Gitiles
Code Review Sign In
korap.ids-mannheim.de / ids-kl / derekovecs / d7760b4d53069026fcb99397a096795a9490c6aa / . / templates / index.html.ep
blob: 418fdc7f3aa054aebe46065f1788e18bba11f3d3 [file] [log] [blame]
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]1<!DOCTYPE html>
2<html>
3 <head>
Marc Kupietzc053d972019-01-10 10:41:51 +0100[diff] [blame]4 <% my $plain_title = $title; $plain_title=~s/<[^>]+>//g; %>
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]5 <title><%= $plain_title %>:<%= $word %> · IDS word vector analysis</title>
Marc Kupietz80bd7b92017-07-04 16:25:54 +0200[diff] [blame]6 <link rel="stylesheet" href="//code.jquery.com/ui/1.12.1/themes/base/jquery-ui.css">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]7 <link href="https://fonts.googleapis.com/css?family=Lato|Roboto+Condensed" rel="stylesheet">
Marc Kupietz80bd7b92017-07-04 16:25:54 +0200[diff] [blame]8 <script src="http://code.jquery.com/jquery-latest.min.js"></script>
Marc Kupietz5f532b32019-01-23 10:03:18 +0100[diff] [blame]9 <script src = "https://cdn.datatables.net/1.10.18/js/jquery.dataTables.min.js"></script>
10 <script src = "https://cdn.datatables.net/fixedcolumns/3.2.5/js/dataTables.fixedColumns.min.js"></script>
11 <script src = "https://cdn.datatables.net/plug-ins/1.10.18/sorting/scientific.js"></script>
12 <script src='https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.5/MathJax.js?config=TeX-MML-AM_CHTML'></script>
13 <link rel="stylesheet" href="https://cdn.datatables.net/1.10.18/css/jquery.dataTables.min.css">
Marc Kupietz4949d232018-03-19 16:43:18 +0100[diff] [blame]14 <link rel="stylesheet" href="/derekovecs/css/derekovecs.css">
Marc Kupietz80bd7b92017-07-04 16:25:54 +0200[diff] [blame]15 <script
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]16 src="http://code.jquery.com/ui/1.12.1/jquery-ui.min.js"
17 integrity="sha256-VazP97ZCwtekAsvgPBSUwPFKdrwD3unUfSGVYrahUqU="
18 crossorigin="anonymous"></script>
Marc Kupietz80bd7b92017-07-04 16:25:54 +0200[diff] [blame]19 <script>
Marc Kupietza0ffb392018-01-25 08:53:43 +0100[diff] [blame]20 MathJax.Hub.Config({
21 config: ["MMLorHTML.js"],
22 jax: ["input/TeX","input/MathML","output/HTML-CSS","output/NativeMML", "output/PreviewHTML"],
23 extensions: ["tex2jax.js","mml2jax.js","MathMenu.js","MathZoom.js", "fast-preview.js", "AssistiveMML.js", "a11y/accessibility-menu.js"],
24 TeX: {
25 extensions: ["AMSmath.js","AMSsymbols.js","noErrors.js","noUndefined.js"]
26 }
27 });
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]28 var urlParams = new URLSearchParams(window.location.search);
Marc Kupietzb3a2e4f2017-12-08 17:25:53 +0100[diff] [blame]29 var currentWords = urlParams.get("word");
Marc Kupietza0ffb392018-01-25 08:53:43 +0100[diff] [blame]30 var CIIsearchWords = (currentWords && currentWords.includes(" ") ? '('+currentWords.replace(/ +/g, " oder ")+')' : currentWords);
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]31 var collocatorTable = null;
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]32 var plainTitle ="<%= $plain_title %>"
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]33 var korapPath="/";
34 if (plainTitle.match(/-en/)) {
35 korapPath="/instance/english";
36 }
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]37 $(document).ready(function() {
Marc Kupietz694610d2017-11-25 18:30:03 +0100[diff] [blame]38
Marc Kupietza0ffb392018-01-25 08:53:43 +0100[diff] [blame]39 $('#firstable').hide();
