Examples of WaybackRequest

Examples of org.archive.wayback.core.WaybackRequest

25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
  //private final Pattern WB_REQUEST_REGEX = Pattern.compile("^id(\\d+)index(\\d+)$");
  private final Pattern WB_REQUEST_REGEX = Pattern.compile("^id(\\d+)index(\\d+)(\\?(.*))*$");
 

  public WaybackRequest parse(String requestPath) {
    WaybackRequest wbRequest = null;
    Matcher matcher = WB_REQUEST_REGEX.matcher(requestPath); 
    if (matcher != null && matcher.matches()) {
      wbRequest = new WaybackRequest();
      String docId = matcher.group(1);     
      wbRequest.put(WaybackConstants.REQUEST_DOC_ID,docId);
      String indexId = matcher.group(2);     
      wbRequest.put(WaybackConstants.REQUEST_INDEX_ID,indexId);
     
      wbRequest.put(WaybackConstants.REQUEST_TYPE, WaybackConstants.REQUEST_REPLAY_QUERY);
    }
    return wbRequest;
  }
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

48
49
50
51
52
53
54
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
  private final static Pattern WB_QUERY_REGEX = Pattern
      .compile("^(\\d{0,13})\\*/(.*[^*])$");

  public WaybackRequest parse(String requestPath) {
   
    WaybackRequest wbRequest = null;
    Matcher matcher = WB_QUERY_REGEX.matcher(requestPath);
    if (matcher != null && matcher.matches()) {

      wbRequest = new WaybackRequest();
      String dateStr = matcher.group(1);
      String urlStr = matcher.group(2);

      String startDate;
      String endDate;     
      if(dateStr.length() == 0) {
        startDate = getEarliestTimestamp();
        //endDate = getLatestTimestamp(); // BUG MC 120608 - nutchwax 0000051
        endDate = Timestamp.padEndDateStr(""); // BUG MC 120608 - nutchwax 0000051       
      } else {
        startDate = Timestamp.parseBefore(dateStr).getDateStr();
        endDate = Timestamp.parseAfter(dateStr).getDateStr();
      }
      wbRequest.put(WaybackConstants.REQUEST_START_DATE,startDate);
      wbRequest.put(WaybackConstants.REQUEST_END_DATE,endDate);
      wbRequest.put(WaybackConstants.REQUEST_TYPE,
          WaybackConstants.REQUEST_URL_QUERY);
      try {
                wbRequest.setRequestUrl(urlStr);
      } catch (URIException e) {
        wbRequest = null;
      }
    }
    return wbRequest;
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
  private final static Pattern WB_QUERY2_REGEX = Pattern
      .compile("^(\\d{1,14})-(\\d{1,14})\\*/(.*[^*])$");
 

  public WaybackRequest parse(String requestPath) {
    WaybackRequest wbRequest = null;
    Matcher matcher = WB_QUERY2_REGEX.matcher(requestPath);
    if (matcher != null && matcher.matches()) {

      wbRequest = new WaybackRequest();
      String startDateStr = matcher.group(1);
      String endDateStr = matcher.group(2);
      String urlStr = matcher.group(3);

      String startDate = Timestamp.parseBefore(startDateStr).getDateStr();
      String endDate = Timestamp.parseAfter(endDateStr).getDateStr();
      wbRequest.put(WaybackConstants.REQUEST_START_DATE,startDate);
      wbRequest.put(WaybackConstants.REQUEST_END_DATE,endDate);
      wbRequest.put(WaybackConstants.REQUEST_TYPE,
          WaybackConstants.REQUEST_URL_QUERY);
      try {
                wbRequest.setRequestUrl(urlStr);
      } catch (URIException e) {
        wbRequest = null;
      }
    }
    return wbRequest;
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

