This repository contains a handful of examples tying together use of the voevent-parse and Comet packages, providing a quick and easy way to get started sending or receiving VOEvents with the 4 Pi Sky VOEvent broker (or any other VOEvent broker, for that matter).
The code provided here makes many assumptions about what you want to do, in order to present a simple interface. Those with more specific requirements should refer to this package as a quick example, and then use voevent-parse and Comet directly (see also: the voevent-parse tutorial, and pysovo).
(For the impatient, who are running Ubuntu/Debian and have the virtualenv tool available)
sudo apt-get install libxml2-dev libxslt-dev
source INSTALL.sh && source RUNME.sh
You'll need to install some Python packages for these scripts to work. Working in a virtualenv is recommended, but not essential. I also recommend use of pip since it allows easy version checking and package uninstallation.
Note that one of the python package dependencies is
lxml,
which has some prerequisites for installation that can cause a
standard pip install
to fail, sometimes with confusing errors.
So, before we do anything else it might be a good idea to try
pip install lxml
and see if that works for you.
On Ubuntu or Debian Linux you can satisfy those requirements by running:
apt-get install libxml2-dev libxslt-dev
before you attempt to install anything else. For more information (e.g. for Mac OSX), see the relevant Notes on VOEvent section or consult the lxml docs.
You can check if LXML is already installed by simply trying
import lxml
from the python interpreter.
Next, grab a local copy of the fourpiskytools
repository:
git clone https://github.com/timstaley/fourpiskytools.git
Then, cd fourpiskytools
and (if you're working with a virtualenv):
pip install .
If you prefer not to use a virtualenv, you can instead install into your user-area:
pip install --user .
If you run into problems, try consulting the installation guides for voevent-parse and Comet and verify that you can install those packages individually.
First, you should verify that you can successfully run the Comet
command
line tool. For example, at the command line try:
twistd -n comet -v --local-ivo=ivo://fpstoolstest/foo#bar --remote=voevent.4pisky.org
(For more detail, see http://voevent.readthedocs.org/en/latest/receiving.html and Comet's own docs.)
Next we need to actually do something with
the VOEvents as they are received. The simplest way to do this is to
write a script that can accept the raw XML via stdin (i.e., so you can
pipe stuff to it via the command line).
Comet comes with a --cmd
option
so you can tell it to run a script of your choice and pipe it the XML contents for every VOEvent received.
We need to write a basic Python script to accept this input,
test it via the command line, and then run Comet with the right command line
options.
The examples folder contains a couple of basic VOEvent processing scripts to get you started. First, take a look at process_voevent_from_stdin_1.py. This very short script just reads in the data from stdin, parses it as a VOEvent object, and writes out a few key items of information to a logfile.
Let's try it out. There's a sample VOEvent XML packet in the examples folder, test_packet.xml. We can use this to run a quick test of our processing script, by simply piping it in using the command line. From the examples directory:
cat test_packet.xml | ./process_voevent_from_stdin_1.py
You should see something like:
At 2015-11-12 13:10:47.278333, received VOEvent with IVORN ivo://voevent.organization.tld/TEST#141113-2118.55_ac97a2b6
Authored by ivo://voevent.organization.tld at 2014-11-13T21:18:55
The same text should also be output to a logfile, script1.log
.
The second script,
process_voevent_from_stdin_2.py,
contains some slightly more interesting logic, showing how to run different
functions depending what sort of VOEvent packet is received. As an aside, note
that if you're on an Ubuntu system, you can pip install pgi
to get pop-up
desktop notifications using the code called from script 2.
The examples folder also contains a shell script that runs Comet and passes
VOEvents to a handler script - see
listen_for_voevents.sh. Note that you won't
see the command line output from the handler script - you'll have to inspect the
logfile instead. (You can track updates to the log by opening another terminal
window and running e.g. tail -f script1.log
.)
To see this in action, you can either wait for a VOEvent to be received from whichever broker you've connected to, or you can inject your own test-packet - try opening another terminal and running the ping_broker.py.
If you want to run Comet and a handler script on a long term basis, you have a few options.
For basic testing, you could simply manually start your version of listen_for_voevents.sh and leave it running in the background, perhaps using a detachable terminal environment such as GNU Screen or tmux.
In the longer run, you will probably want to configure your script to run as a system service, so that it gets restarted if your system is rebooted, etc. Unfortunately the best way to set this up will depend on your flavour of Linux, so you'll need to search for 'custom init script ubuntu 14.04' or something similar.
A full guide is on the To Do list! For now, see send_alert.py.
A quick contents for people who are already familiar with VOEvent, and are just looking to copy-and-paste some code-snippets:
- A function providing a subprocess wrapper about comet-sendvo, to allow publishing of VOEvents from a scripted process.
- A simple (~3 lines) routine for generating unique identifiers to tag your VOEvents with.
- A couple of ready-made functions that take a dictionary defining the sender's 'identity' and provide one-liner methods for composing VOEvent packets.
- A shell script demonstrating how to initialise Comet to receive alerts and pass them on to a Python script for processing.
- A basic utility class for generating GTK notifications (desktop alerts akin to e.g. the 'new email' pop-up that Thunderbird will generate).
Questions? Issues? Happy to help, either drop a note on the issue tracker or send us an email.
If you make use of these examples in work leading to a publication, we ask that you cite the underlying packages as detailed here and here.