SETUP.txt

This all depends on a working neo4j install
  - Install neo4j, set-initial-password, create breedcafs user and set that accounts password

An instance folder is required but not available in the github repo, this should contain
modified version of the config.py file (notes in the template in this folder).

SETUP:

Set up neo4j:
	Install:
		on ubuntu:
			apt-get install neo4j
		on fedora:
			yum install neo4j

	Set initial password:
	    service neo4j stop
		neo4j-admin set-initial-password NEO4J_ROOT_USER-PASSWORD-HERE
		service neo4j start

    Connect to db using cypher shell to create breedcafs user:
        cypher-shell -u neo4j -p NEO4J_ROOT_USER-PASSWORD-HERE
        CALL dbms.security.createUser('breedcafs', BREEDCAFS_USER-PASSWORD-HERE, false);
		# Now that I am using ONgDB for enterprise features we want to set roles for the users
		CALL dbms.security.listRoles();
		#CALL dbms.security.addRoleToUser('editor', 'breedcafs');
		CALL dbms.security.addRoleToUser('admin', 'breedcafs');

        # the 3rd argument in the above call suppresses the request to change password

        Other useful commands:
        CALL dbms.security.changePassword(*password*)
        CALL dbms.security.listUsers()
        CALL dbms.security.deleteUser(*username*)

    Add the breedcafs user and password as environment variables for the web-server and celery instances

	Set import directory (for CSV uploads):
		Edit /etc/neo4j/neo4j.conf to include the path for imports (use same path later in instance/config.py):
			# `LOAD CSV` section of the manual for details.
			dbms.directories.import=/path/to/import

	Add APOC plugin:
		Download the compatible version of APOC (jar file) and place it in the neo4j plugins directory.
		  - https://github.com/neo4j-contrib/neo4j-apoc-procedures
		APOC is currently used for date parsing during CSV import and for clearing schema on reset

		For schema assert add the following to /etc/neo4j/neo4j.conf
            dbms.security.procedures.unrestricted=apoc.schema.assert


Set up the flask app:
    Clone the breedcafs repository into the webserver directory:
        cd /var/www
        git clone https://github.com/marcusmchale/breedcafs

    Add the instance folder, copy the config.py and modify to suit your environment:
        mkdir breedcafs/instance
        cp /var/www/breedcafs/config.py /var/www/breedcafs/instance/config.py

    Set up virtual environment:
        Install:
            on ubuntu:
                apt-get install virtualenv
            on fedora (was already installed for me in v25 but otherwise:
                yum install python-virtualenv

        Initialise:
            virtualenv /var/www/breedcafs/venv
        Activate (from breedcafs folder):
            . ./var/www/breedcafs/venv/bin/activate
        Install requirements:
            pip install -r REQUIREMENTS.txt

    Set permissions so that webserver user has group level read access:
        chgrp -R www-data /var/www/breedcafs
        chmod 755 /var/www/breedcafs
        chmod 2775 /var/www/breedcafs/instance/import
        chmod 2775 /var/www/breedcafs/instance/export


Configure apache2:
	Add the following virtual host to /etc/apache2/sites-available/breedcafs.conf
		<VirtualHost *:80>
			ServerName SERVERNAME-OR-IP
			SetEnv NEO4J_USERNAME neo4j
			SetEnv NEO4J_PASSWORD SECURE-PASSWORD-HERE
			SetEnv MAIL_PASSWORD SECURE-MAIL-PASSWORD-HERE
			SetEnv MAIL_USERNAME your@email.add
			WSGIDaemonProcess breedcafs python-home=/var/www/breedcafs/venv processes=2 threads=15 display-name=%{GROUP}
			WSGIScriptAlias / /var/www/breedcafs/breedcafs.wsgi
			<Directory /var/www/breedcafs>
			    WSGIProcessGroup breedcafs
			    WSGIApplicationGroup %{GLOBAL}
			    Require all granted
			</Directory>
		</VirtualHost>

	Activate the site:
		cd /etc/apache2/sites-available
		a2ensite breedcafs.conf

	Alternative for local install (Werkzeug server):
		export FLASK_APP=run.py
		flask run

SELinux:
	Wherever apache logs are kept needs to be set as http log type:
		chcon method:
			chcon -Rv --type=httpd_log_t /httpd/logs
		semanage method (persists through file system relabel:
			semanage fcontext -a -t httpd_log_t “/httpd/logs(/.*)

Redis:
	This is needed as an in memory data store for asynchronous tasks (like uploading large files)
	Install:
		On fedora 25:
			sudo dnf install redis
	Run:
		On fedora 25:
			sudo systemctl start redis


Celery: 
	This is the python worker that schedules asynchronous jobs on the server
	Install and run as service, modify the celeryd.conf file
	#run for testing
		celery -A app.models worker --loglevel info