Sweet walkthrough

This intro should get you up and running with a basic Blingalytics installation in just a few minutes. Once you’re done, you can check out the documentation for Data Sources and Cache stores to contemplate your options for snappier infrastructure.

If you haven’t done so yet, you’ll need to install the blingalytics package and requirements before beginning. See Installation for details.

One: define a report

With Blingalytics, you use a “report” definition to describe precisely what data you want to look at and how you want to slice it. In a report definition, you’re only required to do two things:

  • List the output columns and where they should get their data
  • Define the key range for the report (explained below)

Baby’s first report

So to start, let’s put together a completely useless simplest-case report:

from blingalytics import base, formats
from blingalytics.sources import key_range, static

class LameReport(base.Report):
    keys = ('lame', key_range.SourceKeyRange)
    columns = [
        ('lame', static.Value(5, format=formats.Integer)),

So what does this report do? It provides one output column, whose value will always be 5. However, that’s not even the most useless property of this report, as this report will actually return zero rows. This is why a report’s keys matter.

Key concept

The keys for a report determine what rows will be in the output. If your website sells doodads, you might want to see how many doodads you sell per day. In this case, the report keys would be the range of days you report on so that you get one row per day. If, on the other hand, you want to see how many doodads each user has bought, you would want one row per user. So the user ID is your report key, and the range would be all your users.

To specify the key range for your report, you set the keys attribute of your report class to a two-tuple. The first item is the label of your key column, and the second item is the type of key range. In our example, the sources.SourceKeyRange tells the report to only include rows returned by the source data. Other key ranges, such as a range of days, can be used to ensure that a row is returned for each key, even if there is no source data.

For advanced use cases, you can even have compound keys. For example, you could have a row per user per day. See Creating and using reports for more.

Build your columns

Columns are defined in a report as a list of two-tuples. Each two-tuple represents a column, in order, by defining a label and its data source. The label should be unique among the columns, and is used by keys and other options to reference that column. The data source defines how that column’s data should be computed, and is covered in more detail in Data Sources.

A slightly realer report

Now that we know a bit more about how this works, let’s define a report that actually does something (to be fair, it’s still pretty useless, but we’re getting closer):

from blingalytics import base, formats
from blingalytics.sources import derived, key_range, static

class RealerReport(base.Report):
    keys = ('prime', key_range.IterableKeyRange([2, 3, 5, 7, 11, 13, 17]))
    columns = [
        ('prime', key_range.Value(format=formats.Integer)),
        ('squared', derived.Value(lambda row: row['prime'] ** 2, format=formats.Integer)),
        ('note', static.Value('Useful data coming soon...', format=formats.String)),
    default_sort = ('prime', 'asc')

So now we’ve got a report with three columns: prime is one of the prime numbers from the key range; note is simply a static string value; and squared is the square of the prime number. OK, time to run it!

Two: run the report

Once you’ve defined a report, such as RealerReport, you can instantiate the report and tell it where to cache the data:

from blingalytics.caches.local_cache import LocalCache
report = RealerReport(LocalCache())

Once you have a report instance, you can run the report:


Retrieving report rows

Now that the report is cached, you can retrieve the data with limits, offsets, column sorting, and so on. But in the simplest case, you can just get all the rows:

rows = report.report_rows()
# rows = [
#    [1, '2', '4', 'Useful data coming soon...'],
#    [2, '3', '9', 'Useful data coming soon...'],
#    [3, '5', '25', 'Useful data coming soon...'],
#    [4, '7', '49', 'Useful data coming soon...']]
#    [5, '11', '121', 'Useful data coming soon...'],
#    [6, '13', '169', 'Useful data coming soon...'],
#    [7, '17', '289', 'Useful data coming soon...'],
# ]

Let’s try sorting and limiting the data:

rows = report.report_rows(sort=('squared', 'asc'), limit=3)
# rows = [
#    [7, '17', '289', 'Useful data coming soon...'],
#    [6, '13', '169', 'Useful data coming soon...'],
#    [5, '11', '121', 'Useful data coming soon...'],
# ]

There are plenty more options for retrieving specific rows. See Report.report_rows for more.

Three: put it in the browser

Working in the Python interpreter is nice and all, but most of us want to insert this thing into a beautiful website and spread it around the web. Good news! Blingalytics comes with the tools to pull this off in just a few lines of code. (For example purposes, this will be shown as a Flask app, but it should be easy enough to insert this into your favorite Python web framework.)

The report app

Let’s assume you’ve already installed Blingalytics and installed Flask.

Now we build a very basic Flask app that has two URLs: a homepage, where we’ll display our report; and an AJAX responder for our report JavaScript to talk to. The report_response helper function makes responding to the AJAX requests easy.

from blingalytics.helpers import report_response
from flask import Flask, request, render_template
from reports import RealerReport # import so it gets registered

app = Flask(__name__)

def index():
    return render_template('index.html')

def report():
    return report_response(request.args)

if __name__ == '__main__':

The index page really just renders the template where you’re going to implement the Blingalytics frontend. The report URL handles AJAX requests from the JavaScript frontend, and the report_response handles all the dirty work of parsing request parameters and interfacing with your report classes.

The report template

The template is even easier. Just include the appropriate CSS and JavaScript on your page, and invoke the Blingalytics jQuery plugin:

<!DOCTYPE html>
<html lang="en">
  <link rel="stylesheet" href="/static/css/blingalytics.css" type="text/css" />
  <div id="report"></div>
  <script src="//ajax.googleapis.com/ajax/libs/jquery/1.6.2/jquery.min.js"></script>
  <script src="/static/js/jquery.dataTables.min.js"></script>
  <script src="/static/js/jquery.blingalytics.js"></script>
      'reportCodeName': 'realer_report'

With the CSS and JavaScript in place, you just have to pass in the 'reportCodeName' option with the code name of the report you want to display.

Partake in the beauty

That’s it! Run your web app. For Flask, this involves running python app.py at the command line, where app.py is the app file you created earlier. Then you should be able to point your browser to localhost:5000 and play with the report in the browser.

Four: pull real data

Be patient... coming soon.