Can't get Metadata from Amazon

[HarryT]

Originally Posted by kiwidude

@HarryT - you could just install another metadata download plugin like the B&N one - which from my experience 99 times out of 100 has a better quality cover than what you get from Amazon.

Thanks!

[chilady1]

Originally Posted by kiwidude

@HarryT - you could just install another metadata download plugin like the B&N one - which from my experience 99 times out of 100 has a better quality cover than what you get from Amazon.

I think I am going to do the same thing and change the main metadata source because I have been experiencing intermittent issues with Amazon's metadata download also. There is no rhyme or reason to it and there is no one time when it is good vs. bad.

[Synamon]

Originally Posted by HarryT

I've just retried downloading covers for the same 6 books that failed last Sunday, and they've all worked fine this time.

Is there any way, by the way, to add additional sources (eg "amazon.co.uk") to the list of sites that are searched for covers?

You probably already know that you can change the Amazon.com source to use the UK site, but for anyone else here's how: Preferences > Metadata download, chose Amazon.com then "configure selected source" and change the "Amazon website to use" at the bottom of the page.

[HarryT]

Originally Posted by Synamon

You probably already know that you can change the Amazon.com source to use the UK site, but for anyone else here's how: Preferences > Metadata download, chose Amazon.com then "configure selected source" and change the "Amazon website to use" at the bottom of the page.

No, I didn't know that. Thank you very much indeed for the useful tip!

[G-Mac]

Harry: A workaround. Go into Amazon, and copy the ASIN number for the book and plug it into Calibre. This seems to let it find the cover.

Hi, I have the ASIN number but I'm unsure where in Calibre to enter it. I've tried the ID section but don't know if it needs a prefix?

I've been trying for days now to send non amazon content to user@kindle.com but when it's uploaded to 'my library' it wont let me send to my other devices (sends direct to kindle no problem) and I'm hoping that the ASIN number will fix this?

[CWatkinsNash]

Originally Posted by G-Mac

Hi, I have the ASIN number but I'm unsure where in Calibre to enter it. I've tried the ID section but don't know if it needs a prefix?

It should be entered in the ID section in the following format:

amazon:B001234ZYX

-----------------------------

Seems today things are a little different - it's getting data from Amazon, but when a book is available in both e- and paper formats, it seems to be only getting matches from the paper versions. This wouldn't be so bad were it not for the fact that, apparently due to layout changes for paper book product pages, the plugin can no longer consistently scrape the product description section. It seems to either only grab the "about the author" if there is one, or nothing at all if there's only a "product description" section.

I was going to try to enter the ASIN manually for these books, but if I'm already on the product page to get the ASIN I might as well just copy what I need while I'm there.

[kovidgoyal]

@CWatkinsNash: Post an example of a book (title, author) for which that happens.

[CWatkinsNash]

Here's (sort of) an example. Griffin's Destiny: Book Three: The Griffin's Daughter Trilogy by Leslie Ann Moore.

http://www.amazon.com/dp/0982514506/ref=cm_sw_su_dp

...is the result in the metadata search. You can see that there is a product description, but the text in the comments on the metadata results screen is only the "about the author" text from the section below "Customers who viewed this..." titled "Editorial Reviews".

Earlier I mistakenly said "about the author" section, when in fact that was actually coming from this "Editorial Review" section. (Apparently it is used for that purpose a lot, which is why I kept seeing "about the author" text.)

If I put in the ASIN ID manually, it will pull the Kindle version result, which gets all the data because ALL of the info for the Kindle version is in the "Editorial Reviews" section - there is no "Product Description" in the page for that version:

http://www.amazon.com/Griffins-Desti...=AG56TWVU5XWC2

[kovidgoyal]

Amazon appears to be suffering from major schizophrenia. They're returning different versions of the same page, I'd guess based on user agent sniffing. See attached screenshots. One page uses javascript and the other doesn't. The same thing is happening with the results page, and explains why the downloads sometimes work and sometimes dont work (calibre uses a randomized user agent). Sigh, user agent sniffing, in 2012! Now to figure out what user agents prompt amazon to use javascript and get the plugin to not use them.

[theducks]

I noticed in one case(#2) you are signed in. Might it be that simple.

I had a similar case on another forum, that would not show me older forum messages until I signed in.

[kovidgoyal]

I doubt it. I can think of no reason to change the underlying technology of the page depending on whether a user is signed in or not. Here's a screenshot showing the different pages with me signed in for both. I dont have the time to investigate this is detail right now, but I really hope it is something explicable, like UA sniffing and not just amazon's servers being flaky.

[CWatkinsNash]

Wow, that's like trying to hit a moving target.

Related query - is there any way to try to force the Amazon metadata plugin to look for an ebook result first? As I mentioned in a previous post, it seems that whenever a paper version is available, it automatically grabs that result. For the purposes of ebook management, the ideal would be for the priority to be reversed.

I don't know if this is strictly something that depends on what Amazon spits out, or if something can be helped by changes to the plugin, but since we're on the topic of Amazon metadata this seemed like the time to ask about that.

I also don't know if it's been this way the whole time, because I didn't pay that much attention to the returned IDs until the other metadata issues started.

[kovidgoyal]

After half a day struggling with amazon's servers, I think I have this licked. calibre should now be able to handle both variants of page as well as the random broken markup that some amazon servers seem to send from time to time. Of course, given how messy the workarounds are, I cannot be sure...try the next release and see.

[HarryT]

Thanks, Kovid!

[CWatkinsNash]