIFEED_CMD="" BUTTON_CMD="--wait-for-button=yes" if [ "$PAGES" == "1" ] ; then BUTTON_CMD="" ; fi ;; esac echo "Scanning ${PAGES} page(s) of ${SIZE} paper ($Y) in ${MODE} at ${RES} dpi, output file is '${1%.*}.pdf'.">&2 if [ -n "$SRC" ] ; then echo "Using Auto Document Feeder (ADF), button disabled." >&2; fi if [ -n "$ROT" ] ; then echo "Image will be rotated." >&2; fi if [ -n "$PAUSE" ] ; then echo "Will pause for editing before compression." >&2; fi if [ -n "${OCRLANG// }" ] ; then echo "OCR in $OCRLANG requested." >&2; fi if (( $PAGES > 1 )) ; then # this is multi-page mode if [ -n "$BUTTON_CMD" ] ; then # non-ADF scan echo >&2 echo "=== Put first page on scanner and press the scan button when you're ready. ===">&2 echo >&2 fi fi # Create a temporary filename # OUT=$(mktemp "$(basename $0).XXXXXXXXXX") # make sure that these files are deleted later on # trap 'rm ${OUT}*' EXIT # If you use a non-standard paper size, "$ADF_CMD --paper-size Custom" # must come BEFORE you specify the page size. # scanimage $ADF_CMD $RES_CMD -x $X -y $Y --mode $MODE $MULTIFEED_CMD \ --batch=${OUT}%03d.pnm --batch-count=$PAGES $BUTTON_CMD # If problem => exit // this does not work when the scanner runs out of pages # # if [ $? != 0 ]; then # echo "*** Some problem occurred while scanning. Abort - status: $?" # exit 1 # fi # Note: In ADF mode, we cannot use errorlevel checking at this point. # this bails out as soon as the scanner runs out of paper ... # exit if some command does not work as intended set -e if [ -n "$PAUSE" ] ; then read -p "Scanning finished, press [Enter] to continue: " >&2; fi && \ # If you copy and edit the pnm files here, you can use the following commands # to convert them to a single pdf afterwards: # # for i in out*.pnm; do convert -page A4 -compress Zip $i ${i%.*}.pdf && rm $i; done # pdftk out*.pdf cat output xx.pdf # pdfopt xx.pdf yy.pdf # rm out*.pdf xx.pdf # pdf2ps yy.pdf yy.ps && ps2pdf yy.ps yy.pdf # rm yy.ps # mv yy.pdf /path/to/file.pdf # measure conversion start time start_time=$(date +%s.%2N) echo -n "Conversion to pdf in progress ... " >&2 && \ if [ -n "${OCRLANG// }" ] ; then # Convert to PNG and run tesseract. # # set OMP_THREAD_LIMIT=1 to limit CPU ressource conflicts on processors that support AVX2 # see https://github.com/the-paperless-project/paperless/issues/438 # export OMP_THREAD_LIMIT=1 # Processing through tesseract will cause the page size to be wrong (too big), # so we need to add the page size with -density and -units before running tesseract: # see https://github.com/tesseract-ocr/tesseract/issues/150 # ls -1 ${OUT}*.pnm | parallel convert $CONTRAST $COMPR -density ${RES}x${RES} -units PixelsPerInch $ROT '{}' '{.}.png' && \ ls -1 ${OUT}*.png | parallel tesseract $TESSOPT -l $OCRLANG '{}' '{.}' pdf && \ rm ${OUT}*.png ${OUT}*.pnm else # just convert to PDF ls -1 ${OUT}*.pnm | parallel convert $CONTRAST $COMPR $ROT -page $SIZE '{}' '{.}.pdf' && \ rm ${OUT}*.pnm fi # if we have more than 1 page, invoke pdftk # if (( $PAGES > 1 )) ; then pdftk ${OUT}*.pdf cat output ${1%.*}.tmp.pdf && \ mv ${1%.*}.tmp.pdf ${1%.*}.pdf && \ rm ${OUT}*.pdf ## ${1%.*}.tmp.pdf else mv ${OUT}001.pdf ${1%.*}.pdf fi echo -n "done. " >&2 # measure conversion end time end_time=$(date +%s.%2N) echo "This took $(echo "scale=2; $end_time - $start_time" | bc) seconds." >&2 echo -n "Compressing file '${1%.*}.pdf' " >&2 && \ if [ -n "${OCRLANG// }" ] ; then echo -n "using Ghostscript ... " >&2 # If OCR was requested, we need to rewrite carefully gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dPDFSETTINGS=/default \ -dNOPAUSE -dQUIET -dBATCH -dDetectDuplicateImages -dNOTRANSPARENCY \ -dCompressFonts=true -sOutputFile=${OUT}.rewrite.pdf ${1%.*}.pdf && \ mv ${OUT}.rewrite.pdf ${1%.*}.pdf else echo -n "... " >&2 # Very efficient in terms of filesize, but destroys all OCR information: pdf2ps ${1%.*}.pdf ${OUT}.ps && \ ps2pdf ${OUT}.ps ${1%.*}.pdf && \ rm ${OUT}.ps fi echo "done.">&2 if [ -n $DO_VIEW ] ; then $VIEWER ${1%.*}.pdf & echo "done.">&2 fi # exit with error status of last cmd # STATUS=$? exit $STATUS