282
283
284
285
286
287
288
289
290
291
292
293
    return (nodes.getLength() <= 0) ? null : nodes;
  }

  protected String getRequestUrl(WaybackRequest wbRequest)
      throws BadQueryException {
    WaybackRequest tmp = wbRequest.clone();
    String type = tmp.get(WaybackConstants.REQUEST_TYPE);
    if(type.equals(WaybackConstants.REQUEST_REPLAY_QUERY)) {
      tmp.put(WaybackConstants.REQUEST_TYPE, WaybackConstants.REQUEST_URL_QUERY);
    }
    return this.searchUrlBase + "?" + tmp.getQueryArguments();
  }
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

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
    arcCacheDir.shutdown();
  }
 
  private WaybackRequest makeCacheWBRequest(URL url, long maxCacheMS,
      boolean bUseOlder) throws URIException {
    WaybackRequest req = new WaybackRequest();
    req.setRequestUrl(url.toString());
    req.put(WaybackConstants.REQUEST_TYPE,
        WaybackConstants.REQUEST_CLOSEST_QUERY);
    req.put(WaybackConstants.REQUEST_EXACT_DATE,
        Timestamp.currentTimestamp().getDateStr());
    Timestamp earliest = null;
    if(bUseOlder) {
      earliest = Timestamp.earliestTimestamp();
    } else {
      Date d = new Date(System.currentTimeMillis() - maxCacheMS);
      earliest = new Timestamp(d);
    }
    req.put(WaybackConstants.REQUEST_START_DATE,earliest.getDateStr());
    // for now, assume all live web requests are only satisfiable by the
    // exact host -- no massaging.
    req.put(WaybackConstants.REQUEST_EXACT_HOST_ONLY,
        WaybackConstants.REQUEST_YES);
    return req;
  }
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

149
150
151
152
153
154
155
156
157
158
159
  private Resource getLocalCachedResource(URL url, long maxCacheMS,
      boolean bUseOlder) throws ResourceNotInArchiveException,
      IOException, LiveDocumentNotAvailableException {
   
    Resource resource = null;
    WaybackRequest wbRequest = makeCacheWBRequest(url,maxCacheMS,bUseOlder);
   
    CaptureSearchResults results = null;
    try {
      SearchResults gresults = index.query(wbRequest);
      if(!(gresults instanceof CaptureSearchResults)) {
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

47
48
49
50
51
52
53
54
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
   */
  private final static Pattern WB_PATH_QUERY_REGEX = Pattern
      .compile("^(\\d{0,13})\\*/(.*)\\*$");

  public WaybackRequest parse(String requestPath) {
    WaybackRequest wbRequest = null;
    Matcher matcher = WB_PATH_QUERY_REGEX.matcher(requestPath);
    if (matcher != null && matcher.matches()) {

      wbRequest = new WaybackRequest();
      String dateStr = matcher.group(1);
      String urlStr = matcher.group(2);

      String startDate;
      String endDate;
      if(dateStr.length() == 0) {
        startDate = getEarliestTimestamp();
        endDate = getLatestTimestamp();
      } else {
        startDate = Timestamp.parseBefore(dateStr).getDateStr();
        endDate = Timestamp.parseAfter(dateStr).getDateStr();
      }

      wbRequest.put(WaybackConstants.REQUEST_START_DATE,
          startDate);
      wbRequest.put(WaybackConstants.REQUEST_END_DATE,endDate);

      wbRequest.put(WaybackConstants.REQUEST_TYPE,
          WaybackConstants.REQUEST_URL_PREFIX_QUERY);
      try {
                wbRequest.setRequestUrl(urlStr);
      } catch (URIException e) {
        wbRequest = null;
      }
    }
    return wbRequest;
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
   */
  private final static Pattern WB_PATH_QUERY2_REGEX = Pattern
      .compile("^(\\d{1,14})-(\\d{1,14})\\*/(.*)\\*$");

  public WaybackRequest parse(String requestPath) {
    WaybackRequest wbRequest = null;
    Matcher matcher = WB_PATH_QUERY2_REGEX.matcher(requestPath);
    if (matcher != null && matcher.matches()) {

