doze

Lazy URL construction


License
BSD-3-Clause
Install
pip install doze==0.5

Documentation

Doze

Build Status

Doze is a URL building package inspired by slumber

This project attempts to take what is awesome about slumber while enabling more flexibility.

Usage

We will demonstrate how to fetch data from the twitter API:

import requests
from doze import url

twitter = url("http://api.twitter.com/1/")
user_lookup = twitter.users("lookup.json")

# generates "http://api.twitter.com/1/lookup.json?screen_name=ericmoritz"
response = requests.get(str(user_lookup(screen_name="ericmoritz")))

if response.status_code == 200:
    data = response.json
else:
    data = None

While this code is a bit more verbose it give us much more flexibility and control over the fetching and serialization of the data.

For instance if we wanted to add a object_hook to the json.loads:

if response.status_code == 200:
    data = json.loads(response.content,
                       object_hook=custom_object_hook)
else:
    data = None

You can also reuse user_lookup to compose multiple requests:

import requests
from doze import url

twitter = url("http://api.twitter.com/1/")
user_lookup = twitter.users("lookup.json")

# Fetch the user_lookup data for each screen name
users = []
for screen_name in ["ericmoritz", "montylounge"]:
    url = str(user_lookup(screen_name=screen_name))
    response = requests.get(url)
    if response.status_code == 200:
        users.append(json.loads(response.content))

If you find the dotted notation awkward you can pass the path items as arguments and the url params as kwargs:

from doze import url

endpoint = url("http://example.com/v1")

# dotted way
def sections(section_id):
    # http://example.com/v1/sections/{section_id}?page=2
    url = str(endpoint.sections(section_id, page=2))
    return requests.get(url).json

# "functional way"
def sections(section_id):
    # http://example.com/v1/sections/{section_id}?page=2
    url = str(endpoint("sections", section_id, page=2))
    return requests.get(url).json

Changelog

v0.3

Backwards compatible change: duplicate query parameters overwrite previous values with later values.

# old behavior
url = endpoint(page=3)
url = endpoint(page=4)
assert "http://example.com/v1?page=3&page=4" == str(url)

# new behavior
assert "http://example.com/v1?page=4" == str(url)