search multiple pdf files for a string

[tlchost]

Trying to assist a local museum in searching pdf files. Need a solution to search multiple pdf files for a specific string.

Users are not technical, so it would be helpful for the script to look in a specific location, i.e. /mnt/home/newsletter, and search for a string, i.e. 'Howard Huges' in a case-insensitive manner. User could select output to screen or a text file.

Thanks

[Burn_IT]

You might find that a lot more difficult than it seems at first.
PDF is an optimised format and you would probably need to expand every document before you can search it.

[mikeslr]

or maybe not: https://www.online-tech-tips.com/comput ... s-at-once/

Of course, Burn_It is correct if the goal is to write your own program, especially as this was posted to the Programming Section. But I'm still working on my first cup of coffee --still more right-brain gestalt than left-brain analytical-- so saw the highlight of problem being "Users are not technical". PDF's having been around for ages, so somebody must have thought about batch searching pdfs before. Googled expecting to find a technical discussion. The above was the first post found.

Why re-invent the wheel? I use FoxitReader all the time. Didn't know it had a batch search capability.

[fabrice_035]

Hello,

You can try https://www.howtogeek.com/228531/how-to ... -in-linux/
(How to Convert a PDF File to Editable Text Using the Command Line in Linux)

https://poppler.freedesktop.org/

Or install with PPM

[technosaurus]

Mupdf has most of what you need in its tools along with a JavaScript interface. The source is pretty easy to follow if you have done any c programming, so you could implement your own tools for specific purposes using the existing ones as a template.

[fabrice_035]

I complete my answer with this (simple) program

Code: Select all

#!/bin/bash
# 
# -> PdfToText mandatory { https://poppler.freedesktop.org/ ] 
# -> PPM / PuppyPacketManager {  POPPLER-UTILS }
#
# search TeXt in pdf with pdftotext tool

trap ctrl_c INT

temp=/tmp/$(head /dev/urandom | tr -dc A-Za-z0-9 | head -c 13 ; echo '')

IFS=$'\n'

sortir() {

rm -f "$temp"
echo -e "\n Bye."
exit
}
export -f sortir

function ctrl_c() {
        sortir
}


binary=$(whereis pdftotext | awk -F: '{print $2}' | tr -d " " )

echo "$binary"
if [ "$binary" != "/usr/bin/pdftotext" ] ; then

echo "/usr/bin/pdftotext not found. End."

sortir

fi

path="$2"

if [ "$path" = "" ] ; then
echo "1) Use default path $PWD, you can also specify folder."
path="$PWD"
else
echo "search path:$2"
fi

if [ -d "$path" ]; then
:
else
echo "Directory not found!"
sortir
fi
if [ "$1" = "" ] ; then
echo "2) This tool search text in PDF : enter an occurrence !"
sortir

else
echo "Search \"$1\" in all .pdf files "
fi


files=$(find $path -iname '*.pdf')

for file in $files
do

echo -e ">Look in:$file"

/usr/bin/pdftotext "$file" "$temp" 
result=$(cat "$temp" | grep $1)

if [ "$result" != "" ] ; then
echo -e "- Found \"$1\" in  $file \nPress [Enter] to continue [o]pen pdf or e[x]it" ; read x
 if [ "$x" = "o" ] ; then
defaultpdfviewer $file & 
 fi
 
  if [ "$x" = "x" ] ; then
  sortir
  fi

fi
done

sortir

[rufwoof]

fabrice_035, does that include PDF's where the text content is being stored in image format? (Suspect not as I don't see any OCR type links/code).

[Burn_IT]

You cannot compare a text string with a graphic and expect to get meaningful results.

[tlchost]

Thanks folks....I'll give some of the suggestions a whirl and let you know which one worked for the museum.

[CatDude]

Hi
I find this quite useful myself: https://pdfgrep.org/

[slavvo67]

Catdude beat me to it. If the PDF is searchable, PDFGrep is a great solution. It should be in most repositories. It comes in RU Xerus as a default because I find it so handy.

All the best,

Slavvo67

[tlchost]

The plot thickens with the search in PDF files. Once the folks at the museum used the advanced search in Acrobat, they posed the question:

If the pdf files were online, would there be a way to search those files?

I can give them all the space they need on one of my servers, but I simply don't have the skills to write a script that would allow the online search. Anyone know of such a script?

thanks

my elemental internet search skills have not discovered one.