Package de.anomic.crawler.ResultURLs

Examples of de.anomic.crawler.ResultURLs.EventOrigin

67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
            post = new serverObjects();
            post.put("process", "0");
        }

        // find process number
        EventOrigin tabletype;
        try {
            tabletype = EventOrigin.getEvent(post.getInt("process", 0));
        } catch (final NumberFormatException e) {
            tabletype = EventOrigin.UNKNOWN;
        }

        if (
            post != null &&
            post.containsKey("autoforward") &&
            tabletype == EventOrigin.LOCAL_CRAWLING &&
            ResultURLs.getStackSize(EventOrigin.LOCAL_CRAWLING) == 0) {
            // the main menu does a request to the local crawler page, but in case this table is empty, the overview page is shown
            tabletype = (ResultURLs.getStackSize(EventOrigin.SURROGATES) == 0) ? EventOrigin.UNKNOWN : EventOrigin.SURROGATES;
        }
       
        // check if authorization is needed and/or given
        if (tabletype != EventOrigin.UNKNOWN ||
            (post != null && (post.containsKey("clearlist") ||
            post.containsKey("deleteentry")))) {
            final String authorization = (header.get(RequestHeader.AUTHORIZATION, "xxxxxx"));
            if (authorization.length() != 0) {
                if (! sb.verifyAuthentication(header, true)){
                    // force log-in (again, because wrong password was given)
                    prop.put("AUTHENTICATE", "admin log-in");
                    return prop;
                }
            } else {
                // force log-in
                prop.put("AUTHENTICATE", "admin log-in");
                return prop;
            }
        }

        if (post != null) {
            // custom number of lines
            if (post.containsKey("count")) {
                lines = post.getInt("count", 500);
            }

            // do the commands
            if (post.containsKey("clearlist")) ResultURLs.clearStack(tabletype);

            if (post.containsKey("deleteentry")) {
                final String hash = post.get("hash", null);
                if (hash != null) {
                    // delete from database
                    sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).remove(hash.getBytes());
                }
            }

            if (post.containsKey("deletedomain")) {
                final String hashpart = post.get("hashpart", null);
                final String domain = post.get("domain", null);
                if (hashpart != null) {
                    // delete all urls for this domain from database
                    try {
                        sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).deleteDomain(hashpart);
                        ResultURLs.deleteDomain(tabletype, domain, hashpart);
                    } catch (IOException e) {
                        Log.logException(e);
                    }
                }
            }

            if (post.containsKey("moreIndexed")) {
                lines = post.getInt("showIndexed", 500);
            }

            if (post.get("si") != null) showInit = !("0".equals(post.get("si")));
            if (post.get("se") != null) showExec = !("0".equals(post.get("se")));
            if (post.get("sd") != null) showDate = !("0".equals(post.get("sd")));
            if (post.get("sw") != null) showWords = !("0".equals(post.get("sw")));
            if (post.get("st") != null) showTitle = !("0".equals(post.get("st")));
            if (post.get("su") != null) showURL = !("0".equals(post.get("su")));
        } // end != null

        // create table
        if (tabletype == EventOrigin.UNKNOWN) {
            prop.put("table", "2");
        } else if (ResultURLs.getStackSize(tabletype) == 0 && ResultURLs.getDomainListSize(tabletype) == 0) {
            prop.put("table", "0");
        } else {
            prop.put("table", "1");
            if (lines > ResultURLs.getStackSize(tabletype)) lines = ResultURLs.getStackSize(tabletype);
            if (lines == ResultURLs.getStackSize(tabletype)) {
                prop.put("table_size", "0");
            } else {
                prop.put("table_size", "1");
                prop.put("table_size_count", lines);
            }
            prop.put("table_size_all", ResultURLs.getStackSize(tabletype));
           
            prop.putHTML("table_feedbackpage", "CrawlResults.html");
            prop.put("table_tabletype", tabletype.getCode());
            prop.put("table_showInit", (showInit) ? "1" : "0");
            prop.put("table_showExec", (showExec) ? "1" : "0");
            prop.put("table_showDate", (showDate) ? "1" : "0");
            prop.put("table_showWords", (showWords) ? "1" : "0");
            prop.put("table_showTitle", (showTitle) ? "1" : "0");
            prop.put("table_showURL", (showURL) ? "1" : "0");

