[Metadata Source Plugin] Barnes & Noble

[kiwidude]

Changes in this release:

• Add option to append TOC from website Features tab to the comments field (available on B&N Textbooks)

[darthyoda6]

Thanks, 'cause they also have TOC's on computer books too.

[kiwidude]

Changes in this release:

• Respond to change to website layout which prevented metadata download working

[kiwidude]

Changes in this release:

• When an ISBN is not directly found, process the search results page

[kiwidude]

Not that anyone else has reported this yet, but it looks like B&N have completely redesigned their website. Unsurpringly this plugin won't work any more. I will start looking at a rewrite now.

[kiwidude]

Ok, given that B&N have made two more structural changes since my last post and the current state of play is an inconsistent mess depending on the book I'm going to give this a day or two to settle down before I publish the next version. I don't know how they do their release process but it clearly isn't anything close to what I would do with a commercial app.

One minor unfortunate impact of their changes is that the plugin now has to do an extra hop to get the large image (if it exists). Maybe they did that intentionally to discourage robots etc. However in its current state it also means that when you click on the cover of a book on their detail page you get a popup window that is blank when they have no image. Utterly crap, so much for their testing process.

[kiwidude]

Attached is a version which works with the B&N website after its revamp. As I mentioned above there is a few things broken on their website that I am not sure if/when they will fix so I want to give it a little more time before I officially release this.

From my quick testing, I "think" the plugin is fully working. There are some changes forced by their website changes. For instance B&N now only publish ratings in the web page rounded to a half star based system, rather than a numerical value like they had previously.

Also I am completely unable to get ratings for textbooks (fiction books are fine). Ratings on textbooks are rendered dynamically using javascript to allow hovering to change the rating etc. If someone fancies themselves a challenge then by all means please tell me where I can scrape the "starting value" from in the page as I lost the will to live rather quickly. I am sure others would be grateful if you can figure it out

As an example, go to this url:
http://www.barnesandnoble.com/w/c-pr...han/1000055175
You can see it has a 4.5 star rating on the right-side - as for how it gets that value I welcome the answer.

Regardless, a version for you to test is attached to this post, let me know of any issues before I officially release it.

[chaley]

xpath something like

"//span[@class="rating_widget"]/span/@class"

or

"//span[contains(@class, "section_updateRating")]/@class"

These will return a string containing "avg-NNN section_updateRating" where NNN is the rating. For the C Programming book, the string is "avg-4h section_updateRating", where the 'h' means 1/2. For another book I see "avg-3 section_updateRating" and it shows 3 stars.

[kiwidude]

@chaley - thx very much, I knew someone could figure it out. I had the content "folded" in notepad++ and had never noticed that inner span.

New plugin version attached to my post above.

[chaley]

Originally Posted by kiwidude

@chaley - thx very much, I knew someone could figure it out. I had the content "folded" in notepad++ and had never noticed that inner span.

Yea, that is a problem. The first thing I do when working with html and xpath is paste the source into an html editor and use its reformat function to prettify it. Without that, the html almost always completely unreadable.

[Jeff_R]

So far, the updated beta is working quite well. I have switch my metadata sources exclusively over to B&N and Goodreads, and the quality of the notes and covers has been incredible!

[kiwidude]

Changes in this release:

• Rewritten to support new B&N website

Thx Jeff for the feedback. Things seem to have stablised the last few days so hopefully they are "done" making changes for now.

Until the next time they break things...

[kiwidude]

Changes in this release:

• Support additional noresults url location after rewrite when lookup by ISBN
• Alter the details URL looked up to prevent an infinite loop on some books due to B&N website error
• If the main format returned is not acceptable (e.g. Audiobook) look for an "Also Available As:" section
• Reorder priority of matching results to those with shortest titles (to de-prioritise box sets)
• Strip '?' from title based lookups
• For non ascii names, ensure the comparison is done with non-asii equivalents

As I wrote on the FantasticFiction update the above changes have come from the opportunity for some "hands-on usage" over the last few days. This should increase both the number and quality of matches against the B&N website for this plugin.

[Caleb666]

Hi kiwidude,

For some reason this plugin downloads the "Details" cover instead of the "Front" one.

For example, if you look up the ISBN 9781593080778 you get to: http://search.barnesandnoble.com/Ori...&pos=-1&ugrp=2 (or this one, ISBN: 9781593082079)

Click on the cover and it will open a new page with the "Details" cover selected, which is useless... but you can choose "Front" and then it will show a good cover.

[kiwidude]

@Caleb666 - yes I had noticed a few of those "odd" covers coming through, you can understand why when for most books there is just a single cover and hence the one that appears on the details page by default is the one it should download.

Looks like I will have to investigate some more complex logic to try to detect this situation and download the desired "front" cover - it is on the list to take a look at sometime. Thanks for the detailed info, requests this detailed make life much easier!