def _build_response_data(self, req, response):
encoding = 'utf8'
unicode_html = u''
try:
unicode_html = response['data'].decode(encoding, 'ignore')
except Exception as e:
logger.warn('failed to decode bytes from url: %s', req.url)
return_type = req.get('return_type') or 'doc'
if return_type == 'doc':
doc = http.Doc(url=req.url, html=unicode_html)
doc.req = req
doc.status.code = response['code']
doc.status.message = response['message']
return doc
elif return_type == 'html':
html = common.DataItem(unicode_html)
html.req = req
html.status = common.DataObject()
html.status.code = response['code']
html.status.message = response['message']
return html
else:
self.scraper.logger.warn('unsupported return_type: %s',
return_type)
return None
def save(self,
record,
filename='result.csv',
max=None,
keys=[],
id=None,
headers=[],
remove_existing_file=True,
always_quoted=True):
#waiting while other thread writing
while self.writingflag:
pass
#hold the flag
self.writingflag = True
path = os.path.join(self.dir, filename)
format = common.DataItem(path).subreg('\.([a-z]{2,5})$--is').lower()
if not self.outdb.get(path):
if os.path.exists(path):
if remove_existing_file:
os.remove(path)
self.outdb.update({
path:
common.DataObject(cnt=0, data=[], ids=[], format=format)
})
trackingobj = self.outdb.get(path)
if keys or id:
id = id or u"".join(
[unicode(record[record.index(key) + 1]) for key in keys])
if id in trackingobj.ids:
self.writingflag = False
return
else:
trackingobj.ids.append(id)
trackingobj.cnt += 1
if format == 'csv':
#for csv format, save to file immediately
common.save_csv(path, record, always_quoted=always_quoted)
elif format in ['xls', 'xlsx']:
#save for later
trackingobj.data.append(record)
if max and trackingobj.cnt == max:
self.flush() #save output files and quit
os._exit(1)
#free the flag
self.writingflag = False
def handler(doc):
page = stats.page
doc.page = page
if verify:
if not verify(
common.DataObject(starturl=common.DataItem(url),
page=page,
doc=doc)):
doc.ok = False
logger.warn("invalid doc at page {0}".format(page))
logger.info('page %s', page)
#download and parse details
if detail:
listings = detail(
common.DataObject(
starturl=common.DataItem(url), page=page,
doc=doc)) if hasattr(detail,
'__call__') else doc.q(detail)
logger.info('details: %s', len(listings))
for listing in listings:
self.downloader.put(Request(url=listing if isinstance(
listing, basestring) else listing.nodevalue(),
cb=parse_detail,
meta=meta,
**options),
onhold=list_pages_first)
done = False
_nexturl = None
_next_post = None
if next:
_nexturl = next(
common.DataObject(
starturl=common.DataItem(url), page=page,
doc=doc)) if hasattr(next, '__call__') else (
next if next.startswith('http') else doc.x(next))
if next_post:
if not next:
#next is not provided, use the original url
_nexturl = doc.url
_next_post = next_post(
common.DataObject(
doc=doc, page=page,
starturl=common.DataItem(url))) if hasattr(
next_post, '__call__') else next_post
if next_post:
if _next_post:
done = False
else:
done = True
else:
if not _nexturl:
done = True
else:
done = False
#if (next and _nexturl ) or (next_post and _next_post):
if not done:
#logger.debug('next_post: %s, _nexturl: %s', _next_post, _nexturl)
stats.page += 1
if max_pages != 0 and stats.page > max_pages:
done = True
else:
self.downloader.put(
Request(_nexturl, _next_post, cb=handler, **options))
else:
done = True
if parse_list:
parse_list(doc)
def pagin(self,
url,
next=None,
post=None,
next_post=None,
parse_list=None,
detail=None,
parse_detail=None,
cc=3,
max_pages=0,
list_pages_first=True,
start_now=False,
debug=True,
verify=None,
meta={},
**_options):
if cc != self.downloader.cc:
self.downloader.set_cc(cc)
options = common.combine_dicts(self.config, _options)
stats = common.DataObject(page=1)
#apply scraper-level options
def handler(doc):
page = stats.page
doc.page = page
if verify:
if not verify(
common.DataObject(starturl=common.DataItem(url),
page=page,
doc=doc)):
doc.ok = False
logger.warn("invalid doc at page {0}".format(page))
logger.info('page %s', page)
#download and parse details
if detail:
listings = detail(
common.DataObject(
starturl=common.DataItem(url), page=page,
doc=doc)) if hasattr(detail,
'__call__') else doc.q(detail)
logger.info('details: %s', len(listings))
for listing in listings:
self.downloader.put(Request(url=listing if isinstance(
listing, basestring) else listing.nodevalue(),
cb=parse_detail,
meta=meta,
**options),
onhold=list_pages_first)
done = False
_nexturl = None
_next_post = None
if next:
_nexturl = next(
common.DataObject(
starturl=common.DataItem(url), page=page,
doc=doc)) if hasattr(next, '__call__') else (
next if next.startswith('http') else doc.x(next))
if next_post:
if not next:
#next is not provided, use the original url
_nexturl = doc.url
_next_post = next_post(
common.DataObject(
doc=doc, page=page,
starturl=common.DataItem(url))) if hasattr(
next_post, '__call__') else next_post
if next_post:
if _next_post:
done = False
else:
done = True
else:
if not _nexturl:
done = True
else:
done = False
#if (next and _nexturl ) or (next_post and _next_post):
if not done:
#logger.debug('next_post: %s, _nexturl: %s', _next_post, _nexturl)
stats.page += 1
if max_pages != 0 and stats.page > max_pages:
done = True
else:
self.downloader.put(
Request(_nexturl, _next_post, cb=handler, **options))
else:
done = True
if parse_list:
parse_list(doc)
##### end of the handler function ##################################################
#start the initial url
self.downloader.put(Request(url, post, cb=handler, **options))
if start_now:
self.downloader.start()
