Examples of RSSFeed

Examples of net.yacy.cora.document.RSSFeed

66
67
68
69
70
71
72
73
74
75
76
            rssReader = new RSSReader(RSSFeed.DEFAULT_MAXSIZE, source, RSSReader.Type.none);
        } catch (IOException e) {
            throw new Parser.Failure("Load error:" + e.getMessage(), url, e);
        }
       
        final RSSFeed feed = rssReader.getFeed();
        //RSSMessage channel = feed.getChannel();
        final List<Document> docs = new ArrayList<Document>();
        MultiProtocolURI uri;
        Set<String> languages;
        Map<MultiProtocolURI, Properties> anchors;
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

435
436
437
438
439
440
441
442
443
444
445
446
            }
        }
        if (seed == null) return false;

        // we know a peer which should provide remote crawl entries. load them now.
        final RSSFeed feed = yacyClient.queryRemoteCrawlURLs(this.sb.peers, seed, 60, 8000);
        if (feed == null || feed.isEmpty()) {
            // something is wrong with this provider. To prevent that we get not stuck with this peer
            // we remove it from the peer list
            this.sb.peers.peerActions.peerDeparture(seed, "no results from provided remote crawls");
            // try again and ask another peer
            return remoteCrawlLoaderJob();
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

74
75
76
77
78
79
80
81
82
83
84
        }
        if (rss == null) {
            Log.logWarning("Load_RSS", "no rss for url " + this.urlf.toNormalform(true, false));
            return;
        }
        final RSSFeed feed = rss.getFeed();
        indexAllRssFeed(this.sb, this.urlf, feed);

        // add the feed also to the scheduler
        recordAPI(this.sb, null, this.urlf, feed, 7, "seldays");
    }
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
                RSSMessage message;
                int maximumRecords = maximumRecordsInit;
                long timeout = timeoutInit;
                mainloop: while (timeout > 0 && maximumRecords > 0) {
                    final long st = System.currentTimeMillis();
                    RSSFeed feed;
                    try {
                        feed = loadSRURSS(urlBase, query, timeout, startRecord, recordsPerSession, verify, global, userAgent);
                    } catch (final IOException e1) {
                        //e1.printStackTrace();
                        break mainloop;
                    }
                    if (feed == null || feed.isEmpty()) break mainloop;
                    maximumRecords -= feed.size();
                    innerloop: while (!feed.isEmpty()) {
                        message = feed.pollMessage();
                        if (message == null) break innerloop;
                        try {
                            queue.put(message);
                        } catch (final InterruptedException e) {
                            e.printStackTrace();
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

208
209
210
211
212
213
214
215
216
217
218
            final RSSReader reader = RSSReader.parse(RSSFeed.DEFAULT_MAXSIZE, result);
            if (reader == null) {
                throw new IOException("cora.Search failed asking peer '" + uri.getHost() + "': probably bad response from remote peer (1), reader == null");
            }
            final RSSFeed feed = reader.getFeed();
            if (feed == null) {
                // case where the rss reader does not understand the content
                throw new IOException("cora.Search failed asking peer '" + uri.getHost() + "': probably bad response from remote peer (2)");
            }
            return feed;
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
                messageMaxCount--;
                continue channelIteration;
            }

            // read the channel
            final RSSFeed feed = EventChannel.channels(channel);
            if (feed == null || feed.isEmpty()) continue channelIteration;

            RSSMessage message = feed.getChannel();
            if (message != null) {
                prop.putXML("channel_title", message.getTitle());
                prop.putXML("channel_description", message.getDescription());
                prop.put("channel_pubDate", message.getPubDate());
            }
            while (messageMaxCount > 0 && !feed.isEmpty()) {
                message = feed.pollMessage();
                if (message == null) continue;

                // create RSS entry
                prop.putXML("item_" + messageCount + "_title", channelName + ": " + message.getTitle());
                prop.putXML("item_" + messageCount + "_description", message.getDescription());
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
            Log.logException(e);
        }

