Active File Processor

(Full Documentation for this will be written shortly.)

What is it?

It's designed to sit on the backend of a website that accepts images and videos from users, and transodes them to sizes suitable for the web, and to flash video.

The code is however more general, and could in fact be used for anything that needs to watch a bunch of directories and do interesting things (hence the name of this page).

What does it do right now?

At present, the code does this:

• Has directory watchers
• When a file is added to the source directory, the first processor splits it off into different directories based on file type
• Another does image size conversion, cropping etc,
• Another does video conversion to flash video.
• The result is images & videos placed into a directory for moderation.

It's designed to run as a backend system that sits in the background running continuously.

Draft Installation Instructions

Install Image Magick as normal.

Download and install libamr

This is to support transcoding of video from mobile phones.

Originally from http://www.penguin.cz/~utx/amr

curl -O http://ftp.penguin.cz/pub/users/utx/amr/amrnb-7.0.0.2.tar.bz2
tar jxvf amrnb-7.0.0.2.tar.bz2
cd amrnb-7.0.0.2/
./configure --libdir /usr/lib/
make  #   -- requires wget to build
sudo su
make install
ldconfig

Download and install ffmpeg

This is used by the transcode engine to transcode videos. You can however plug something in its place relatively easily.

svn co svn://svn.mplayerhq.hu/ffmpeg/trunk ffmpeg-trunk
cd ffmpeg-trunk/
./configure --enable-nonfree --enable-libamr-nb
make
sudo su
make install
ldconfig

Download, install and configure the transcode engine.

Grab the code and install it

curl -O http://www.kamaelia.org/release/Kamaelia-FileProcessor-0.1.0.tar.gz
tar zxvf Kamaelia-FileProcessor-0.1.0.tar.gz
cd Kamaelia-FileProcessor-0.1.0
sudo python setup.py install

Create uploads and moderation directories for transcode/moderation pipeline

cd /tmp # or whereever you prefer - you'll need to change a config option no matter what :-)
mkdir incoming
mkdir -p uploads/images uploads/videos
mkdir -p moderate/images moderate/videos

Make everything writeable by the webuser.

Configure it
Copy the default config to one you want to edit

cp /etc/batch_converter.conf.dist /etc/batch_converter.conf

Then edit /etc/batch_converter.conf such that the following lines/config options ...

# main_incoming_queue /tmp/uploads
# image_queue /tmp/uploads/images 
# video_queue /tmp/uploads/videos 
# image_moderation_queue /tmp/moderate/images
# video_moderation_queue /tmp/moderate/videos

... get changed to:

# main_incoming_queue /tmp/incoming
# image_queue /tmp/uploads/images 
# video_queue /tmp/uploads/videos 
# image_moderation_queue /tmp/moderate/images
# video_moderation_queue /tmp/moderate/video

... or whatever you changed it to...

Grab and install PIL ("python imaging library")
Download from

http://effbot.org/downloads/Imaging-1.1.6.tar.gz

Unpack and install

tar zxvf Imaging-1.1.6.tar.gz
cd Imaging-1.1.6/
sudo python setup.py install --force

Note: PIL requires the ability to build tcl/tk apps for odd reasons. You may find you need to install that dependency before PIL will build. Hence the recommendation to use am .rpm or .deb file if you can.