Portality - a web frontend based on elasticsearch.
This repo uses submodules - for more info: http://git-scm.com/book/en/Git-Tools-Submodules
This repo is a collection of useful things - it requires customisation in order to get working. Make sure you READ BELOW to get it up and running properly.
More information including a description of the various available default views is available at http://cottagelabs.com/software/portality
sudo apt-get install python-pip python-dev build-essential
sudo pip install --upgrade pip
sudo pip install --upgrade virtualenv
sudo apt-get install git
virtualenv -p python2.7 portality –no-site-packages
git submodule init git submodule update
pip install gunicorn
# python setup.py install pip install -e .
By default portality assumes an elasticsearch backend for the dao to talk to. There are numerous ways to install elasticsearch, here is an example:
tar -xzvf elasticsearch-0.18.2.tar.gz
ln -s elasticsearch-0.18.2 elasticsearch
Then you should install the elasticsearch service, and create a user to run it on your system if you so desire. Then you should ensure that user has unlimited ability to lock memory - to do this for root, it looks like this:
ulimit -l unlimited
You should also set some config before you start elasticsearch - in the conf/elasticsearch.yml uncomment bootstrap.mlockall to enable it. Also set the RAM to as high a number as you can afford on your machine. Then to ensure you do not accidentally contaminate elasticsearch with indices from other running instances on the same network, set the elasticsearch group name to something that is not the default. That way, elasticsearch will only try to auto-shard between instances that you manually set to have that same group name. Then once you are ready to go, just do
sudo /etc/init.d/elasticsearch start
Important configuration options
When you customise your app.cfg (read below) you should ensure you set a random long string as your secret key, and that you properly set the location of your elasticsearch instance, and any running info such as ports - making sure not to clash with other services running on your server. When you first start your system, you should create a new user with the same name as the SUPERUSER config setting - that account will then have full frontend admin control. If you want custom mappings in the elasticsearch index for your models, ensure they are defined in your config and they will be created when the system starts, if they are not already there - note also the config setting for the name you wish to use for your elasticsearch index.
Portality provides various useful default bits and pieces - you need to choose which ones you want and enable them. In the portality folder you will find a web.py file and a default_settings.py file - these are configured to setup everything that portality has, so you should copy web.py to a different file - say app.py - and then customsise it to only run the things you need. You should also make a copy of default_settings.py and call it app.cfg and put it in the folder above - the root folder of the portality repo. Then you can overwrite the settings and they will be read from there on startup.
For data access and modelling, the dao.py file defines the data access model which you can leave as is unless you want to change the backend, and the default_models.py file has some example data models that inherit from the dao. You should symlink models.py to default_models.py if you want to use it as is, or else create a new models.py and write your own models into it.
When you first clone portality, it includes default_auth, default_view and default_templates folders - if you want to use their contents as is, you should symlink to each of them from folders called auth, view and templates. Alternatively, create folders called auth, view and templates and then copy/ paste the ones you want, and write your own. Each view that needs templates should have corresponding default templates in the templates folder.
Once you are fully setup with app.cfg, models.py, auth, view and templates, you can try running your app by calling your app file - using the example name of app.py, do the following:
A note about feeds
There is a view called feed.py which provides data feeds, however it requires lxml - however as this is a bit more complex to install, it is not called by the default web.py and lxml is commented out in the setup.py. So if you want to use the feed view, install the lxml requirements on your local system then uncomment lxml in the setup.py and re-install. You will then be able to import the feed view and use it as required. Some example code for getting lxml on an ubuntu machine (sudo / permissions required):
sudo apt-get install libxml2-dev libxslt-dev
sudo apt-get install lib32z1-dev
sudo apt-get install python-lxml
And if you get a weird error even then, note this particular problem with some versions of virtualenv on ubuntu - which I had even after upgrading virtualenv to the most recent version:
Doing the symlink mentioned in the first answer solved it for me.
sudo ln -s plat-x86_64-linux-gnu/_sysconfigdata_nd.py .
When you want to customise portality, you should start your own new git repo and then add portality as an upstream source. Then you can merge it with your own local master.
git remote add upstream https://github.com/CottageLabs/portality.git git fetch upstream git merge upstream/master
Next you should create real versions of the default files and folders as described above. Then you can edit them as you require, and when you next want to sync with portality, you can just fetch and merge from upstream again without conflicting with your local changes.
NOTE that if you do overwrite any of the other files in your own repo copy, you will need to check for conflicts after a fetch and merge, and fix them.
Check out the Flask docs for further info about the framework.
If you want to contribute your changes back up the fork and into portality, make sure you prepare a branch of your local repo that conforms to the portality structure, ignore anything that is relevant only to your own specific instance, and send a pull request. (https://help.github.com/articles/using-pull-requests)
Run it with proper deployment
Use your preferred web server setup to expose your website. For example, nginx proxy passing to the python script, which itself can be run using supervisord.
You should install nginx with a sites-available folder and a sites-enabled folder. There is an example default nginx config in the deploy folder of the portality repo. Copy or symlink it into the sites-available folder of nginx, then symlink that in sites-enabled and restart nginx.
sudo ln -s /opt/portality/src/portality/deploy/portality_nginx_config portality
sudo ln -s ../sites-available/portality portality
sudo /etc/init.d/nginx restart
Now nginx will proxy-pass to your app, so you just need to have your app running. If you did the pip install gunicorn in the virtualenv install above, then you can do this using the supervisord config if you like (or your own preference).
sudo apt-get install supervisor
sudo ln -s /opt/portality/src/portality/deploy/portality_supervisord_config
sudo supervisorctl reread
sudo supervisorctl update
Auto-deployment and backups
And coming soon a default view that acts as a git webhook to auto-deploy.
And perhaps some built-in backup.