[cc-commits] [CC SVN] r10976 - metadata_scraper/trunk/src/cc/deedscraper
webmaster at creativecommons.org
webmaster at creativecommons.org
Tue Oct 14 13:38:56 EDT 2008
Author: nyergler
Date: 2008-10-14 17:38:56 +0000 (Tue, 14 Oct 2008)
New Revision: 10976
Modified:
metadata_scraper/trunk/src/cc/deedscraper/app.py
Log:
Use a set-based triple sink.
Modified: metadata_scraper/trunk/src/cc/deedscraper/app.py
===================================================================
--- metadata_scraper/trunk/src/cc/deedscraper/app.py 2008-10-14 17:38:00 UTC (rev 10975)
+++ metadata_scraper/trunk/src/cc/deedscraper/app.py 2008-10-14 17:38:56 UTC (rev 10976)
@@ -32,9 +32,7 @@
from decorator import decorator
from support import LOG
-if sys.version < (2,5):
- # import set support
- from sets import Set as set
+from rdfadict.sink import DictSetTripleSink
FOLLOW_PREDICATES = (
'http://www.w3.org/1999/02/22-rdf-syntax-ns#seeAlso',
@@ -79,6 +77,10 @@
contents= opener.open(request).read()
subjects.append(url)
+ # default to a set-based triple sink
+ if sink is None:
+ sink = DictSetTripleSink()
+
triples = parser.parse_string(contents, url, sink)
# look for possible predicates to follow
@@ -112,6 +114,11 @@
# parse the RDFa from the document
triples = self._load_source(url)
+ # post-process the Object sets into lists
+ for s in triples.keys():
+ for p in triples[s].keys():
+ triples[s][p] = list(triples[s][p])
+
ns_cc = 'http://creativecommons.org/ns#'
ns_wr = 'http://web.resource.org/cc/'
More information about the cc-commits
mailing list