Basic Installation¶

This guide will take you through setting up A_A in development mode. Analytics automated is a Django application and has a couple of required dependencies

Python3
postgreSQL
Redis
Celery 4.x
Django >1.10

Code Dependancies¶

1. Get Python 3¶

You may also wish to sort out virtualenv and run the application within that, however that is beyond the scope of this installation guide

https://www.python.org/downloads/

2. Install postgreSQL¶

If you’re on a mac we advise using brew

brew install postgres

If you’re in a linux env some manner of

yum install postgres

or

apt-get install postgres

3. Install Redis¶

One of

brew install redis

yum install redis

apt-get install redis

4. Get the Analytics Automated¶

git clone https://github.com/AnalyticsAutomated/analytics_automated.git

This will place the code in a dir called analytics_automated/

5. Install the python packages needed¶

cd analytics_automated/
pip install -r requirements/dev.txt

Pre-launch configuration¶

1. Set Up Postgres¶

You need to setup some bits and pieces in postgres before we start

You may need to initialise postgres:
```
initdb -D [SOME_PATH]
```

Start the postgres daemon:

pg_ctl start -l /scratch0/NOT_BACKED_UP/dbuchan/postgres/logfile -D scratch0/NOT_BACKED_UP/dbuchan/postgres/

Then login:
```
psql -h localhost -d postgres
```

Now create the django db user for A_A

CREATE ROLE a_a_user WITH LOGIN PASSWORD 'thisisthedevelopmentpasswordguys';
CREATE DATABASE analytics_automated_db;
GRANT ALL PRIVILEGES ON DATABASE analytics_automated_db TO a_a_user;
ALTER USER a_a_user CREATEDB;

2. Now configure Django¶

We maintain the idea of separate secrets files which only you have control of. You need to create these and populate them. base_secrets.json are site wide settings which dev and production will use. dev_secrets.json are settings which only the dev installation will will access. A production system will need a production_secrets.json

Create the files we need:

cd analytics_automated_project/settings
touch base_secrets.json`
touch dev_secrets.json`

If you’re using bugsnag add your bugsnag key to base_secrets.json:
```
{
  "BUGSNAG": "YOUR KEY HERE"
}
```

Add the dev_secrets.json settings needed to start in developments mode. The postgres login credentials and the secret key:

{
  "USER": "a_a_user",
  "PASSWORD": "thisisthedevelopmentpasswordguys",
  "SECRET_KEY": "VERY LONG KEY HERE"
}

Next open the base settings files in analytics_automated_project/settings/base.py In here you’ll find a section at the top labelled “Required A_A user settings”. These are all the things you need set for the app to run. We prefer to keep theses settings in dev.py and production.py files. Then we can start the server in different configs for different purposes. Either uncomment all these in base.py or move them to dev.py or production.py and set them there. You can leave the smtp settings commented if you do not wish to send alerts via email to your users.

3. Starting A_A in development localhost mode¶

Start Redis:
```
redis-server
```
Start the celery workers, from the root dir of A_A. Note that we have to specify the queues the workers read from (-Q), for the basic settings we’ll have these workers just watch all the default queues, note that the workers are watching the low priority, normal priority and high priority. In a more complex set up you can have different worker pools on different machines watch specific queues. localhost queues:
```
cd analytics_automated/
celery --app=analytics_automated_project.celery:app worker --loglevel=INFO -Q low_localhost,localhost,high_localhost,low_GridEngine,GridEngine,high_GridEngine,low_R,R,high_R,low_Python,Python,high_Python
```

Run the Django migrations to configure the database:

cd analytics_automated/
python manage.py migrate --settings=analytics_automated_project.settings.dev

Add an admin user to the Django application:

cd analytics_automated/
python manage.py createsuperuser

Now start A_A, again from the root dir of the app. Note we’ll start it assuming you put the users settings in settings/dev.py:
```
cd analytics_automated/
python manage.py runserver --settings=analytics_automated_project.settings.dev
```
ALTERNATIVELY We also provide some scripts for bash and OSX in the utilities/ directory which will start all the components on one machine.
Scheduled tasks. If you are going to user celery-beat then you may want to add a queue name for scheduled tasks to the workers and starting the beat service.

4. Config complete¶

You should now be running all the components of A_A on a single machine with a set of workers watching the localhost queue. This means we can now configure data analysis pipelines which run code on the machine which the workers are running on. This is the most basic setup we can run Redis, the web app, the database and the workers on completely separate machines and even run multiple instances of the workers watching the same queue. We’ll deal with this set in the Advanced Uses tutorial.

Basic Installation¶

Code Dependancies¶

1. Get Python 3¶

2. Install postgreSQL¶

3. Install Redis¶

4. Get the Analytics Automated¶

5. Install the python packages needed¶

Pre-launch configuration¶

1. Set Up Postgres¶

2. Now configure Django¶

3. Starting A_A in development localhost mode¶

4. Config complete¶

5. Now move on to How The System Works ¶

analytics_automated

Navigation

Related Topics

Basic Installation¶

Code Dependancies¶

1. Get Python 3¶

2. Install postgreSQL¶

3. Install Redis¶

4. Get the Analytics Automated¶

5. Install the python packages needed¶

Pre-launch configuration¶

1. Set Up Postgres¶

2. Now configure Django¶

3. Starting A_A in development localhost mode¶

4. Config complete¶

5. Now move on to How The System Works¶

5. Now move on to How The System Works ¶