Examples of CaptureSearchResult

Examples of org.archive.wayback.core.CaptureSearchResult

65
66
67
68
69
70
71
72
73
74
75
76
   * @param results
   */
  public void filter(CaptureSearchResults results) {
    Iterator<CaptureSearchResult> itr = results.iterator();
    while(itr.hasNext()) {
      CaptureSearchResult result = itr.next();
      String captureDate = result.getCaptureTimestamp();
      if((captureDate.compareTo(startDateStr) >= 0)
          && (captureDate.compareTo(endDateStr) < 0)) {
        matches.add(result);
      }   
    }
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

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
    } catch (IllegalArgumentException e) {
      LOGGER.warning("Skipping unrecognized record type : " + typeStr);
      return null;
    }

    CaptureSearchResult result = genericResult(rec);

    switch (type) {
    case response:
      String mime = annotater.transformHTTPMime(header.getMimetype());
      if(mime != null && mime.equals("text/dns")) {
        // close to complete reading, then the digest is legit
        // TODO: DO we want to use the WARC header digest for this?
        rec.close();
        result.setDigest(transformWARCDigest(rec.getDigestStr()));
        result.setMimeType(mime);
      } else {
        result = adaptWARCHTTPResponse(result,rec);
      }
      break;

     
    case revisit:
      // also set the mime type:
      result.setMimeType("warc/revisit");
      break;
     
    case request:
      if(processAll) {
        // also set the mime type:
        result.setMimeType("warc/request");
      } else {
        result = null;
      }
      break;
     
    case metadata:
      if(processAll) {
        // also set the mime type:
        result.setMimeType("warc/metadata");
      } else {
        result = null;
      }
      break;
     
    case warcinfo:
      result.setMimeType(WARC_FILEDESC_VERSION);
      break;
     
    default:
      LOGGER.info("Skipping record type : " + type);
      break;
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

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
   * 
   *    file, offset, timestamp, digest, urlKey, originalUrl
   */
  private CaptureSearchResult genericResult(WARCRecord rec) {

    CaptureSearchResult result = new CaptureSearchResult();

    result.setMimeType(DEFAULT_VALUE);
    result.setHttpCode(DEFAULT_VALUE);
    result.setRedirectUrl(DEFAULT_VALUE);

    ArchiveRecordHeader header = rec.getHeader();

    String file = transformWARCFilename(header.getReaderIdentifier());
    long offset = header.getOffset();
   
    result.setCaptureTimestamp(transformWARCDate(header.getDate()));
    result.setFile(file);
    result.setOffset(offset);
    result.setDigest(transformWARCDigest(header.getHeaderValue(
        WARCRecord.HEADER_KEY_PAYLOAD_DIGEST)));
   
    String origUrl = header.getUrl();
    if(origUrl == null) {
      String type = header.getHeaderValue(WARCConstants.HEADER_KEY_TYPE).toString();
      if(type.equals(WARCConstants.WARCRecordType.warcinfo)) {
        String filename = header.getHeaderValue(
            WARCConstants.HEADER_KEY_FILENAME).toString();
        result.setOriginalUrl("filedesc:"+filename);
        result.setUrlKey("filedesc:"+filename);       
      } else {
        result.setOriginalUrl(DEFAULT_VALUE);
        result.setUrlKey(DEFAULT_VALUE);
      }

     
    } else {
      result.setOriginalUrl(origUrl);
      try {
        String urlKey = canonicalizer.urlStringToKey(origUrl);
        result.setUrlKey(urlKey);
      } catch (URIException e) {
        String shortUrl =
          (origUrl.length() < 100)
          ? origUrl
          :origUrl.substring(0,100);
        LOGGER.warning("FAILED canonicalize(" + shortUrl + "):" +
            file + " " + offset);
        result.setUrlKey(origUrl);
      }
    }
    return result;
  }
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

30
31
32
33
34
35
36
37
38
39
40
41
        Iterator<CaptureSearchResult> itr = results.iterator();
        previous = null;
        next = null;
        this.result = result;
        while(itr.hasNext()) {
                CaptureSearchResult cur = itr.next();
                if(cur.isClosest()) {
                        break;
                }
                previous = cur;
        }
        if(itr.hasNext()) {
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
  private void createNext() {
    if(cachedNext == null) {
      if(peek.hasNext()) {
        // populate
        CaptureSearchResult captureResult = peek.next();
        String currentKey = captureResult.getUrlKey();
        String originalUrl = captureResult.getOriginalUrl();
        String firstCapture = captureResult.getCaptureTimestamp();
        LOGGER.info("Creating new UrlResult:" + currentKey + " " +
            firstCapture);
        String lastCapture = firstCapture;
        HashMap<String,Object> digests = new HashMap<String,Object>();
        digests.put(captureResult.getDigest(),null);
        int numCaptures = 1;

        cachedNext = new UrlSearchResult();
        cachedNext.setUrlKey(currentKey);
        cachedNext.setOriginalUrl(originalUrl);

