EbookFoundation
/
regluit
mirror of https://github.com/EbookFoundation/regluit.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

add unt and ub

pull/94/head
eric 2020-07-31 16:18:07 -04:00
parent 9d48f8a081
commit 48af441a27
1 changed files with 19 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
core/loaders/harvest.py
View File

@ -70,6 +70,8 @@ def harvesters(ebook):
yield ebook.url.find('digitalcommons.usu.edu') > 0, harvest_usu yield ebook.url.find('digitalcommons.usu.edu') > 0, harvest_usu
yield ebook.provider == 'libros.fahce.unlp.edu.ar', harvest_fahce yield ebook.provider == 'libros.fahce.unlp.edu.ar', harvest_fahce
yield ebook.provider == 'fedoabooks.unina.it', harvest_fedoabooks yield ebook.provider == 'fedoabooks.unina.it', harvest_fedoabooks
yield ebook.provider == 'digital.library.unt.edu', harvest_unt
yield ebook.provider == 'diposit.ub.edu', harvest_ub
def ebf_if_harvested(url): def ebf_if_harvested(url):
onlines = EbookFile.objects.filter(source=url) onlines = EbookFile.objects.filter(source=url)
@ -472,6 +474,7 @@ def harvest_fahce(ebook):
return doc.select_one('div.publicationFormatLink a[href]') return doc.select_one('div.publicationFormatLink a[href]')
return harvest_one_generic(ebook, selector) return harvest_one_generic(ebook, selector)
def harvest_fedoabooks(ebook): def harvest_fedoabooks(ebook):
def selector(doc): def selector(doc):
return doc.select('a.cmp_download_link[href]') return doc.select('a.cmp_download_link[href]')
@ -479,3 +482,19 @@ def harvest_fedoabooks(ebook):
return url.replace('view', 'download') + '?inline=1' return url.replace('view', 'download') + '?inline=1'
return harvest_multiple_generic(ebook, selector, dl=dl) return harvest_multiple_generic(ebook, selector, dl=dl)
UBPDF = re.compile(r'/dspace/bitstream/.*\.pdf')
def harvest_ub(ebook):
def selector(doc):
return doc.find(href=UBPDF)
return harvest_one_generic(ebook, selector)
# won't harvest page-image books
def harvest_unt(ebook):
def selector(doc):
return doc.select_one('#link-pdf-version[href]')
return harvest_one_generic(ebook, selector)