        // index all selected items: description only
        if (rss != null && post.containsKey("indexSelectedItemContent")) {
            final RSSFeed feed = rss.getFeed();
            loop: for (final Map.Entry<String, String> entry: post.entrySet()) {
                if (entry.getValue().startsWith("mark_")) try {
                    final RSSMessage message = feed.getMessage(entry.getValue().substring(5));
                    final DigestURI messageurl = new DigestURI(message.getLink());
                    if (RSSLoader.indexTriggered.containsKey(messageurl.hash())) continue loop;
                    if (sb.urlExists(Segments.Process.LOCALCRAWLING, messageurl.hash()) != null) continue loop;
                    sb.addToIndex(messageurl, null, null);
                    RSSLoader.indexTriggered.insertIfAbsent(messageurl.hash(), new Date());
                } catch (final IOException e) {
                    Log.logException(e);
                } catch (final Failure e) {
                    Log.logException(e);
                }
            }
        }

        if (rss != null && post.containsKey("indexAllItemContent")) {
            record_api = true;
            final RSSFeed feed = rss.getFeed();
            RSSLoader.indexAllRssFeed(sb, url, feed);
        }

        if (record_api && rss != null && rss.getFeed() != null && rss.getFeed().getChannel() != null) {
            // record API action
            RSSLoader.recordAPI(sb, post.get(WorkTables.TABLE_API_COL_APICALL_PK, null), url, rss.getFeed(), repeat_time, repeat_unit);
        }

        // show items from rss
        if (rss != null) {
            prop.put("showitems", 1);
            final RSSFeed feed = rss.getFeed();
            final RSSMessage channel = feed.getChannel();
            prop.putHTML("showitems_title", channel == null ? "" : channel.getTitle());
            String author = channel == null ? "" : channel.getAuthor();
            if (author == null || author.length() == 0) author = channel == null ? "" : channel.getCopyright();
            Date pubDate = channel == null ? null : channel.getPubDate();
            prop.putHTML("showitems_author", author == null ? "" : author);
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

51
52
53
54
55
56
57
58
59
60
61
        if (post != null) {
            if (post.containsKey("retrieve")) {
                final String peerhash = post.get("peer", null);
                final Seed seed = (peerhash == null) ? null : sb.peers.getConnected(peerhash);
                final RSSFeed feed = (seed == null) ? null : Protocol.queryRemoteCrawlURLs(sb.peers, seed, 20, 60000);
                if (feed != null) {
                    for (final Hit item: feed) {
                        //System.out.println("URL=" + item.getLink() + ", desc=" + item.getDescription() + ", pubDate=" + item.getPubDate());
                       
                        // put url on remote crawl stack
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

435
436
437
438
439
440
441
442
443
444
445
446
            }
        }
        if (seed == null) return false;

        // we know a peer which should provide remote crawl entries. load them now.
        final RSSFeed feed = Protocol.queryRemoteCrawlURLs(this.sb.peers, seed, 60, 8000);
        if (feed == null || feed.isEmpty()) {
            // something is wrong with this provider. To prevent that we get not stuck with this peer
            // we remove it from the peer list
            this.sb.peers.peerActions.peerDeparture(seed, "no results from provided remote crawls");
            // try again and ask another peer
            return remoteCrawlLoaderJob();
View Full Code Here

Examples of net.yacy.cora.document.RSSFeed

74
75
76
77
78
79
80
81
82
83
84
        }
        if (rss == null) {
            Log.logWarning("Load_RSS", "no rss for url " + this.urlf.toNormalform(true, false));
            return;
        }
        final RSSFeed feed = rss.getFeed();
        indexAllRssFeed(this.sb, this.urlf, feed);

        // add the feed also to the scheduler
        recordAPI(this.sb, null, this.urlf, feed, 7, "seldays");
    }
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.