root/gs3-extensions/html-to-expeditee/trunk/src/src/cgi-bin/html-to-expeditee.pl.in @ 24934

Revision 24934, 9.0 KB (checked in by davidb, 8 years ago)

Introduction of new Perl module to handle taking the JSON version of the traversed HTML tree and output it as an expeditee frame

Line 
1#!@FULL_PERL_EXE@ -w
2
3# Need to specify the full path of Perl above, e.g. for Windows something like
4#!C:\\Perl32\\bin\\perl -w
5
6
7use strict;
8
9# Set this to 1 to work around IIS 6 craziness
10my $iis6_mode = 0;
11
12
13# IIS 6: for some reason, IIS runs this script with the working directory set to the Greenstone
14#   directory rather than the cgi-bin directory, causing lots of stuff to fail
15if ($iis6_mode)
16{
17    # Change into cgi-bin directory
18    chdir("cgi-bin");
19}
20
21
22# We use require and an eval here (instead of "use") to catch any errors loading the module (for IIS)
23eval("require \"gsdlCGI.pm\"");
24if ($@)
25{
26    print STDOUT "Content-type:text/plain\n\n";
27    print STDOUT "ERROR: $@\n";
28    exit 0;
29}
30
31
32sub generate_html_form
33{
34    my ($isGSDL2,$site,$collect,$cl) = @_;
35
36    my $html_form = <<EOT;
37<html>
38  <head>
39    <title>HTML to Expeditee Frames</title>
40
41    <base href=".." />
42
43    <link type="text/css" href="ext/html-to-expeditee/jquery/css/le-frog/jquery-ui-1.8.16.custom.css" rel="stylesheet" />
44    <script type="text/javascript" src="ext/html-to-expeditee/jquery/js/jquery-1.6.2.min.js"></script>
45    <script type="text/javascript" src="ext/html-to-expeditee/jquery/js/jquery-ui-1.8.16.custom.min.js"></script>
46    <script type="text/javascript" src="ext/html-to-expeditee/js/gsajax-min.js"></script>
47    <script type="text/javascript" src="ext/html-to-expeditee/js/html-to-expeditee.js"></script>
48
49
50
51  </head>
52  <body>
53    <form class="ui-widget">
54      Convert the collection
55        <input type="text"
56           class="ui-corner-all" style="padding: 4px;"
57           name="collect" value="$collect" id="collect" />
58      to Expeditee frames by traversing the classifier
59        <input type="text"
60           class="ui-corner-all" style="padding: 4px;"
61               name="cl" value="$cl" id="cl" />
62
63        <input value="$site" id="site" type="hidden">
64        <input value="Go" id="go"
65               class="ui-button ui-widget ui-state-default ui-corner-all"
66               type="submit">
67
68    </form>
69    <script type="text/javascript">
70      var docOIDs = [];
71      var numDocOIDs;
72
73      \$(function() {
74      \$('#progress').progressbar();
75     
76      \$('#go').button().click(function() {
77          var collect = document.getElementById("collect").value;
78          if (collect.match(/^\\s*\$/)) {
79          alert("No collection specified");
80          return false;
81          }
82
83          var cl = document.getElementById("cl").value;
84          if (cl.match(/^\\s*\$/)) {
85          alert("No classifier specified");
86          return false;
87          }
88          if (cl.match(/^\\d+\$/)) {
89              cl = "CL" + cl;
90          }
91         
92          var gs2=$isGSDL2;
93
94          var url;
95          if (gs2) {
96              url = "library.cgi";
97          url += "?c="+collect +"&a=d&cl=" + cl;
98          }
99          else {
100              url = "dev";
101              url += "?c="+collect +"&a=b&rt=s&s=ClassifierBrowse&cl=" + cl;
102              url += "&excerptid=gs_content";
103          }
104
105
106          /* processing animation */
107         
108          docOIDs = [];
109          var outstandingURLs = [];
110          outstandingURLs.push(url);
111
112          while (outstandingURLs.length>0) {
113          url = outstandingURLs.shift();
114         
115          var clHtml = urlGetSync(url);
116                 
117          var aElems;
118          var workingTrav = document.getElementById("workingTraverse");
119          workingTrav.innerHTML = clHtml;
120          aElems = workingTrav.getElementsByTagName("a");
121         
122         
123          /* any links with a=d ... cl=??? => outstandingURLS */
124          /* any links with a=d ... d=???  => docOIDS */
125         
126          var actionRE = new RegExp("(\\\\?|&)a=(?:d|b)(&|\$)");
127          var clRE     = new RegExp("(\\\\?|&)cl=" + cl + "\\\\.");
128          var docRE    = new RegExp("(?:\\\\?|&)d=(.*?)(?:&|\$)");
129
130          for (var i=0; i<aElems.length; i++) {
131              var aElem = aElems[i];
132              var href=aElem.href;
133              if (href && href.match(actionRE)) {
134              if (href.match(clRE)) {
135                  outstandingURLs.push(href);
136              }
137              else if (href.match(docRE)) {
138
139                  var docMatch = docRE.exec(href);
140                  var docOID = docMatch[1];
141
142                  docOIDs.push(docOID);
143              }
144              }
145          }
146          }
147
148          numDocOIDs = docOIDs.length;
149
150          var iframe = document.getElementById("iframe");
151
152          var docOID = docOIDs.shift();
153          console.log("doc oid = " + docOID);
154         
155          var url;
156          if (gs2) {
157              url = "library.cgi";
158              url += "?c="+collect +"&a=d&d=" + docOID;
159          }
160          else {
161              url = "dev";
162              url += "?c="+collect +"&a=d&d=" + docOID;
163              url += "&ed=1&dt=hierarchy";
164              // url += "&excerptid=gs_content";
165          }
166         
