def check_duplicate( self, doc ):
# For dates before the site change we should try to verify
# the document duplication by other means (since the 'claint' changed
# on the new site
if doc['date'] < datetime.datetime(2014,9,19):
# Does the current doc_type have synonyms?
doc_types = [ doc['doc_type'].lower() ]
for sn in synonyms:
if doc['doc_type'].lower() in sn:
doc_types = sn
# Create a query for the synonyms:
dt_qs = Q( doc_type__iexact = doc_types[0] )
for dt in doc_types[1:]:
dt_qs = dt_qs | Q( doc_type__iexact = dt )
dl = Document.objects.filter(
date__exact = doc['date'] ).filter(
dt_qs ).filter(
number__iexact = doc['number'] ).filter(
series__exact = doc['series'] )
if len(dl) > 1:
# We have a number of documents that, for a given date, have
# duplicates with the same number and type. The dates can be
# listed with:
# select
# count(*), date, doc_type, number
# from
# dreapp_document
# where
# date < '2014-9-18'
# group by
# date, doc_type, number
# having
# count(*) > 1;
logger.error('Duplicate document in the database: %(doc_type)s %(number)s %(date_st)s' % doc)
raise DREScraperError('More than one doc with the same number and type.')
if len(dl) == 1:
doc['document'] = dl[0]
raise DREDuplicateError('Duplicate document')
# For other dates we simply use the db integrity checks to spot a
# duplicate
document = doc['document']
try:
sid = transaction.savepoint()
document.save()
transaction.savepoint_commit(sid)
logger.debug('ID: %d http://dre.tretas.org/dre/%d/' % (document.id, document.id) )
except IntegrityError:
# Duplicated document
transaction.savepoint_rollback(sid)
doc['document'] = Document.objects.get(claint = doc['id'] )
raise DREDuplicateError('Duplicate document')
def get_html(self):
allowed_areas = (
( 45, 40, 295, 780),
(297, 40, 550, 780),
)
txt = convert_pdf_to_txt_layout(self.filename, allowed_areas)
txt = self.cut_out_doc(txt)
if not txt:
logger.error('CACHEPDF No text for doc id=%d', self.doc.id)
return parse_generic_document.run(txt)
def save_file(filename, url):
k = 1
while True:
try:
url, data_blob, cookies = fetch_url( url )
break
except urllib2.HTTPError:
logger.error('Could not read PDF: %s DOC: %s' % ( url, filename))
k += 1
if k == MAX_ATTEMPTS:
raise DREError('Couldn\'t get the PDF: %s' % url )
logger.debug('Sleeping 2 secs...')
time.sleep(2)
with open(filename, 'wb') as f:
f.write(data_blob)
f.close()
