Data visualization (tree data-structure) web app, backend is built on node.js for rest api and websocket communication, while admin backend is written in python (Django/Flask), client is based on angularjs,d3.js.

clone the git module with: git clone https://github.com/dpiscia/paudm_web

install the module dependencies

cd paudm_web
npm install

You need to prepare or modify the standard one the config file for DB connection, the dbs have to be two in sqlite mode , and can be only one in postgres.

If you run grunt, then two sqlite db files will be created in test folder, you can eventually use these dbs as starting point

var config = {}
//if postgresql use the below config scheme
//config.job = {client : "pg" , host : "host name", user : "user" , port : 5432 , password : "pwd" , name : "db_name" }; 

//config.pau= {client : "pg" , host : "host name", user : "user" , port : 5432 ,password : "pwd" , name : "db_name" }; 
//if sqlite use the scheme belowe
config.job = {client : "sqlite3", name : "test/test_DB_job"};
config.pau = {client : "sqlite3", name : "test/test_DB_pau"};
config.session_store = false;
config.port = 3000;
config.redis = {port :6379, host : 'localhost'};
config.sync = false; //disabled if postgresql is used with two phases transaction or with sqlite3
module.exports = config;

module.exports = config;

in order to synchronize server to DB , some triggers have to be set on the DB side (it works only with postgresql) :

this sync is incompatible with two phase commits transaction. In this situation an alternative can be represented by using REDIS as message broker.

CREATE FUNCTION notify_trigger() RETURNS trigger AS $$
DECLARE
BEGIN
  PERFORM pg_notify('watchers', TG_OP ||  ',id,' || NEW.id );
  RETURN new;
END;
$$ LANGUAGE plpgsql;

CREATE TRIGGER watched_table_trigger AFTER INSERT ON job 
FOR EACH ROW EXECUTE PROCEDURE notify_trigger();

CREATE TRIGGER update_table_trigger AFTER UPDATE ON job 
FOR EACH ROW EXECUTE PROCEDURE notify_trigger();

CREATE FUNCTION old_notify_trigger() RETURNS trigger AS $$
DECLARE
BEGIN
  PERFORM pg_notify('watchers', TG_OP ||  ',id,' || OLD.id );
  RETURN new;
END;
$$ LANGUAGE plpgsql;

CREATE TRIGGER delete_table_trigger AFTER DELETE ON job 
FOR EACH ROW EXECUTE PROCEDURE old_notify_trigger();

install grunt:

npm install -g grunt-cli

testing can be done by typing:

grunt

Basic server side test coverage is provided, but it is far from optimal/complete.

##Continue Testing

Continue testing is done through Travis-ci https://travis-ci.org/dpiscia/paudm_web

(Coming soon)

In lieu of a formal styleguide, take care to maintain the existing coding style. Add unit tests for any new or changed functionality. Lint and test your code using Grunt.

##Admin

the admin is implemented by Django/bootsrapped, or by flask, a django-inspired framewrok not-opinionated and orm-flexible At the moment session will not be shared between the two backends, but it's possibile to use REdis and share it. From the flask admin it's possibile to re-use 100% of the sqlalchemy model already defined for scientific pipelines.

pre-release(pre-vacations) v 0.0.1

release (almost) v 0.1

https://github.com/dpiscia/paudm_web/releases

Copyright (c) 2013 PAUDM-team
Licensed under the MIT license.