PDF grepping.

pdf_grep.sh

#!/bin/bash

###########################################################################################
# This script is used to comb a pdf file for a specific phrase. It is bad, inefficient
#  script, but I haven't been able to figure out how to properly do it yet (and I've been
#  looking for several months - combing S.O. and freshmeat/freecode for proper index/search
#  solutions.)
#
# Usage: 
#      Ensure script is executable:       chmod +x /path/to/pdf_grep.sh
#      and then:
#        find /path/to/pdfs/ -type f -exec /path/to/pdf_grep.sh '{}' 'harmonic oscillator' \; 
#
# Dependencies:
#
#   -find       (replace with locate at your leisure.)
#   -grep       (or a suitable replacement.)  
#   -pdftotext  (comes with the poppler library, or pdf-utils in some distros.)
#
# Caveats:
#
# The pdf text is grepped, so be CAREFUL WITH LONG PHRASES which might be split across two
#  lines. The output is a line of text which matches the pattern. 
#
# How should you use this script? Search "en masse" through a collection of documents. If
#  any interesting leads appear, follow up with a proper search in a proper pdf viewer. If 
#  anything, this script will help narrow down searches accross libraries of texts.
#
# Adapted from a (gross) script at 
#  http://stackoverflow.com/questions/4643438/how-to-search-contents-of-multiple-pdf-files 
#  by hal clark in November of 2011.
###########################################################################################


# First argument is the pdf file.
THEFILE="$1"
shift

# All other terms are the search terms. Spaces will remain intact.
THETERMS="$@"

# PDFtoTEXT options.
PDFTOTEXTOPTS=" -q -layout -enc ASCII7 " # quiet, maintain layout, specify encoding.

# Grep options.
# output line numbers, ignore case, color when possible,
# contextual lines printed.
GREPOPTS=" -n -i --color=auto -C 2 "


# File existence check.
if [ ! -f "${THEFILE}" ] ; then
    echo "File ${THEFILE} was not found. Exiting."
    exit
fi

# Begin the actual searching. We have to convert the entire document to text 
# before searching, so grab some popcorn or break out the knitting needles.
echo "### Searching for phrase \"${THETERMS}\" in file \"${THEFILE}\"."
pdftotext ${PDFTOTEXTOPTS}  "${THEFILE}"  -  |  grep ${GREPOPTS}  "${THETERMS}"


# Output the total number of lines of text to give some reference to the output. Will 
# probably double execution time...
# echo "Total lines in document: $( pdftotext ${PDFTOTEXTOPTS}  "${THEFILE}"  -  |  wc -l )"