            boolean dark = true;
            String urlstr, urltxt;
            yacySeed initiatorSeed, executorSeed;
            URIMetadataRow urle;
            URIMetadataRow.Components metadata;

            int cnt = 0;
            final Iterator<Map.Entry<String, InitExecEntry>> i = ResultURLs.results(tabletype);
            Map.Entry<String, InitExecEntry> entry;
            while (i.hasNext()) {
                entry = i.next();
                try {
                    urle = sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).load(UTF8.getBytes(entry.getKey()));
                    if (urle == null) {
                        Log.logWarning("PLASMA", "CrawlResults: URL not in index with url hash " + entry.getKey());
                        urlstr = null;
                        urltxt = null;
                        metadata = null;
                        continue;
                    }
                    metadata = urle.metadata();
                    urlstr = metadata.url().toNormalform(false, true);
                    urltxt = nxTools.shortenURLString(urlstr, 72); // shorten the string text like a URL
                   
                    initiatorSeed = entry.getValue() == null || entry.getValue().initiatorHash == null ? null : sb.peers.getConnected(ASCII.String(entry.getValue().initiatorHash));
                    executorSeed = entry.getValue() == null || entry.getValue().executorHash == null ? null : sb.peers.getConnected(ASCII.String(entry.getValue().executorHash));

                    prop.put("table_indexed_" + cnt + "_dark", (dark) ? "1" : "0");
                    prop.put("table_indexed_" + cnt + "_feedbackpage", "CrawlResults.html");
                    prop.put("table_indexed_" + cnt + "_tabletype", tabletype.getCode());
                    prop.put("table_indexed_" + cnt + "_urlhash", entry.getKey());

                    if (showInit) {
                        prop.put("table_indexed_" + cnt + "_showInit", "1");
                        prop.put("table_indexed_" + cnt + "_showInit_initiatorSeed", (initiatorSeed == null) ? "unknown" : initiatorSeed.getName());
                    } else
                        prop.put("table_indexed_" + cnt + "_showInit", "0");

                    if (showExec) {
                        prop.put("table_indexed_" + cnt + "_showExec", "1");
                        prop.put("table_indexed_" + cnt + "_showExec_executorSeed", (executorSeed == null) ? "unknown" : executorSeed.getName());
                    } else
                        prop.put("table_indexed_" + cnt + "_showExec", "0");

                    if (showDate && urle != null) {
                        prop.put("table_indexed_" + cnt + "_showDate", "1");
                        prop.put("table_indexed_" + cnt + "_showDate_modified", daydate(urle.moddate()));
                    } else
                        prop.put("table_indexed_" + cnt + "_showDate", "0");

                    if (showWords && urle != null) {
                        prop.put("table_indexed_" + cnt + "_showWords", "1");
                        prop.put("table_indexed_" + cnt + "_showWords_count", urle.wordCount());
                    } else
                        prop.put("table_indexed_" + cnt + "_showWords", "0");

                    if (showTitle) {
                        prop.put("table_indexed_" + cnt + "_showTitle", (showTitle) ? "1" : "0");
                            prop.put("table_indexed_" + cnt + "_showTitle_available", "1");

                            if (metadata == null || metadata.dc_title() == null || metadata.dc_title().trim().length() == 0)
                                prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "0");
                            else {
                                prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "1");
                                prop.putHTML("table_indexed_" + cnt + "_showTitle_available_nodescr_urldescr", metadata.dc_title());
                            }

                            prop.put("table_indexed_" + cnt + "_showTitle_available_urlHash", entry.getKey());
                            prop.putHTML("table_indexed_" + cnt + "_showTitle_available_urltitle", urlstr);
                    } else
                        prop.put("table_indexed_" + cnt + "_showTitle", "0");

                    if (showURL) {
                        prop.put("table_indexed_" + cnt + "_showURL", "1");
                            prop.put("table_indexed_" + cnt + "_showURL_available", "1");

