montana / prowl-dgraph

Documentation on using Dgraph with Prowl

Using Dgraph with Prowl & Docker Installation

If you're not familiar with Dgraph, visit their website at https://dgraph.io/. Look through their documentation, but in this instance, this documentation will be walking you through how to install what you need for this to run on the Prowl framework, and to get it running.

First access the Prowl VPS that all Prowl employees should haves access to. Make sure it's running the right kernel you can do this by running

uname

I suggest you use these flags when running uname

m, o

So the command should be

uname -mo

It obviously should be running the same kernel, since the VPS is the same all accross the board for each Prowl employee. Depending on what Prowl VPS you're on, you may need to update Docker, so run

sudo apt-get remove docker docker-engine

The contents of /var/lib/docker/, including images, containers, volumes, and networks, are preserved. The Docker CE package is now called docker-ce, and the Docker EE package is now called docker-ee.

Now run to install Docker

sudo apt-get update
sudo apt-get install \
linux-image-extra-$(uname -r) \
linux-image-extra-virtual

Now we are going to add the Docker repo, since all our VPS's are on 64-bit, you need to run

sudo add-apt-repository \
"deb [arch=amd64] https://download.docker.com/linux/ubuntu \
(lsb_release -cs) \
stable"

Now update,

sudo apt-get update

Make sure Docker is installed via

docker -v

Alternatively you can run

sudo docker run hello-world

Make sure you're connected to your VPS you're assigned to, we call this at Prowl the "VPS Cluster", this is all stated in our private repo of information, it should look something like this in the interface

VPS	Private IP address (ethN)	VPN IP address (wg0)
VPS1	10.8.23.93	10.0.1.1
VPS2	10.8.23.94	10.0.1.2
Docker	10.8.23.95	10.0.1.3
Docker2	10.8.23.95	10.0.1.3

Dgraph Installation

Clear all the pre existing Docker information, now run

curl https://get.dgraph.io > /tmp/get.sh
vim /tmp/get.sh  # Inspect the script
sh /tmp/get.sh   # Execute the script

In some cases vim /tmp/get.sh doesn't work, for some reason, so try

vi /tmp/get.sh

Assuming you have Docker installed and followed earlier instructions, pull Dgraph from Docker

docker pull dgraph/dgraph

Now you can actually run the dgraph command, by simply running

dgraph

In this case, we are going to be binding dgraph on port 4040, run

mkdir -p ~/dgraph
# Mapping port 4040 from within the container to 9090  of the instance
docker run -it -p 9090:4040 -v ~/dgraph:/dgraph dgraph/dgraph dgraph --bindall=true

Now via the information we get from the Prowlbox, we can search using Dgraph via mutation. An example of this, would be the following JSON contents

{
    "me": [
        {
            "artist": [
                {
                    "name": "Weiss"
                }
            ],
            "song_title": "Man Gone",
            "release_date": "2016-05-21",
            "running_time": 4:29,
                {
                    "label": "Toolroom"
                },
                {
                    "device_name": "JakesiPhone"
                },
                {
                    "playlist_boolean": "no"
                }
            ]
        },
        {
                }
            ]
        }
    ]
}

Some of the Docker engines have problems with Dgraph, so troubleshoot via

docker run -it dgraph/dgraph bash

Then try running Dgraph again

dgraph

How it should look

This is how it should look if you followed the instructions correctly

Hooks

Model hooks are available on both serialization and deserialization. These can be used to control what of the model data gets serialized (eg. sent to the client) or what will gets written into the model after deserialization.

Here are the available callbacks (can be directly implemented on the DataObject or in a DataExtension)

Signature	Parameter type	Info
onBeforeDeserialize(&$data)	string	Called before the raw JSON is being parsed. You get access to the raw JSON data sent by the client.
onAfterDeserialize(&$data)	array	Called after JSON has been deserialized into an array map. You can modify this array to prevent incoming values to be applied to your model (sanitize incoming data).

This of course is to parse the JSON information you are getting from Dgraph, this is a feature within Prowl. In conclusion, Dgraph is a quick way to search for information Prowl collects, it wouldn't be our main way to search for data, but for on the go situations, this is very ideal.