scan_docs
A utility I wrote that handles scanned files created by my Brother
ADS-1500W.
You can modify it to work with any scanner that outputs files
ImageMagick can read, but that will require at least modifying the
regexes for matching its output files.
The Brother scanner supports CIFS/Samba, so I have a share that it
puts the files in. This script is told to monitor that directory,
and works accordingly.
USAGE:
Configure the documents you want it to support by following the
comments and the examples in the included config file.
Start the program by running:
scan_docs <directory_to_watch> [config_file]
If you don't specify config_file, it will look for scanner_config.yml
in the same directory as the script itself.
The program will then check the directory every 10 seconds for new
scans. If it finds one, it will convert it to pnm, run tesseract on
it, and then parse the text so it (hopefully) knows what directory to
put the file in.
If the file is unscannable, it'll put it in unscannable/ If it is
scannable, but doesn't match any configured documents, it will put it
in other/<time>.pdf.
I run this script from cron (it is safeguarded against more than one
copy running at once), so I just put:
* * * * * root scan_docs /docs
In my crontab. Then, when it successfully works on a file, it will
do its thing, and exit, and I get an email of what happened so I can
see if I need to take further action.
(If you just want to run it all of the time and output it to a log
file or something, you can remove the line that says 'exit;' below so
it won't quit every time it successfully does any work.)
Using this app and my brother scanner, my house has gone paperless.
I have scanned all of my documents, and this program categorised the
vast majority for me, making it super duper easy.
It also tries to be smart about monitoring files to make sure they've
stopped changing size before it works on them (just in case the file
is slow to copy or very large).
Also, do not use this on a multi-user system without modification
first -- it just scans to /tmp/scanned.pnm. This is both a possible
race-condition, as well as security vulnerability since there could
be sensitive information in the file.
Finally, if you want to write a subroutine to handle the filename
generation for a document, define it between the {} containing
'ScanSubs' below, and follow the examples in the config file
included.
Patches/pull requests welcome. Also, you could ask me nicely to make
changes if anyone who isn't me decides to use this thing.
DEPENDENCIES
Depends on the following perl modules:
* Fcntl
* YAML
* List::Util
* File::Path
* POSIX
* Date::Parse
* FindBin
Also depends on the following apps:
* Tesseract
* ImageMagick
WARRANTY/LICENSE
Program is licensed under the Perl Artistic license. You can
distribute it under the same terms as Perl itself.
Program has no warranty. Make sure it does what you want before you
shred your documents.