40 //Set up a callback to hear back when MathJax is done rendering the equations
41 // it finds
42 $('#ccd').load(
43 '@Url.Action("ActionResultMethod","ControllerName",{controller parameters})',
44 function () {
45 MathJax.Hub.Queue(
46 ["Typeset",MathJax.Hub,"ccd"],
47 function () {
Marc Kupietz7bd55232019-01-22 15:29:06 +0100[diff] [blame]48 $("#mi_tt").attr("title",$("#pmi_ttt").html());
Marc Kupietza0ffb392018-01-25 08:53:43 +0100[diff] [blame]49 $("#lfmd_tt").attr("title",$("#lfmd_ttt").html());
Marc Kupietz7bd55232019-01-22 15:29:06 +0100[diff] [blame]50 $("#md_tt").attr("title",$("#md_ttt").html());
Marc Kupietza0ffb392018-01-25 08:53:43 +0100[diff] [blame]51 $("#npmi_tt").attr("title",$("#npmi_ttt").html());
Marc Kupietzf19814b2019-01-21 16:50:19 +0100[diff] [blame]52 $("#ll_tt").attr("title",$("#ll_ttt").html());
Marc Kupietz7bd55232019-01-22 15:29:06 +0100[diff] [blame]53 $("#logdice_tt").attr("title",$("#logdice_ttt").html());
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]54 $("#logdiceaf_tt").attr("title",$("#logdiceaf_ttt").html());
Marc Kupietza0ffb392018-01-25 08:53:43 +0100[diff] [blame]55 }
56 );
57 });
58
59 //set things up so that we can shove raw html into what is shown in the tooltip;
60 // in this case, we will have already put into the title attribute the html that
61 // contains the MathJax rendered equations (via what we do in the callback).
62 $(function () {
63 $(document).tooltip({
64 content: function () {
65 return $(this).prop('title');
66 }
67 });
68 });
69
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]70 $("input").bind("keydown", function(event) {
71 // track enter key
72 var keycode = (event.keyCode ? event.keyCode : (event.which ? event.which : event.charCode));
73 if (keycode == 13) { // keycode for enter key
74 // force the 'Enter Key' to implicitly click the Update button
75 document.getElementById('SEARCH').click();
76 return false;
77 } else {
78 return true;
79 }});
Marc Kupietzdab9f222017-11-29 14:22:59 +0100[diff] [blame]80
Marc Kupietze871abd2018-01-25 16:18:27 +0100[diff] [blame]81 var collocatorTable_activated = false;
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]82 $( "#tabs" ).on( "tabsactivate", function( event, ui ) {
83 if (localStorage) localStorage['tab'] = ui.newTab.index();
Marc Kupietzd7760b42019-02-21 09:01:44 +0100[diff] [blame^]84 if(ui.newTab.index() == 3 && !collocatorTable_activated) {
Marc Kupietze871abd2018-01-25 16:18:27 +0100[diff] [blame]85 classicCollocatorTable.columns.adjust();
86 collocatorTable.columns.adjust();
87 collocatorTable_activated = true;
88 }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]89 });
90
91 $(function(){
92 $("#SEARCH").click(function() {
93 window.open($(location).attr('pathname')+'?'+$('form').serialize(), "_self");
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]94 });
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]95 });
Marc Kupietz0af83e32017-11-27 09:31:37 +0100[diff] [blame]96
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]97 function changeCharColor(txt, heat, word) {
98 var newText = "";
99 for (var i=0, l=txt.length; i<l; i++) {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]100 newText += (i == 5 ? txt.charAt(i) : '<a href="http://korap.ids-mannheim.de' + korapPath + '/?ql=cosmas2&q=' +
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]101 CIIsearchWords + ' /' + (i > 5? '%2B' : '-') + 'w' +