                            prop.put("table_indexed_" + cnt + "_showURL_available_urlHash", entry.getKey());
                            prop.putHTML("table_indexed_" + cnt + "_showURL_available_urltitle", urlstr);
                            prop.put("table_indexed_" + cnt + "_showURL_available_url", urltxt);
                    } else
                        prop.put("table_indexed_" + cnt + "_showURL", "0");

                    dark = !dark;
                    cnt++;
                } catch (final Exception e) {
                    Log.logSevere("PLASMA", "genTableProps", e);
                }
            }
            prop.put("table_indexed", cnt);
           
            cnt = 0;
            dark = true;
            final Iterator<String> j = ResultURLs.domains(tabletype);
            String domain;
            while (j.hasNext() && cnt < 100) {
                domain = j.next();
                if (domain == null) break;
                prop.put("table_domains_" + cnt + "_dark", (dark) ? "1" : "0");
                prop.put("table_domains_" + cnt + "_feedbackpage", "CrawlResults.html");
                prop.put("table_domains_" + cnt + "_tabletype", tabletype.getCode());
                prop.put("table_domains_" + cnt + "_domain", domain);
                prop.put("table_domains_" + cnt + "_hashpart", DigestURI.hosthash6(domain));
                prop.put("table_domains_" + cnt + "_count", ResultURLs.domainCount(tabletype, domain));
                dark = !dark;
                cnt++;
            }
            prop.put("table_domains", cnt);
        }
        prop.put("process", tabletype.getCode());
        // return rewrite properties
        return prop;
    }
View Full Code Here
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
        return new indexingQueueEntry(in.process, in.queueEntry, documents, null);
    }

    private Document[] parseDocument(final Response response) throws InterruptedException {
        Document[] documents = null;
        final EventOrigin processCase = response.processCase(this.peers.mySeed().hash);

        if (this.log.isFine()) {
            this.log.logFine("processResourceStack processCase=" + processCase +
                ", depth=" + response.depth() +
                ", maxDepth=" + ((response.profile() == null) ? "null" : Integer.toString(response.profile().depth())) +
View Full Code Here
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
    private void storeDocumentIndex(final Segments.Process process, final Response queueEntry, final Document document, final Condenser condenser, final SearchEvent searchEvent, final String sourceName) {

        // CREATE INDEX
        final String dc_title = document.dc_title();
        final DigestURI referrerURL = queueEntry.referrerURL();
        EventOrigin processCase = queueEntry.processCase(this.peers.mySeed().hash);
        if (process == Segments.Process.SURROGATES) {
            processCase = EventOrigin.SURROGATES;
        }

        if (condenser == null || document.indexingDenied()) {
View Full Code Here
803
804
805
806
807
808
809
810
811
812
813
        // 2) result of search queries, some indexes are here (not possible here)
        // 3) result of index transfer, some of them are here (not possible here)
        // 4) proxy-load (initiator is "------------")
        // 5) local prefetch/crawling (initiator is own seedHash)
        // 6) local fetching for global crawling (other known or unknwon initiator)
        EventOrigin processCase = EventOrigin.UNKNOWN;
        // FIXME the equals seems to be incorrect: String.equals(boolean)
        if (initiator() == null || initiator().length == 0 || ASCII.String(initiator()).equals("------------")) {
            // proxy-load
            processCase = EventOrigin.PROXY_LOAD;
        } else if (UTF8.String(initiator()).equals(mySeedHash)) {
View Full Code Here
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
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
            post = new serverObjects();
            post.put("process", "0");
        }

        // find process number
        EventOrigin tabletype;
        try {
            tabletype = EventOrigin.getEvent(post.getInt("process", 0));
        } catch (final NumberFormatException e) {
            tabletype = EventOrigin.UNKNOWN;
        }

        if (
            post != null &&
            post.containsKey("autoforward") &&
            tabletype == EventOrigin.LOCAL_CRAWLING &&
            ResultURLs.getStackSize(EventOrigin.LOCAL_CRAWLING) == 0) {
            // the main menu does a request to the local crawler page, but in case this table is empty, the overview page is shown
            tabletype = (ResultURLs.getStackSize(EventOrigin.SURROGATES) == 0) ? EventOrigin.UNKNOWN : EventOrigin.SURROGATES;
        }

