1 /**
2 * This file Copyright (c) 2008-2010 Magnolia International
3 * Ltd. (http://www.magnolia-cms.com). All rights reserved.
4 *
5 *
6 * This file is dual-licensed under both the Magnolia
7 * Network Agreement and the GNU General Public License.
8 * You may elect to use one or the other of these licenses.
9 *
10 * This file is distributed in the hope that it will be
11 * useful, but AS-IS and WITHOUT ANY WARRANTY; without even the
12 * implied warranty of MERCHANTABILITY or FITNESS FOR A
13 * PARTICULAR PURPOSE, TITLE, or NONINFRINGEMENT.
14 * Redistribution, except as permitted by whichever of the GPL
15 * or MNA you select, is prohibited.
16 *
17 * 1. For the GPL license (GPL), you can redistribute and/or
18 * modify this file under the terms of the GNU General
19 * Public License, Version 3, as published by the Free Software
20 * Foundation. You should have received a copy of the GNU
21 * General Public License, Version 3 along with this program;
22 * if not, write to the Free Software Foundation, Inc., 51
23 * Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
24 *
25 * 2. For the Magnolia Network Agreement (MNA), this file
26 * and the accompanying materials are made available under the
27 * terms of the MNA which accompanies this distribution, and
28 * is available at http://www.magnolia-cms.com/mna.html
29 *
30 * Any modifications to this file must keep this entire header
31 * intact.
32 *
33 */
34 package info.magnolia.module.rssaggregator.importhandler;
35
36 import com.sun.syndication.feed.synd.SyndFeed;
37 import com.sun.syndication.fetcher.FeedFetcher;
38 import com.sun.syndication.fetcher.impl.FeedFetcherCache;
39 import com.sun.syndication.fetcher.impl.HashMapFeedInfoCache;
40 import com.sun.syndication.fetcher.impl.HttpURLFeedFetcher;
41 import org.slf4j.Logger;
42 import org.slf4j.LoggerFactory;
43
44 import java.net.URL;
45 import java.util.Set;
46
47 /**
48 * A simple single threaded {@link RSSFeedFetcher} for feed retrieval <code>http</code> that follows redirects and
49 * supports conditional gets. <p/> Will attempt to fetch all {@link FeedChannel feed channels} defined in the provided
50 * {@link AggregateFeed aggregate feeds}. Failures to fetch a feed result for a given feed channel are logged at ERROR
51 * level. <p/> Because of the single threaded nature of this implementation, it is suited for situations where the
52 * number feed channels to retrieve is small and the time window between update checks is large enough. <p/> For more
53 * demanding situations, use a more performant RSSFeedFetcher implementation.
54 *
55 * @author Rob van der Linden Vooren
56 * @see RSSFeedFetcher
57 */
58 public class SimpleRSSFeedFetcher implements RSSFeedFetcher {
59
60 private final FeedFetcherCache feedInfoCache;
61
62 private static final Logger log = LoggerFactory.getLogger(SimpleRSSFeedFetcher.class);
63
64 public SimpleRSSFeedFetcher() {
65 this.feedInfoCache = HashMapFeedInfoCache.getInstance();
66 }
67
68 public Set<AggregateFeed> fetchAggregateFeeds(Set<AggregateFeed> aggregateFeeds) {
69 for (AggregateFeed aggregateFeed : aggregateFeeds) {
70 for (FeedChannel channel : aggregateFeed.getChannels()) {
71 log.debug("Fetching feed channel '{}' for aggregateFeed '{}'", channel.getUrl(), aggregateFeed.getName());
72 SyndFeed feed = fetchFeedChannel(channel);
73 if (feed == null) {
74 continue;
75 }
76 channel.setFeed(feed);
77 }
78 }
79 return aggregateFeeds;
80 }
81
82 // Helper methods
83
84 /**
85 * Fetch the given feed channel. A failed fetch results in a <code>null</code> {@link FeedChannel#feed feed}
86 * property set.
87 *
88 * @param channel the feed channel to fetch
89 * @return the feed channel with its {@link FeedChannel#feed feed} property set to the result of the fetch
90 */
91 protected SyndFeed fetchFeedChannel(FeedChannel channel) {
92 try {
93 final FeedFetcher fetcher = new HttpURLFeedFetcher(feedInfoCache);
94 return fetcher.retrieveFeed(new URL(channel.getUrl()));
95 } catch (Exception e) {
96 log.error("Failed to store fetch result for channel '{}': {}", new Object[] {channel, e});
97 return null;
98 }
99 }
100 }