New metadata API in 0.8 questions

[kovidgoyal]

Yeah the pubdate is looked up via xisbn during the merging process and the earliest pubdate is used.

Note that merging of results means that it is hopeless to try to use a "particular edition" in the new code. If you want details for a particular edition, you are going to have to enter them manually.

Take a look at the merging code it isn;t too mysetrious (I hope) its in sources.identify

[kovidgoyal]

I've committed code to merge results if they have a common identifier. Its turned off for now as I dont feel like testing it. If you decide you want it, turn it on and see how it goes.

[kiwidude]

Cool, will do. I've just been trying not to get too far from the sides of the pool

A question. Would I be correct in thinking that there is no access to the db from the configuration screen? I was going to reuse the multicomplete combo with a list of the tags in the database but methinks the best I could do is build a list from my own default tags in the plugin and the user manually replaces them with their bionic memories.

Any chance of a gui or db parameter coming through to config_widget()? I can appreciate you don't want it (ab)used by the plugin with the threading issues, but this is just for the configuration...

[kovidgoyal]

Sigh...

Just committed,

calibre.gui2.ui.get_gui()

[kiwidude]

Originally Posted by kovidgoyal

Sigh...

Just committed,

calibre.gui2.ui.get_gui()

Go on, you love it when I ask for something new, really...

Thanks dude.

[kovidgoyal]

ON second thoughts, I added a get_current_db method that is more robust if all you want is access to the current db

from calibre.gui2 import get_current_db

[ldolse]

I'm having a very weird problem with ISBN parsing/populating metadata. Namely it causes the download to fail even though it's successful.

Aaside from saying the download failed, all the metadata was clearly successfully downloaded, including the ISBN, which show up as populated in the debug log

Here's an example:

Starting job: Download metadata for 1 books 
Running identify query with parameters: 
{u'authors': [u'Angie Sage'], u'identifiers': {}, u'timeout': 30, u'title': u'The Magykal Papers'} 
Using plugins: Overdrive 
The log from individual plugins is below 

****************************** Overdrive ****************************** 
Request extra headers: [('User-agent', 'Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.11) Gecko/20101012 Firefox/3.6.11')] 
Found 1 results 
Downloading from Overdrive took 9.11680293083 


--- 
Title               : The Magykal Papers
Author(s)           : Angie Sage
Publisher           : HarperCollins
Tags                : Fiction, Juvenile Fiction, Science Fiction & Fantasy
Series              : Septimus Heap #1
Language            : English
Published           : 2009-06-30T16:00:00+00:00
Identifiers         : overdrive:2fd07b37-0944-4f58-91e4-5331ac663f46, isbn:9780061919923
Comments            : 
Enter the world of Septimus Heap with this collection of previously unpublished papers. 
This rich compendium includes: 
     The private journals of Septimus, Jenna, and Marcia Overstrand.  
     The best–and worst–places to eat as described in The Egg-on-Toast Restaurant Guide.  
     Sirius Weazal's Speedy Guides to the Palace, the Wizard Tower, and Wizard Way.  
     Excerpts from the Pigeon Post Biography series and the Heaps of History series.  
     Alther Mella's Guide to Being Dead: Ten Handy Rules for New Ghosts.  
     Beautiful maps, quirky flyers, funny letters, and much more!  
 

******************************************************************************** 
The identify phase took 9.17 seconds 
The longest time (9.116803) was taken by: Overdrive 
Merging results from different sources and finding earliest publication dates 
Failed to download metadata for The Magykal Papers

****************************** Overdrive Covers ****************************** 
Request extra headers: [('User-agent', 'Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.11) Gecko/20101012 Firefox/3.6.11')] 
Downloaded cover: 510x680 
Took 18.3445880413 seconds 
No cached cover found, running identify
Downloading cover from: http://images.contentreserve.com/ImageType-100/0293-1\{2FD07B37-0944-4F58-91E4-5331AC663F46}Img100.jpg 

******************************************************************************** 
Download complete, with 1 failures

Here's the specific lines of code which actually scrapes/populates the ISBN field:

        ebook_isbn = root.xpath("//td/label[@id='ctl00_ContentPlaceHolder1_lblIdentifier']/text()")

        if ebook_isbn:
            print "ebook isbn is "+str(ebook_isbn[0])
            isbn = check_isbn(ebook_isbn[0].strip())
            if isbn:
                self.cache_isbn_to_identifier(isbn, ovrdrv_id)
                mi.isbn = isbn

The ISBN appears to be successfully populated based on the debug logs - I have no idea why it 'fails'.

If I comment that out the overall metadata download succeeds.

Also when this fails none of the metadata from this source is used, including the cover which is successfully downloaded from a separate function - is this intentional?

I'm wondering if the failure has to do with the merging of that ISBN, since this an ebook isbn which isn't represented in all the ISBN databases... If I make an xisbn request this seems to be supported there at least:

http://xisbn.worldcat.org/webservices/xid/isbn/9780061919923?method=getEditions&format=json&fl=form,year,lang,ed

{
 "stat":"ok",
 "list":[{
	"isbn":["9780061919923"],
	"year":"2009",
	"form":["BA",
	 "DA"],
	"lang":"eng"},
  {
	"ed":"1st ed.",
	"isbn":["9780061704161"],
	"year":"2009",
	"form":["BA"],
	"lang":"eng"},
  {
	"isbn":["9781408800560"],
	"year":"2009",
	"form":["BB"],
	"lang":"eng"},
  {
	"isbn":["9782226193643"],
	"year":"2009",
	"form":["BA"],
	"lang":"fre"}]}

[kovidgoyal]

It certainly looks like it is failing in the merge stage.Can you send me your plugin and I will debug.

[ldolse]

There's a merge request with the plugin in my branch.