Introduction | FetchBookFacts |
AWS Programmer API | IsASINInStock |
Types of Links | Amazon in India |
Other Countries | Tips |
Sample Code For Amazon AWS | Links |
AwsHandlerResolver |
Amazon.com keeps a database of all its products including books, electronics and DVDs (Digital Video Discs). an enormous online bookstore that carries many Java related books. It includes pictures and prices, but not detailed descriptions. Each country has its own separate database. They sometimes use the same ASIN number for the same product. You can use this data to embed links to Amazon stores on your own webpage. When people buy products you get a small commission.
On the mindprod.com website, I use a variety of affiliate links to the Amazon.com website:
Types of Links To Amazon.com | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
text linkto a specific book |
Java Concurrency in Practice | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
image linkto a specific book |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
CMP (Canadian Mind Products) Book Macros |
|
Screen scraping will drive you insane at Amazon because the pages are formatted inconsistently, they make ever increasing use of JavaScript and they change the layouts frequently just to keep things interesting.
Thankfully, they have created a programmer’s API. called AWS or more recently the PA-API (Product Advertising Application Programming Interface). Actually there are dozens of dozens of them actually, to let you do all manner of things including getting the information about a book in computer-friendly and reasonably stable form — XML (extensible Markup Language) / SOAP (Simple Object Access Protocol). Each country has its own servers, own accounts, own passwords, own XML schema… Unfortunately for batch processes, they don’t want you to do more than 200 inquiries per hour. Oddly, the problem is too much documentation and too many samples. You are overwhelmed. Don’t try to make sense of the complicated stuff until you gave found introductory docs and mastered the basics. Read the Getting Started Guide materials first. To help avoid overloading Amazon’s servers, find a response format that includes just what you need. Amazon issues you user IDs and passwords to use on all queries so they can track what you are up to with the API. They also insist that all queries be digitally signed and sent with HTTPS (Hypertext Transfer Protocol over SSL (Secure Socket Layer)). You can sign by computing your own RFC 2104-compliant HMAC-SHA256 digital signature, or you can use Oasis WSS (Web Services Security) SOAP -Style signing, using your own X.509 certificate or one provided by Amazon. I guess they are preparing for a sophisticated DOS (Denial of Service attack) attack.
To use it you will need:
Convert the *.wsdl to Java source code with:
You don’t need AWS SDK (Software Development Kit). All you need is AWSECommerceService.wsdl, and Java’s built-in JAX features.
There is a utility called the Amazon Product Advertising API Signed Requests Helper that helps you understand how queries are constructed. It sorts the parameters alphabetically, adds a timestamp and digitally signs the request to form a GET URL (Uniform Resource Locator). It is a little out of date. The URLs it uses do not match those in the JAX spec. Here is what a typical query looks like:
Fields to use in your inquiries you may find of interest include:
Fields of interest in the responses include:
You do not directly specify which fields you can in response, but you can specify constellations of names called ResponseGroups.
The XML response has a complex tree structure, it is not just a matter of requesting the field you want.
You can batch up to ten queries in on request.
They also have SDK s for half a dozen languages. I don’t know much about it.
Country | Port Name | Endpoint |
---|---|---|
Canada | https://webservices.amazon.ca/onca/soap | AWSECommerceServicePortCA |
China | https://webservices.amazon.cn/onca/soap | AWSECommerceServicePortCN |
France | https://webservices.amazon.fr/onca/soap | AWSECommerceServicePortFR |
Germany | https://webservices.amazon.de/onca/soap | AWSECommerceServicePortDE |
Great Britain | https://webservices.amazon.co.uk/onca/soap | AWSECommerceServicePortUK |
Italy | https://webservices.amazon.it/onca/soap | AWSECommerceServicePortIT |
Japan | https://webservices.amazon.jp/onca/soap | AWSECommerceServicePortJP |
Spain | https://webservices.amazon.es/onca/soap | AWSECommerceServicePortES |
USA | https://webservices.amazon.com/onca/soap | AWSECommerceServicePort or AWSECommerceServicePortUS |
You don’t use SNI (Server Name Indication). You must have Symantec Class 3 Secure Server CA - G4 root cert installed in your JVM (Java Virtual Machine).
