#Pyler - A python tile server

##What's all this about then? Tile rendering is a fairly heavy task. Let's make this whole process more "cloudy". Rather than rendering the tile from a module on the webserver, let's split the rendering from the request. Push it off to another machine, who's only purpose is to render. Now we can have many small servers rather than a big beefy server. The Database is still the main bottleneck unfortunately, but maybe this can ease the pain a little.

##Plan:

• Use Tornado instead of flask
• Use Protobuf to create a metatile storage format (store 64 - 8x8 tiles in one file)
• Render tiles as an 8x8 metatile
• Distribute render tasks with Celery
• User Redis as a lock service to prevent the same metatile being rendered multiple times
• Figure out a way to configure/read styles to use
• Assuming both worker and server have access to the same folder. NFS works for distributed, by probably not optimal. Work out the best way to store the data. Database? Distributed FS? Memcached or similar?

##Installing You'll need mapnik2 installed, along with python bindings. Currently only tested with python 2.7, if I can get the mapnik bindings for python 3 installed, I'll test with those. You'll also need redis

###For Ubuntu 14.04:

sudo add-apt-repository ppa:mapnik/v2.2.0
sudo apt-get update
sudo apt-get install libmapnik libmapnik-dev mapnik-utils python-mapnik

To install redis

sudo apt-get install redis-server

You also need the requirements specified in requirements.txt. pip is the easiest way to install these.

pip install -r requirements.txt

###Map Styles In config.py, edit the MAPS list. Add a tuple with the mapname, and the path to the mapnik style xml.

To convert a CartoCSS style to mapnik xml, use the carto command from Mapbox.

##Running

$ python pyler.py

$ celery -A tasks.celery worker --loglevel=INFO --concurrency=3