        // now rip through the rest until we find either the last
        // in the iterator, or the first having a different urlKey:
        while((captureResult = peek.peekNext()) != null) {
          String urlKey = captureResult.getUrlKey();
          if(currentKey.equals(urlKey)) {
            // remove from iterator, and accumulate:
            peek.next();
            numCaptures++;
            digests.put(captureResult.getDigest(), null);

            String captureTS = captureResult.getCaptureTimestamp();
            if(captureTS.compareTo(firstCapture) < 0) {
              firstCapture = captureTS;
            }
            if(captureTS.compareTo(lastCapture) > 0) {
              lastCapture = captureTS;
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

398
399
400
401
402
403
404
405
406
407
408
        Resource payloadResource = createTestHtmlResource("20100601000000", "hogheogehoge\n".getBytes("UTF-8"));
        CaptureSearchResults results = setupCaptures(
                0,
                payloadResource
                );
        CaptureSearchResult closest = results.getClosest();
       
        // when closest's timestamp == request's timestamp,
        // it gets ReplayRenderer with replay.getRenderer(wbRequest, closest, httpHeaderResource, payloadResource),
        // and calls renderResource() on it.
        EasyMock.expect(replay.getRenderer(wbRequest, closest, payloadResource, payloadResource)).andReturn(replayRenderer);
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
        byte[] payload = "hogehogehogehoge\n".getBytes("UTF-8");
        Resource payloadResource = createTestHtmlResource("20100501000001", payload);
        Resource headerResource = createTestRevisitResource("20100601000000", payload.length, true);
        CaptureSearchResults results = setupCaptures(1, payloadResource, headerResource);

        CaptureSearchResult previous = results.getResults().get(0);
        CaptureSearchResult closest = results.getClosest();
//        previous.setFile("aaa.warc.gz");
//        previous.setOffset(0);
        closest.flagDuplicateDigest(previous); // right? TODO: could be done in setupCaptures()
        assertTrue(closest.isDuplicateDigest());
        assertTrue(closest.getDuplicatePayloadFile() != null);
        assertTrue(closest.getDuplicatePayloadOffset() != null);

        EasyMock.expect(replay.getRenderer(wbRequest, closest, headerResource, payloadResource)).andReturn(replayRenderer);
        // calls replayRenderer.renderResource(...)
        replayRenderer.renderResource(httpRequest, httpResponse, wbRequest,
                closest, headerResource, payloadResource, cut.getUriConverter(),
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

550
551
552
553
554
555
556
557
558
559
560
561
                1,
                createTestHtmlResource("http://test.example.com/style.css",
                        "20100501000000", "hogheogehoge\n".getBytes("UTF-8")),
                createTest502Resource()
                );
        CaptureSearchResult closest = results.getClosest();
        assertTrue(closest.isHttpError());
       
        // or wbRequest.setBestLatestReplayRequest();
        final String expectedRedirectURI = "/web/20100501000000cs_/http://test.example.com/style.css";
        httpResponse.setHeader("Location", expectedRedirectURI);
        httpResponse.setStatus(302);
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
        // - calls queryIndex(), which calls collection.resourceIndex.query(),
        //     which returns CaptureSearchResults
        //   (unexpected object from queryIndex() results in WaybackException("Unknown index format").
        //    this is considered to be a programming/configuration error. not tested.)
        CaptureSearchResults results = new CaptureSearchResults();
        CaptureSearchResult result = new CaptureSearchResult();
        results.setClosest(result);
        EasyMock.expect(resourceIndex.query(wbRequest)).andReturn(results);
        // - calls MementoUtils.printTimemapResponse(results, wbRequest, httpResponse) instead
        //     if wbRequst.isMementoTimemapRequest() (N/A here) (TODO: can we move this to
        //     QueryRenderer implementation?)
        // - calls query.renderCaptureResults(...)
        query.renderCaptureResults(httpRequest, httpResponse, wbRequest, results, cut.getUriConverter());
       
        EasyMock.replay(httpRequest, httpResponse, resourceIndex, query);
       
        cut.init();
        boolean r = cut.handleRequest(httpRequest, httpResponse);
       
        EasyMock.verify(query);
       
        // result shall have closest flag set (FIrefox proxy plugin expects this)
        assertTrue("closest flag", result.isClosest());
       
        assertTrue("handleRequest return value", r);
       
    }
View Full Code Here

Examples of org.archive.wayback.core.CaptureSearchResult

146
147
148
149
150
151
152
153
154
155
156
157
158
    assertTrue("subpath",isBlocked(filter,"http://www.peagreenboat.com/foo"));
    assertTrue("emptypath",isBlocked(filter,"http://www.peagreenboat.com/"));
  }
 
  private boolean isBlocked(ObjectFilter<CaptureSearchResult> filter, String url) throws URIException {
    CaptureSearchResult result = new CaptureSearchResult();
    result.setOriginalUrl(url);
    result.setUrlKey(canonicalizer.urlStringToKey(url));
    int filterResult = filter.filterObject(result);
    if(filterResult == ObjectFilter.FILTER_EXCLUDE) {
      return true;
    }
    return false;
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.