Since I work as a freelance, I have to deal with a lot of pdf files I receive from clients. Some of them contain uncompressed images that make the files bigger than they need to be.
On this stackexchange question, a user named Marco gives us a nice way to use ghostscript to compress any pdf file.
There is a loss of quality in the compression, but most of the time it is acceptable.
Here is the script:
#!/bin/sh
INPUTPDF=$1
OUTPUTPDF=$2
TMPPDF=$(mktemp)
METADATA=$(mktemp)
# save metadata
pdftk "$INPUTPDF" dump_data_utf8 > "$METADATA"
# compress
gs \
-q \
-sOutputFile="$TMPPDF" \
-sDEVICE=pdfwrite \
-dNOPAUSE \
-dBATCH \
-dPDFSETTINGS=/ebook \
"$INPUTPDF"
# restore metadata
pdftk "$TMPPDF" update_info_utf8 "$METADATA" output "$OUTPUTPDF"
# clean up
rm -f "$TMPPDF" "$METADATA"