Planet Generator

Easily create a blog roll or planet out of multiple feeds on a similar subject.

How to use

Change the $csv_url in find_feeds to point to a CSV file that lists the sites you want to aggregate. If you don't want to download blogs.csv from somewhere, just put it in the work dir as blogs.csv and disable the fetch_csv function call.

Run download_feeds to actually download the content of each feed in feeds.csv, the file that was generated by find_feeds.

Edit the $planet_XXX and @keywords variables in generate_planet to your needs. The @keywords array are words the feed items must match to be included (at least one). Run the script to create planet.xml in the work dir. Symlink this file into your website somewhere.

Install appropriate entries in your crontab to make it run on a regular basis.

Creating a wrapper script similar to this (I usually name it run.sh) is probably the simplest way to use it:

#!/bin/bash

cd $HOME/src/planet_generator
mkdir -p work
./find_feeds
./download_feeds
./generate_planet
if [ -r work/planet.xml ]; then
    cp work/planet.xml /somewhere/else/planet.xml
fi
cd - >/dev/null

Using Docker

Run ./build_docker_image once to create the Docker image planet_generator.

Creating a wrapper script similar to this (I usually name it run.sh) is probably the simplest way to use it:

#!/bin/bash

pushd $HOME/src/planet_generator >/dev/null
mkdir -p work
./run_docker_container
if [ -r work/planet.xml ]; then
    cp work/planet.xml /somewhere/else/planet.xml
fi
popd >/dev/null

The Docker image entrypoint uses gosu from https://github.com/tianon/gosu. The bundled version is compiled for Linux amd64.

Dependencies

Install these packages with your favorite Perl CPAN client.

• Text::CSV
• XML::Feed
• Feed::Find
• URI
• LWP::Protocol::https
• File::Slurp
• DateTime
• Regexp::Assemble

Copyright

Robin Smidsrød robin@smidsrod.no 2013-2015

License

Same as Perl 5.