Webalyzer helps you collect HTML from live web pages and run analysis on the the HTML in batches for reporting purposes.

• PostgreSQL
• Redis
• Python & pip
• node & npm

1. Start your PostgreSQL and Redis servers

2. Create and activate a virtual environment:

    virtualenv venv
    source venv/bin/activate
   

3. Install python requirements from the requirements.txt file:

    pip install -r requirements.txt
   

Ideally you should use peep to assure you don't install anything that hasn't been vetted for:

    pip install peep
    peep install -r requirements.txt

4. Install node requirements:

    npm install -g crass yuglify
   

5. Create postgres database:

    createdb webalyzer
   

6. Add default environment variable settings in .env file:

    SECRET_KEY=terrible-secret
    DEBUG=True
    ALLOWED_HOSTS=*
    DATABASE_URL=postgresql://<username>@localhost:5432/webalyzer
   

7. Run migrations:

    python manage.py migrate
   

Run the django server:

    python manage.py runserver

Run the message queue worker:

    python manage.py celeryd

By default the message queue (using celery) uses the same database as the site depends on. This makes it easy for local development.

To override this to use a RabbitMQ server or something like that set the environment variable BROKER_URL. For example:

    export BROKER_URL="amqp://guest:guest@localhost:5672//"

1. Add a <script> tag to your page for collect.js:

    <script>
    (function() {
    var s = document.createElement('script');
    s.src = 'http://127.0.0.1:8000/static/collect.js';
    s.async=true; s.onload=function() {
        Webalyzer('example.com');
    } document.head.appendChild(s)})();
    </script>
   

2. Hit your page(s)

3. Go to the collected analysis page: http://127.0.0.1:8000/analyzer/

There are configuration options available for when you initialize the Webalyzer in that script snippet. For example, instead of:

    Webalyzer('example.com');

You can do this:

    Webalyzer('example.com', {
         doSend: function(html, url) {
             if (getWeather() === 'raining') {
                 return false;
             }
             return true;
         },
         beforeSendHTML: function(html, url) {
             return html.replace(/negative/g, 'positive');
         },
    });

The doSend function gives you an opportunity to cancel the sending, by returning false. You could potentially use this to throttle some of the traffic or perhaps you only want it sent depending on some other condition.

The beforeSendHTML gives you an opportunity to change the HTML that's going to be sent. For example, you might scrub something in the content.