Skip to content

A super awesome Python wrapper for the Amazon Product Advertising API.

License

Notifications You must be signed in to change notification settings

patrickdessalle/bottlenose

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Bottlenose

Version License Versions

Bottlenose is a thin, well-tested, maintained, and powerful Python wrapper over the Amazon Product Advertising API. There is practically no overhead, and no magic (unless you add it yourself).

Before you get started, make sure you have both Amazon Product Advertising and AWS accounts (yes, they are separate -- confusing, I know).

Features

  • Compatible with Python versions 2.4 and up
  • Support for CA, CN, DE, ES, FR, IN, IT, JP, UK, and US Amazon endpoints
  • No requirements, except simplejson for Python pre-2.6
  • Configurable query parsing
  • Configurable throttling for batches of queries
  • Configurable query caching
  • Configurable error handling and retry

Communication

  • If you need help or would like to ask a general question, use Stack Overflow. Apply the 'bottlenose' tag to your question to get help faster.
  • If you found a bug or have a feature request, open an issue.
  • If you want to contribute, submit a pull request. If it's a big change, please open an issue first to discuss implementation.

Usage

1. Available Search Methods:

Required
amazon = bottlenose.Amazon(AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_ASSOCIATE_TAG)
Search for a Specific Item
response = amazon.ItemLookup(ItemId="B007OZNUCE")
Search for Items by Keywords
response = amazon.ItemSearch(Keywords="Kindle 3G", SearchIndex="All")
Search for Images for an item
response = amazon.ItemLookup(ItemId="1449372422", ResponseGroup="Images")
Search for Similar Items
response = amazon.SimilarityLookup(ItemId="B007OZNUCE")

2. Available Shopping Related Methods:

Required
amazon = bottlenose.Amazon(AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_ASSOCIATE_TAG)
Create a cart
response = amazon.CartCreate(...)
Adding to a cart
response = amazon.CartAdd(CartId, ...)
Get a cart by ID
response = amazon.CartGet(CartId, ...)
Modifying a cart
response = amazon.CartModify(ASIN, CartId,...)
Clearing a cart
response = amazon.CartClear(CartId, ...)

3. Sample Code

import bottlenose
amazon = bottlenose.Amazon(AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_ASSOCIATE_TAG)
response = amazon.ItemLookup(ItemId="0596520999", ResponseGroup="Images",
    SearchIndex="Books", IdType="ISBN")
print(response)
# <?xml version="1.0" ?><ItemLookupResponse xmlns="http://webservices.amazon...

Here is another example.

response = amazon.ItemSearch(Keywords="Kindle 3G", SearchIndex="All")
# <?xml version="1.0" ?><ItemSearchResponse xmlns="http://webservices.amazon...

Bottlenose can also read your credentials from the environment automatically; just set $AWS_ACCESS_KEY_ID, $AWS_SECRET_ACCESS_KEY and $AWS_ASSOCIATE_TAG.

Any valid API call from the following is supported (in addition to any others that may be added in the future). Just plug in appropriate request parameters for the operation you'd like to call, and you're good to go.

BrowseNodeLookup
CartAdd
CartClear
CartCreate
CartGet
CartModify
ItemLookup
ItemSearch
SimilarityLookup

You can refer here for a full listing of API calls to be made from Amazon.


For more information about these calls, please consult the Product Advertising API Developer Guide.

Parsing

By default, API calls return the response as a raw bytestring. You can change this with the Parser constructor argument. The parser is a callable that takes a single argument, the response as a raw bytestring, and returns the parsed response in a format of your choice.

For example, to parse responses with BeautifulSoup:

import bottlenose
from bs4 import BeautifulSoup

amazon = bottlenose.Amazon(
    AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_ASSOCIATE_TAG,
    Parser=lambda text: BeautifulSoup(text, 'xml')
)
results = amazon.ItemLookup(ItemId="0198596790", ResponseGroup="SalesRank")

print(results.find('SalesRank').string)
# 168088

Throttling/Batch Mode

Amazon strictly limits the query rate on its API (by default, one query per second per associate tag). If you have a batch of non-urgent queries, you can use the MaxQPS argument to limit them to no more than a certain rate; any faster, and bottlenose will sleep() until it is time to make the next API call.

Generally, you want to be just under the query limit, for example:

amazon = bottlenose.Amazon(MaxQPS=0.9)

If some other code is also querying the API with your associate tag (for example, a website backend), you'll want to choose an even lower value for MaxQPS.

Caching

You can often get a major speedup by caching API queries. Use the CacheWriter and CacheReader constructor arguments.

CacheWriter is a callable that takes two arguments, a cache url, and the raw response (a bytestring). It will only be called after successful queries.

CacheReader is a callable that takes a single argument, a cache url, and returns a (cached) raw response, or None if there is nothing cached.

The cache url is the actual query URL with authentication information removed. For example:

http://webservices.amazon.com/onca/xml?Keywords=vacuums&Operation=ItemSearch&Region=US&ResponseGroup=SearchBins&SearchIndex=All&Service=AWSECommerceService&Version=2013-08-01

Example code:

def write_query_to_db(cache_url, data):
    ...

def read_query_from_db(cache_url):
    ...

amazon = bottlenose.Amazon(CacheWriter=write_query_to_db,
                           CacheReader=read_query_from_db)

Note that Amazon's Product Advertising API Agreement only allows you to cache queries for up to 24 hours.

Error Handling

Sometimes the Amazon API returns errors; for example, if you have gone over your query limit, you'll get a 503. The ErrorHandler constructor argument gives you a way to keep track of such errors, and to retry queries when you receive a transient error.

ErrorHandler should be a callable that takes a single argument, a dictionary with these keys:

  • api_url: the actual URL used to call the API
  • cache_url: api_url minus authentication information
  • exception: the exception raised (usually an HTTPError or URLError)

If your ErrorHandler returns true, the query will be retried. Here's some example code that does exponential backoff after throttling:

import random
import time
from urllib2 import HTTPError

def error_handler(err):
    ex = err['exception']
    if isinstance(ex, HTTPError) and ex.code == 503:
        time.sleep(random.expovariate(0.1))
        return True

amazon = bottlenose.Amazon(ErrorHandler=error_handler)

License

Apache License, Version 2.0. See LICENSE for details.

About

A super awesome Python wrapper for the Amazon Product Advertising API.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 98.2%
  • Shell 1.3%
  • Makefile 0.5%