        // check if authorization is needed and/or given
        if (tabletype != EventOrigin.UNKNOWN ||
            (post != null && (post.containsKey("clearlist") ||
            post.containsKey("deleteentry")))) {
            final String authorization = (header.get(RequestHeader.AUTHORIZATION, "xxxxxx"));
            if (authorization.length() != 0) {
                if (! sb.verifyAuthentication(header)){
                    // force log-in (again, because wrong password was given)
                    prop.put("AUTHENTICATE", "admin log-in");
                    return prop;
                }
            } else {
                // force log-in
                prop.put("AUTHENTICATE", "admin log-in");
                return prop;
            }
        }

        if (post != null) {
            // custom number of lines
            if (post.containsKey("count")) {
                lines = post.getInt("count", 500);
            }

            // do the commands
            if (post.containsKey("clearlist")) ResultURLs.clearStack(tabletype);

            if (post.containsKey("deleteentry")) {
                final String hash = post.get("hash", null);
                if (hash != null) {
                    // delete from database
                    sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).remove(hash.getBytes());
                }
            }

            if (post.containsKey("deletedomain")) {
                final String domain = post.get("domain", null);
                final String hashpart = domain == null ? null : DigestURI.hosthash6(domain);
                if (hashpart != null) {
                    // delete all urls for this domain from database
                    try {
                        sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).deleteDomain(hashpart);
                        ResultURLs.deleteDomain(tabletype, domain, hashpart);
                    } catch (final IOException e) {
                        Log.logException(e);
                    }
                }
            }

            if (post.containsKey("moreIndexed")) {
                lines = post.getInt("showIndexed", 500);
            }

            if (post.get("si") != null) showInit    = !("0".equals(post.get("si")));
            if (post.get("se") != null) showExec    = !("0".equals(post.get("se")));
            if (post.get("sd") != null) showDate    = !("0".equals(post.get("sd")));
            if (post.get("sw") != null) showWords   = !("0".equals(post.get("sw")));
            if (post.get("st") != null) showTitle   = !("0".equals(post.get("st")));
            if (post.get("sc") != null) showCountry = !("0".equals(post.get("sc")));
            if (post.get("sp") != null) showIP      = !("0".equals(post.get("sp")));
            if (post.get("su") != null) showURL     = !("0".equals(post.get("su")));
        } // end != null

        // create table
        if (tabletype == EventOrigin.UNKNOWN) {
            prop.put("table", "2");
        } else if (ResultURLs.getStackSize(tabletype) == 0 && ResultURLs.getDomainListSize(tabletype) == 0) {
            prop.put("table", "0");
        } else {
            prop.put("table", "1");
            if (lines > ResultURLs.getStackSize(tabletype)) lines = ResultURLs.getStackSize(tabletype);
            if (lines == ResultURLs.getStackSize(tabletype)) {
                prop.put("table_size", "0");
            } else {
                prop.put("table_size", "1");
                prop.put("table_size_count", lines);
            }
            prop.put("table_size_all", ResultURLs.getStackSize(tabletype));

            prop.putHTML("table_feedbackpage", "CrawlResults.html");
            prop.put("table_tabletype", tabletype.getCode());
            prop.put("table_showInit",    (showInit) ? "1" : "0");
            prop.put("table_showExec",    (showExec) ? "1" : "0");
            prop.put("table_showDate",    (showDate) ? "1" : "0");
            prop.put("table_showWords",   (showWords) ? "1" : "0");
            prop.put("table_showTitle",   (showTitle) ? "1" : "0");
            prop.put("table_showCountry", (showCountry) ? "1" : "0");
            prop.put("table_showIP",      (showIP) ? "1" : "0");
            prop.put("table_showURL",     (showURL) ? "1" : "0");