102 Math.abs(i-5) + ':' + Math.abs(i-5) + ' ' + word +
103 '" target="korap"><span style="background-color:' +
104 getHeatColor(heat[i]/maxHeat)+'">'+txt.charAt(i)+'</span></a>');
Marc Kupietzb6c615d2017-12-02 10:38:20 +0100[diff] [blame]105 }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]106 return newText;
107 }
Marc Kupietzb6c615d2017-12-02 10:38:20 +0100[diff] [blame]108
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]109 function getHeatColor(value) {
110 var hue=((1-value)*120).toString(10);
111 return ["hsl(",hue,",90%,70%)"].join("");
112 }
113
Marc Kupietz29f5bbd2019-01-24 15:12:59 +0100[diff] [blame]114 function bitmask2window(autofocus, win) {
115 var af_str = autofocus.toString(2).padStart(10, "0")
116 .replace(/^([0-9]{5})/, '$1 ')
117 .replace(/0/g, '·')
118 .replace(/1/g, '◼');
119 var w_str = win.toString(2).padStart(10, "0")
120 .replace(/^([0-9]{5})/, '$1 ')
121 .replace(/0/g, '·')
122 .replace(/1/g, 'o');
123 for (var i=0, l=af_str.length; i<l; i++) {
124 if(af_str.charAt(i) == '·' && w_str.charAt(i) == 'o' ) {
125 af_str = af_str.substring(0, i) + "◽" + af_str.substring(i+1);
126 }
127 }
128 return af_str;
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]129 }
130
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]131 function bitvec2window(n, heat, word) {
132 var str = n.toString(2).padStart(10, "0")
133 .replace(/^([0-9]{5})/, '$1x')
134 .replace(/0/g, '·')
135 .replace(/1/g, '+');
136 return changeCharColor(str, heat, word);
137 }
138
Marc Kupietz1d96a082019-02-18 09:29:06 +0100[diff] [blame]139 function toFixed(n, d) {
140 if(typeof n != "number") {
141 return n;
142 } else {
143 return n.toFixed(d);
144 }
145 }
146
147 function toExponential(n, d) {
148 if(typeof n != "number") {
149 return n;
150 } else {
151 return n.toExponential(d);
152 }
153 }
154
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]155 % use Mojo::ByteStream 'b';
156 var paraResults = <%= b(Mojo::JSON::to_json($lists)) %>;
157 var urlprefix = new URLSearchParams(window.location.search);
158 if (paraResults.length > 0 && paraResults[0] != null) {
159 var nvecs = [],
160 nwords = [],
161 nranks = [],
162 nmarked = [];
163 for(var i = 0; i < paraResults.length; i++) {
164 nwords = nwords.concat(paraResults[i].map(function(a){return a.word;}));
165 nvecs = nvecs.concat(paraResults[i].map(function(a){return a.vector;}));
166 nranks = nranks.concat(paraResults[i].map(function(a){return a.rank;}));
167 nmarked = nmarked.concat(paraResults[i].map(function(a){return a.marked;}));
Marc Kupietz0d4c0ca2017-12-04 09:18:56 +0100[diff] [blame]168 }
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]169 showMap({target: " "+urlParams.get('word')+" ", mergedEnd: <%= $mergedEnd %>, words: nwords, vecs: nvecs, ranks: nranks, marked: nmarked} );
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]170 var t = $('#firsttable').DataTable({
Marc Kupietzee894d52019-01-09 14:55:14 +0100[diff] [blame]171 data: [].concat.apply([], paraResults),
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]172 "sScrollY": "780px",
173 "bScrollCollapse": true,
174 "bPaginate": false,
175 "bJQueryUI": true,
176 "dom": '<"top">rt<"bottom"flp><"clear">',
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]177 "initComplete":function(settings, json) {
