flattery - fast flattening and unflattening of nested Python data

This library exposes a fast C implementation for flattening and unflattening hierarchical Python data structures. A unit test suite is included.

Examples

Synopsis

#!/usr/bin/env python
from flattery import flatten, unflatten

data = {
  "x.y.0": "zero",
  "x.y.1": "one",
  "x.z" : 42
}

print unflatten(data)
>>>
{
  'x' : {
    'y' : [ 'zero', 'one' ],
    'z': 42
  }
}
<<<

assert( data == flatten(unflatten(data)) )

Processing hierarchical records

#!/usr/bin/env python
import sys
from flattery import unflatten

cols = [ "time", "request.method", "request.uri", "response.status", "response.size" ]

for line in sys.stdin:
  fields = line.rstrip('\r\n').split(len(cols)-1)
  values = dict([(cols[i],fields[i]) for i in xrange(len(cols))])
  record = unflatten(values)
  # do something with the record...
  print record
>>>
{ 'time': '12/12/2012 12:00:00',
  'request': { 'method': 'GET', 'uri': '/stuff/' },
  'response': { 'status': '200', 'size': '40391' } }
...

Web form processing: grouping data

Or suppose you have a web form for collecting several distinct blobs of data at once, like a payment form:

#!/usr/bin/html
<form method="post" action="">
  <div>
    First name: <input type="text" name="contact.firstname" value="" />
    Last name:  <input type="text" name="contact.lastname" value="" />
    Email:      <input type="text" name="contact.email" value="" />
  </div>
  <div>
    Credit Card Type: <select name="payment.cctype"> <option> ... </option> </select>
    Number:           <input type="text" name="payment.ccnumber" value="" />
    CCV:              <input type="text" name="payment.ccv" value="" />
    Expiration Month: <select name="payment.ccmonth"> <option> ... </option> </select>
    Expiration Year:  <select name="payment.ccyear"> <option> ... </option> </select>
  </div>
  <div>
    <input type="submit" name="submit" />
  </div>
</form>

In the form processing code, you can expand the key value form data pairs into a nested object:

#!/usr/bin/env python
from flattery import unflatten

params = formdata()   # however you get a dictionary of form data...
data = unflatten(params)
print data
>>>
{ 'contact':
    { 'lastname':'Doe',
      'firstname': 'John',
      'email':'jdoe@example.com' },
  'payment':
    { 'cctype': 'amex',
      'ccnumber': '4111111111111111',
      'ccv': '4321',
      'ccmonth': '12' ,
      'ccyear' : '2020' } }

(But be careful with multiply-valued form data.)

Web form processing: tabular data

Another web example, where a user is editing tabular data:

#!/usr/bin/html
<form method="post" action="">
  <div>
    <ul class="table">
      <ul class="row">
        <li><input type="checkbox" name="rows.0.delete" /></li>
        <li><input type="text" name="rows.0.name" value="" /></li>
        <li><input type="text" name="rows.0.email" value="" /></li>
      </ul>
      <ul class="row">
        <li><input type="checkbox" name="rows.1.delete" /></li>
        <li><input type="text" name="rows.1.name" value="" /></li>
        <li><input type="text" name="rows.1.email" value="" /></li>
      </ul>
      ...
      <ul class="row">
        <li><input type="checkbox" name="rows.99.delete" /></li>
        <li><input type="text" name="rows.99.name" value="" /></li>
        <li><input type="text" name="rows.99.email" value="" /></li>
      </ul>
    </ul>
  </div>
  <div>
    <input type="submit" name="submit" />
  </div>
</form>

(You should validate that the embedded indices are below some reasonable limit to avoid a memory DoS.)

In the form processing code:

#!/usr/bin/env python
from flattery import unflatten

params = formdata()   # however you get a dictionary of form data...
data = unflatten(params)
print data
>>>
{ 'rows':
  [
    'delete': '1',
    'name': 'John Doe',
    'email': 'johndoe@example.com',
  ],
  [
    'delete': '0',
    'name': 'Suzy Q',
    'email': 'suzyq@example.com',
  ],
  ...
  [
    'delete': '0',
    'name': 'Charlie Chaplin',
    'email': 'charliechaplin@example.com',
  ],
}

Installation

Ubuntu / Debian users:

#!/bin/sh
fakeroot ./debian/rules binary
dpkg -i ../python-flattery*.deb

If there's no "real" packaging for your system yet:

#!/bin/sh
./setup.py build_ext --inplace
./test.py
./setup.py build
./setup.py install