dalton 0.2

An httplib injection library for recording and playing back HTTP interactions.

“I want you to be nice, until it’s time to not be nice.”

– Dalton, Road-House

An httplib injection library for recording and playing back HTTP interactions.

Dalton monkey-patches two methods of httplib’s HTTPConnection class to intercept request/response interactions, and can play them back based on a Dalton recording. To ease testing, the recording is generated Python code to ease in customization of the response and to allow for branches in the playback path.

Monkey-patched methods of HTTPConnection:

• request

• getresponse

Using the more verbose method to send/recieve requests with HTTPConnection is not supported at this time.

Note: This is a first and early release, mainly so that I could use it with mechanize to record/playback interactions. As mechanize only uses the request/getresponse API on HTTPConnection, I have no interest in adding intercept to the rest. Please feel free to fork this to add additional features as I don’t plan on adding them myself (though I will happily pull bug fixes and feature additions with unit tests).

Warning: Dalton uses inspect.currentframe magic to derive the caller which may only work on CPython (PyPy and Jython is untested).

Example

Since dalton monkey-patches httplib, no modification is necessary of libraries that utilize the supported methods.

import dalton
dalton.inject() # monkey-patch httplib

from httplib import HTTPConnection
h = HTTPConnection('www.google.com')

# when recording, httplib capture is restricted by caller
recorder = dalton.Recorder(caller=h)

# record httplib calls in this block
with recorder.recording():
    h.request('GET', '/')
    resp = h.getresponse()
    body = resp.read()

# save the interaction
recorder.save('google')

A folder called google will be created in the current directory for use with dalton’s playback facility.

Playing it back:

import dalton
dalton.inject() # monkey-patch httplib

from httplib import HTTPConnection
h = HTTPConnection('www.google.com')

# load the player
player = dalton.Player(caller=h, playback_dir='google')

# run httplib calls against the player
with player.playing():
    h.request('GET', '/')
    resp = h.getresponse()
    body = resp.read()

# body is now the same as it was recorded, no calls to www.google.com
# were made

This generates a directory google with the following layout:

• __init__.py

• step_0_response.txt

The contents of __init__.py contain the following generated playback information:

import os
import dalton
from dalton import FileWrapper

here = os.path.abspath(os.path.dirname(__file__))

class StepNumber0(object):
    recorded_request = {
        'headers':  {},
        'url': '/',
        'method': 'GET',
        'body': None,
    }
    recorded_response = {
        'headers':  [('x-xss-protection', '1; mode=block'),
                     ('transfer-encoding', 'chunked'),
                     (                    'set-cookie',
                                          'PREF=ID=ff; expires=Thu, 11-Apr-2013 20:19:35 GMT; path=/; domain=.google.com, NID=45=fU; expires=Wed, 12-Oct-2011 20:19:35 GMT; path=/; domain=.google.com; HttpOnly'),
                     ('expires', '-1'),
                     ('server', 'gws'),
                     ('cache-control', 'private, max-age=0'),
                     ('date', 'Tue, 12 Apr 2011 20:19:35 GMT'),
                     ('content-type', 'text/html; charset=ISO-8859-1')],
        'body': FileWrapper('step_0_response.txt', here),
        'status': 200,
        'reason': 'OK',
        'version': 11,
    }
    next_step = 'None'

    def handle_request(self, request):
        assert dalton.request_match(request, self.recorded_request)
        return (self.next_step, dalton.create_response(self.recorded_response))

This file can be modified after recordings to customize the playback, add additional branches, etc.