      wbRequest = new WaybackRequest();
      String startDateStr = matcher.group(1);
      String endDateStr = matcher.group(2);
      String urlStr = matcher.group(3);
      String startDate = Timestamp.parseBefore(startDateStr).getDateStr();
      String endDate = Timestamp.parseAfter(endDateStr).getDateStr();
      wbRequest.put(WaybackConstants.REQUEST_START_DATE,
          startDate);
      wbRequest.put(WaybackConstants.REQUEST_END_DATE,endDate);

      wbRequest.put(WaybackConstants.REQUEST_TYPE,
          WaybackConstants.REQUEST_URL_PREFIX_QUERY);
      try {
                wbRequest.setRequestUrl(urlStr);
      } catch (URIException e) {
        wbRequest = null;
      }
    }
    return wbRequest;
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

50
51
52
53
54
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
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
   */
  private final Pattern WB_REQUEST_REGEX = Pattern
      .compile("^(\\d{1,14})/(.*)$");

  public WaybackRequest parse(String requestPath) {
    WaybackRequest wbRequest = null;
    Matcher matcher = WB_REQUEST_REGEX.matcher(requestPath);
    String urlStr = null;
    if (matcher != null && matcher.matches()) {
      wbRequest = new WaybackRequest();
      String dateStr = matcher.group(1);
      urlStr = matcher.group(2);
      if (!urlStr.startsWith("http://")) {
        urlStr = "http://" + urlStr;
      }             

      // The logic of the classic WM wrt timestamp bounding:
      // if 14-digits are specified, assume min-max range boundaries
      // if less than 14 are specified, assume min-max range boundaries
      // based upon amount given (2001 => 20010101... - 20011231...)
      // AND assume the user asked for the LATEST possible date
      // within that range...
      //
      // ...don't ask me, I just work here.

      String startDate = null;
      String endDate = null;
      if (dateStr.length() == 14) {
        startDate = getEarliestTimestamp();
        endDate = getLatestTimestamp();
      } else {

        // classic behavior:
        // startDate = Timestamp.parseBefore(dateStr).getDateStr();
        // endDate = Timestamp.parseAfter(dateStr).getDateStr();
        // dateStr = endDate;

        // "better" behavior:
        startDate = getEarliestTimestamp();
        endDate = getLatestTimestamp();
        dateStr = Timestamp.parseAfter(dateStr).getDateStr();

      }
      wbRequest.put(WaybackConstants.REQUEST_EXACT_DATE, dateStr);
      //wbRequest.put(WaybackConstants.REQUEST_START_DATE, startDate); BUG MC 120608
      //wbRequest.put(WaybackConstants.REQUEST_END_DATE, endDate); BUG MC 120608

      wbRequest.put(WaybackConstants.REQUEST_TYPE,
          WaybackConstants.REQUEST_REPLAY_QUERY);

      try {
//        String wbPrefix = wbRequest.getDefaultWaybackPrefix();
//        if (urlStr.startsWith(wbPrefix)) {
//          wbRequest.setBetterRequestURI(urlStr);
//        }
        wbRequest.setRequestUrl(urlStr);
      } catch (URIException e) {
        if(urlStr != null) {
          LOGGER.severe("Failed parse of url(" + urlStr + ")");
        }
        e.printStackTrace();
View Full Code Here

Examples of org.archive.wayback.core.WaybackRequest

67
68
69
70
71
72
73
74
75
76
77
78
79
      return null;
    }
    String requestPath = origRequestPath.substring(contextPath.length());   
    requestPath=EscapeDecoder.urlUnescape(requestPath); // BUG 0000069 (nutchwax) and BUG 0000075 (nutchwax)
   
    WaybackRequest wbRequest = parse(requestPath);   
    if(wbRequest != null) {
      //wbRequest.fixup(httpRequest); BUG 120608
      wbRequest.setResultsPerPage(maxRecords);
    }   

    return wbRequest;
  }
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.