167          /* load iframe with document drawn from Greenstone collection */
168          iframe.src = url;
169         
170          var progressbar = document.getElementById("progressbar");
171          progressbar.style.display = "block";
172
173
174          return false; });
175     });
176     
177      function pageLoaded()
178      {
179      var iframe = document.getElementById("iframe");     
180     
181      if (iframe.src) {
182          if (iframe.style.display != "block") {
183          iframe.height = "90%";
184          iframe.style.display = "block";
185          }
186
187          var iframeDoc;
188          if ( iframe.contentDocument )
189          { /* FF */
190            iframeDoc = iframe.contentDocument;
191          }
192          else if ( iframe.contentWindow )
193          { /* IE */
194            iframeDoc = iframe.contentWindow.document;
195          }
196
197          var frameID = (numDocOIDs - docOIDs.length);
198          var progressPercent = frameID/numDocOIDs * 100;
199
200          \$(function() {
201          \$('#progressbar').progressbar({ value: progressPercent })
202         });
203
204          var gsContent = iframeDoc.getElementById("gs_content");
205
206          /*
207          var gsContentPos = getElementPosition(gsContent);
208          var pxl = gsContentPos.xl;
209          var pxr = gsContentPos.xr;
210          */
211
212          var expFrameTree = htmlToExpeditee(gsContent);
213
214          var expFrame = JSON.stringify(expFrameTree);
215          //alert(expFrame);
216
217          var collect = document.getElementById("collect").value;
218          var site    = document.getElementById("site").value;
219
220          var url = "cgi-bin/html-to-expeditee.pl";
221          var params = "c=" + collect;
222          if (site.match(/\\w/)) {
223        params += "&site=" + site;
224          }
225          params += "&a=generate-frame&fn=" + frameID;
226          params += "&json=" + escape(expFrame);
227
228          var clHtml = urlPostSync(url,params);
229
230          if (!clHtml.match(/html-to-expeditee saved frame/)) {
231          alert("Error processing url: " + url);
232          }
233
234          if (docOIDs.length>0) {
235          var docOID = docOIDs.shift();
236         
237          // console.log("doc oid = " + docOID);
238
239         // is the following line used ????
240          var cl = document.getElementById("cl").value;
241         
242          var gs2 = $isGSDL2;
243              var url;
244
245              if (gs2) {
246                  url = "library.cgi";
247                  url += "?c="+collect +"&a=d&d=" + docOID;
248              }
249              else {
250                  url = "dev";
251                  url += "?c="+collect +"&a=d&d=" + docOID;
252                  url += "&ed=1&dt=hierarchy";
253                  // url += "&excerptid=gs_content";
254              }
255
256          iframe.src = url;
257          }
258          else {
259          var progressbar = document.getElementById("progressbar");
260          progressbar.style.display = "none";
261
262          iframe.style.display = "none";
263          delete iframe.src;
264          }
265      }
266     
267      }
268
269    </script>
270
271    <div id="progressbar" width="100%"
272          style="display: none; margin: 10px; height: 10px;"></div>
273    <div id="workingTraverse" style="display: none"></div>
274
275    <hr style="margin: 10px;">
276
277    <iframe width="100%" id="iframe" style="display: none;"
278            onload="pageLoaded()"></iframe>
279
280  </body>
281</html>
282EOT
283
284    print "Content-type:text/html\n\n";
285    print $html_form;
286}
287
288sub main
289{
290
291#    $ENV{'QUERY_STRING'} = "a=...";set-import-metadata&c=espresso-music&d=HASH012d6f72cde5dc48162f4a1d.1&metaname=annotation&metapos=0&metavalue=adfadfad";
292#    $ENV{'REQUEST_METHOD'} = "GET";
293
294    my $gsdl_cgi = new gsdlCGI();
295
296    # Load the Greenstone modules that we need to use
297    $gsdl_cgi->setup_gsdl();
298
299    my $gsdlhome = $ENV{'GSDLHOME'};   
300    $gsdl_cgi->checked_chdir($gsdlhome);
301
302    require cgiactions::HtmlToExpediteeAction;
303   
304    # Useful debug statement for seeing what packages have been included
305####    printf("%-45s%-s\n",$_,$INC{$_}) foreach (sort keys %INC);
306   
307    $gsdl_cgi->parse_cgi_args();
308
309    # We don't want the gsdlCGI module to return errors and warnings in XML
310    $gsdl_cgi->{'xml'} = 0;
311
312    my $fn = $gsdl_cgi->clean_param("fn");     # frame number
313
314    if (defined $fn) {
315
316    my $action = new HtmlToExpediteeAction($gsdl_cgi,$iis6_mode);
317
318    $action->do_action();
319
320    }
321    else {
322    # generate form, pre-filled out with any useful values such
323    # as the collection and classifier value
324
325    my $collect = $gsdl_cgi->clean_param("c");   
326    my $cl = $gsdl_cgi->clean_param("cl");   
327
328    # Establish collect_dir using defining 'site' along the way if GS3
329
330    my $site = undef;
331    my $isGSDL2 = undef;
332
333    if ($gsdl_cgi->greenstone_version() == 2) {
334        $isGSDL2 = 1;
335    }
336    else {
337        $isGSDL2 = 0;
338       
339        # GS3 (and possible future versions) make use of 'site'
340        $site = $gsdl_cgi->clean_param("site");   
341        if (!defined $site) {
342        $gsdl_cgi->generate_error("No site specified.");
343        }
344    }
345
346    generate_html_form($isGSDL2,$site,$collect,$cl);
347    }
348}
349
350
351
352&main();
Note: See TracBrowser for help on using the browser.