            boolean dark = true;
            String urlstr, urltxt;
            Seed initiatorSeed, executorSeed;
            URIMetadataRow urle;
            URIMetadataRow.Components metadata;

            int cnt = 0;
            final Iterator<Map.Entry<String, InitExecEntry>> i = ResultURLs.results(tabletype);
            Map.Entry<String, InitExecEntry> entry;
            while (i.hasNext()) {
                entry = i.next();
                try {
                    urle = sb.indexSegments.urlMetadata(Segments.Process.LOCALCRAWLING).load(UTF8.getBytes(entry.getKey()));
                    if (urle == null) {
                        Log.logWarning("PLASMA", "CrawlResults: URL not in index with url hash " + entry.getKey());
                        urlstr = null;
                        urltxt = null;
                        metadata = null;
                        continue;
                    }
                    metadata = urle.metadata();
                    urlstr = metadata.url().toNormalform(false, true);
                    urltxt = nxTools.shortenURLString(urlstr, 72); // shorten the string text like a URL

                    initiatorSeed = entry.getValue() == null || entry.getValue().initiatorHash == null ? null : sb.peers.getConnected(ASCII.String(entry.getValue().initiatorHash));
                    executorSeed = entry.getValue() == null || entry.getValue().executorHash == null ? null : sb.peers.getConnected(ASCII.String(entry.getValue().executorHash));

                    prop.put("table_indexed_" + cnt + "_dark", (dark) ? "1" : "0");
                    prop.put("table_indexed_" + cnt + "_feedbackpage", "CrawlResults.html");
                    prop.put("table_indexed_" + cnt + "_tabletype", tabletype.getCode());
                    prop.put("table_indexed_" + cnt + "_urlhash", entry.getKey());

                    if (showInit) {
                        prop.put("table_indexed_" + cnt + "_showInit", "1");
                        prop.put("table_indexed_" + cnt + "_showInit_initiatorSeed", (initiatorSeed == null) ? "unknown" : initiatorSeed.getName());
                    } else
                        prop.put("table_indexed_" + cnt + "_showInit", "0");

                    if (showExec) {
                        prop.put("table_indexed_" + cnt + "_showExec", "1");
                        prop.put("table_indexed_" + cnt + "_showExec_executorSeed", (executorSeed == null) ? "unknown" : executorSeed.getName());
                    } else
                        prop.put("table_indexed_" + cnt + "_showExec", "0");

                    if (showDate && urle != null) {
                        prop.put("table_indexed_" + cnt + "_showDate", "1");
                        prop.put("table_indexed_" + cnt + "_showDate_modified", daydate(urle.moddate()));
                    } else
                        prop.put("table_indexed_" + cnt + "_showDate", "0");

                    if (showWords && urle != null) {
                        prop.put("table_indexed_" + cnt + "_showWords", "1");
                        prop.put("table_indexed_" + cnt + "_showWords_count", urle.wordCount());
                    } else
                        prop.put("table_indexed_" + cnt + "_showWords", "0");

                    if (showTitle) {
                        prop.put("table_indexed_" + cnt + "_showTitle", (showTitle) ? "1" : "0");
                        prop.put("table_indexed_" + cnt + "_showTitle_available", "1");

                        if (metadata == null || metadata.dc_title() == null || metadata.dc_title().trim().length() == 0)
                            prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "0");
                        else {
                            prop.put("table_indexed_" + cnt + "_showTitle_available_nodescr", "1");
                            prop.putHTML("table_indexed_" + cnt + "_showTitle_available_nodescr_urldescr", metadata.dc_title());
                        }

                        prop.put("table_indexed_" + cnt + "_showTitle_available_urlHash", entry.getKey());
                        prop.putHTML("table_indexed_" + cnt + "_showTitle_available_urltitle", urlstr);
                    } else
                        prop.put("table_indexed_" + cnt + "_showTitle", "0");

                    if (showCountry && urle != null) {
                        prop.put("table_indexed_" + cnt + "_showCountry", "1");
                        prop.put("table_indexed_" + cnt + "_showCountry_country", metadata.url().getLocale().getCountry());
                    } else
                        prop.put("table_indexed_" + cnt + "_showCountry", "0");

                    if (showIP && urle != null) {
                        prop.put("table_indexed_" + cnt + "_showIP", "1");
                        prop.put("table_indexed_" + cnt + "_showIP_ip", metadata.url().getInetAddress().getHostAddress());
                    } else
                        prop.put("table_indexed_" + cnt + "_showIP", "0");

                    if (showURL) {
                        prop.put("table_indexed_" + cnt + "_showURL", "1");
                        prop.put("table_indexed_" + cnt + "_showURL_available", "1");

