Hi.

I try to create a metadata download plugin for a Hungarian online shop called libri.hu.
First I have made a python script that downloads the metadata and print it in the console.
The I try to convert it to a plugin for Calibre, but I stucked.
In debug mode I can see it prints out the data for the book but at the end of the run of the plugin I get an error message:
Traceback (most recent call last):
File "threading.py", line 525, in __bootstrap_inner
File "D:\Calibre Libary\calibre\src\calibre\gui2\dialogs\fetch_meta data.py", l
ine 38, in run
File "D:\Calibre Libary\calibre\src\calibre\ebooks\metadata\fetch.p y", line 32
0, in search
File "D:\Calibre Libary\calibre\src\calibre\ebooks\metadata\fetch.p y", line 28
7, in filter_metadata_results
AttributeError: 'str' object has no attribute 'publisher'

I'm new in python and I have just basic programing skills.
My code is now just a basic scratch version, and I try to do fine tuning and make some error handling in the future.
The metadata it scrapes mostly contains special Hungarian characters.
Here is my plugins code:
# -*- coding: utf-8 -*-

import lxml, sys
import lxml.html as lh
#import traceback

from calibre import browser
from calibre.customize import Plugin
from calibre.ebooks.metadata.fetch import MetadataSource
from calibre.ebooks.metadata import MetaInformation

def get_social_metadata(title, authors, publisher, isbn):
print '4'
mi = MetaInformation(title, authors)
SEARCH_URL = "http://www.libri.hu/reszletes_kereso"
print '5'
br = browser()
br.open(SEARCH_URL)
br.select_form(name = "detailed_search_form")
br['cim'] = title
br['szerzo'] = authors
# br['isbn'] = isbn
search_page = br.submit()

book_page = br.follow_link(url_regex='konyv', nr=0)

print search_page.geturl()
print book_page.geturl()

doc = lh.parse(book_page)

book_title = ''.join(doc.xpath('//*[@id="book"]/div/h1/text()')).replace("\t",'').replace("\n",'')
print book_title
mi.title = book_title

book_author = ''.join(doc.xpath('//*[@id="book"]//*[@class="authors"]//text()')).replace("\t",'').replace("\n",'')
print book_author
mi.authors = book_author

book_property = doc.xpath('//*[@class="props"]/text()')
book_property = [text for text in book_property if text != '\r\n\t\t\t\t\t\r\n\t\t\t' and text != '\r\n\t\t\t\t' and text != '\r\n\t\t\t\t\r\n\t\t\t\t\t\r\n\t\t\t\t' and text != ' ' and text != ': ']

book_publish_date = ''.join(book_property[6]).replace(" ",'').replace("\t",'').replace("\n",'')
print book_publish_date
mi.pubdate = book_publish_date

book_publisher = ''.join(doc.xpath('//*[@class="props"]/a/text()'))
print book_publisher
mi.publisher = book_publisher

#mi.series = series[0].strip()

book_comments = ''.join(doc.xpath('//*[@id="tab_content_lead"]//text()')).replace("\t",'').replace("\n",'')
print book_comments
mi.comments = book_comments

book_isbn = ''.join(book_property[5]).replace(" ",'').replace("\t",'').replace("\n",'')
print book_isbn
mi.isbn=book_isbn

return mi

class LibraryThing(MetadataSource):

author = 'Csaba Hoffer'
metadata_type = 'basic'
name = 'Libri_Hu'
description = _('Downloads metadata from Libri.hu')
version = (1, 0, 0)
supported_platforms = ['windows', 'osx', 'linux']

print '1'

def fetch(self):
print '2'
try:
self.results = get_social_metadata(self.title, self.book_author, self.publisher, self.isbn)
print '8'
#print self.results
print '9'

except Exception, e:
print '6'
self.exception = e
self.tb = traceback.format_exc()

def main(args=sys.argv):
print get_social_metadata(None, None, None, None)
return 0

if __name__ == '__main__':
sys.exit(main())

Please, if anyone have some hint for me what could be the problem share it with me.:help:
Thanks!

Welcome to MobileRead, Daermond

Help should arrive soon.

decide if your plugin is used to download social metadata or basic metadata. And note that self.results has to be set to a list of Metadata objects as noted in the documentation of the MEtadataSource class

Thanks Kovid!:thanks:
I have missed that self.results should be a list object.
Now it working! :)
Thanks for the tip!

It is possible that one plugin could used for basic and social metadata downloading?

no you have to make two

Hi.
The plugin downloads both type of metadata...

I have an other problem. I use the latest 0.7.24 version on WinXP and on Win7 too. When I try to fetch metadata from edit metadata option it do the job fine. Downloads all the basic and social metadata. But when I try to download metadata from the main page right click on the book and edit metadata and download metadat (all the options) does not find the same book. When I run it in debug mode it seems that calibare does not give the data to self. variables. I try to print the self.title and the debug console says None. It is a plugin problem or calibare problem?

By the way thanks for the whole Calibre!

That's weird, open a ticket for it.

I have found the problem.
The matadata type metadata_type = 'basic' was not defined in the MetadataSource class.
It is now set, and working fine both ways.
So these was the plugin faliure.