The web is littered with obsolete lists of the endpoint URLs. You can find the most recent list inside AWSECommerceService.wsdl. You don’t use the endpoints directly unless you are experimenting with Amazon Product Advertising API Signed Requests Helper. Otherwise you use the corresponding port getter e.g. AWSECommerceService.getAWSECommerceServicePortCA() or AWSECommerceServicegetAWSECommerceServicePort() ; and JAX looks effectively looks up the corresponding end point in AWSECommerceService.wsdl (using pre-generated code). Note that Amazon in Spain (Amazon.es) and Amazon in India (Junglee.com) originally did not support AWS, but now it does do. Unlike the other Amazon stores, Junglee will not let you register as an affiliate unless you are resident of India. I suspect they will not let you probe their AWS database either. Amazon in Australia (Amazon.com.au), Amazon in Brazil (Amazon.com.br) and Amazon in Mexico (Amazon.com.mx) still do not. Amazon in Australia and Amazon in Mexico have no paper books or electronics, just kindles and kindle eBooks. I don’t include them on my site book referrals.
It took weeks to get this little sample program to use the Amazon AWS API working. Even the Signed Requests Helper utility took days to get working. The problem is Amazon makes their API more and more complicated each year, but do not update the documentation and they refuse to provide complete example code. I have never in my career of 55 years seen a more incompetently documented API. Further, the API is not consistent. Further, it is often idiotic, e.g. returning the number of results as a BigInteger when even a short would have done fine.
It uses SSL (Secure Sockets Layer) but not SNI. You must have the root certificate for Symantec Class 3 Secure Server CA - G4 installed in your cacerts.
I downloaded this AwsHandlerResolver code from the Amazon website, then modified it to use the more compact CMP Base64 classes. AwsHandlerResolver handles the poorly documented stuff like adding a timestamp, digitally signing your request and Base64 ASCII-armouring the signature.
The following is a more complicated example. You give it an ISBN-13 (aka EAN (European Article Number) ) on the command line and it probes the Amazon API to find the title, author, publisher, published date. It also probes for the ISBN-13s of alternate bindings, e.g. Paperback, Hardcover, Kindle, Audio. It creates a text file (in the form of an HTML (Hypertext Markup Language) macro) summarising this information and downloads the large bookcover image. It does not include the author birth/death dates since Amazon does not track them. On my website, those fields are added later by code part of HTMLMacros.
The following takes either an isbn13 or ASIN and probes the Amazon.com AWS API to find out if the product is in stock. The results do not correspond well with the information on the website. I don’t know which is actually correct.
Amazon in India is puzzling. Instead of being called Amazon.in they called it Junglee.com. You could not probe its inventory with AWS using getAWSECommerceServicePortIN even though that port was defined in the to AWSECommerceService.wsdl file. You had to screen scrape the consumer interface. You could not sign up as an affiliate unless you lived in India. I get perhaps a third of my traffic from India, but I was still excluded. To my surprise, Junglee had the largest inventory of any Amazon store.
Then on 2014-08-23 a new, separate website appeared called Amazon.in alongside Junglee.com without any explanation of its relationship to Junglee.com.
I tried to contact Amazon.in via their website for clarification. They refused to talk to me unless I provided them with a fake cell number. They asked me a number of impertinent questions, then to cap it off gave me no space to pose my questions. I am soured on them already.
Convert the *.wsdl to Java source code with:
This page is posted |
http://mindprod.com/jgloss/amazonproductadvertising.html | |
Optional Replicator mirror
|
J:\mindprod\jgloss\amazonproductadvertising.html | |
Please read the feedback from other visitors,
or send your own feedback about the site. Contact Roedy. Please feel free to link to this page without explicit permission. | ||
Canadian
Mind
Products
IP:[65.110.21.43] Your face IP:[18.217.17.242] |
| |
Feedback |
You are visitor number | |