                        prop.put("table_indexed_" + cnt + "_showURL_available_urlHash", entry.getKey());
                        prop.putHTML("table_indexed_" + cnt + "_showURL_available_urltitle", urlstr);
                        prop.put("table_indexed_" + cnt + "_showURL_available_url", urltxt);
                    } else
                        prop.put("table_indexed_" + cnt + "_showURL", "0");

                    dark = !dark;
                    cnt++;
                } catch (final Exception e) {
                    Log.logSevere("PLASMA", "genTableProps", e);
                }
            }
            prop.put("table_indexed", cnt);

            cnt = 0;
            dark = true;
            final Iterator<String> j = ResultURLs.domains(tabletype);
            String domain;
            while (j.hasNext() && cnt < 100) {
                domain = j.next();
                if (domain == null) break;
                prop.put("table_domains_" + cnt + "_dark", (dark) ? "1" : "0");
                prop.put("table_domains_" + cnt + "_feedbackpage", "CrawlResults.html");
                prop.put("table_domains_" + cnt + "_tabletype", tabletype.getCode());
                prop.put("table_domains_" + cnt + "_domain", domain);
                prop.put("table_domains_" + cnt + "_count", ResultURLs.domainCount(tabletype, domain));
                dark = !dark;
                cnt++;
            }
            prop.put("table_domains", cnt);
        }
        prop.put("process", tabletype.getCode());
        // return rewrite properties
        return prop;
    }
View Full Code Here
802
803
804
805
806
807
808
809
810
811
812
        // 2) result of search queries, some indexes are here (not possible here)
        // 3) result of index transfer, some of them are here (not possible here)
        // 4) proxy-load (initiator is "------------")
        // 5) local prefetch/crawling (initiator is own seedHash)
        // 6) local fetching for global crawling (other known or unknwon initiator)
        EventOrigin processCase = EventOrigin.UNKNOWN;
        // FIXME the equals seems to be incorrect: String.equals(boolean)
        if (initiator() == null || initiator().length == 0 || ASCII.String(initiator()).equals("------------")) {
            // proxy-load
            processCase = EventOrigin.PROXY_LOAD;
        } else if (UTF8.String(initiator()).equals(mySeedHash)) {
View Full Code Here
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
        return new indexingQueueEntry(in.process, in.queueEntry, documents, null);
    }

    private Document[] parseDocument(final Response response) throws InterruptedException {
        Document[] documents = null;
        final EventOrigin processCase = response.processCase(this.peers.mySeed().hash);

        if (this.log.isFine()) {
            this.log.logFine("processResourceStack processCase=" + processCase +
                ", depth=" + response.depth() +
                ", maxDepth=" + ((response.profile() == null) ? "null" : Integer.toString(response.profile().depth())) +
View Full Code Here
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
        // CREATE INDEX
        final String dc_title = document.dc_title();
        final DigestURI url = new DigestURI(document.dc_source());
        final DigestURI referrerURL = queueEntry.referrerURL();
        EventOrigin processCase = queueEntry.processCase(this.peers.mySeed().hash);
        if (process == Segments.Process.SURROGATES) {
            processCase = EventOrigin.SURROGATES;
        }

        if (condenser == null || document.indexingDenied()) {
View Full Code Here
TOP

Related Classes of de.anomic.crawler.ResultURLs.EventOrigin

  • CrawlResults
  • de.anomic.crawler.retrieval.Response
  • de.anomic.search.Switchboard
  • net.yacy.search.Switchboard
    • Copyright © 2018 www.massapicom. 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.