178 $('td.paradigmator a').on('mousedown', function(e) {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]179 return paradigmatorClick(e, paraResults[0][0].word, this.childNodes["0"].textContent);
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]180 });
181 },
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]182 "columns": [
183 { "data": "rank", type: "allnumeric" },
184 { "data": "dist", render: function ( data, type, row ) {return data.toFixed(3) }},
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]185 { "data": "word", class: "paradigmator", render: function ( data, type, row ) {
186 urlprefix.set("word", data); return '<a class="' + getMergedClass(row.rank) + '" href="?' + urlprefix + '">' + data + '</a>'
187 }}
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]188 ],
189 "columnDefs": [
190 { className: "dt-right", "targets": [0,1] },
191 { "searchable": false,
192 "orderable": false,
193 "targets": 0
194 },
195 { "orderSequence": [ "desc" ], "targets": [ 1 ] },
196 { "orderSequence": [ "asc", "desc" ], "targets": [ 2 ] },
197 ],
198 "oLanguage": {
199 "sSearch": "Filter: "
200 },
201 "order": [[ 1, 'desc' ]],
202 } );
Marc Kupietz0d4c0ca2017-12-04 09:18:56 +0100[diff] [blame]203
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]204 t.on( 'order.dt search.dt', function () {
205 t.column(0, {order:'applied'}).nodes().each( function (cell, i) {
206 cell.innerHTML = i+1;
207 } );
208 } ).draw();
209
210 $( "#first" ).clone().prependTo( "#tabs-2" );
211
212 }
213
214
215 var collocatorData = <%= b(Mojo::JSON::to_json($collocators)) %>;
216 var maxHeat; // = Math.max.apply(Math,collocatorData.map(function(o){return o.cprob;}))
217
Marc Kupietzd7760b42019-02-21 09:01:44 +0100[diff] [blame^]218 vocabDistanceTable = $('#vocabdistt').DataTable({
219 ajax: {
220 method: "GET",
221 url: baseURL + '/getBiggestVocabDistances',
222 dataType: 'json',
223 dataSrc: function (result) {
224 return result;
225 },
226 timeout: 30000,
227 },
228 "initComplete":function(settings, json){
229 vocabDistanceTable.columns.adjust().draw();
230 },
231 "createdRow": function (row, data, rowIndex) {
232 $.each($('td.collocator', row), function (colIndex) {
233 $(this).attr('title', "f("+data.word+")="+data.f2.toLocaleString("en-GB") + " f1: "+ccResult.f1+ " total: "+ccResult.N);
234 });
235 },
236 "sScrollY": "780px",
237 "bScrollCollapse": true,
238 "bPaginate": false,
239 "bJQueryUI": true,
240 "dom": '<"top">rt<"bottom"flp><"clear">',
241 "columns": [
242 { "data": "rank", type: "allnumeric" },
243 { "data": "dist", render: function ( data, type, row ) {return data.toFixed(3) }},
244 { "data": "word", class: "paradigmator", render: function ( data, type, row ) {
245 urlprefix.set("word", data); return '<a class="' + getMergedClass(row.rank) + '" href="?' + urlprefix + '">' + data + '</a>'
246 }}
247 ],
248 "columnDefs": [
249 { className: "dt-right", "targets": [0,1] },
250 { "searchable": false,
251 "orderable": false,
252 "targets": 0
253 },
254 { "orderSequence": [ "desc" ], "targets": [ 1 ] },
255 { "orderSequence": [ "asc", "desc" ], "targets": [ 2 ] },
256 ],
257 "oLanguage": {
258 "sSearch": "Filter: "
259 },
260 "order": [[ 1, 'desc' ]]
261 });
262
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]263 if (collocatorData != null) {
264 maxHeat = Math.max.apply(Math,collocatorData.map(function(o){return Math.max.apply(Math,o.heat);}))
265 collocatorTable = $('#secondtable').DataTable({
266 data: collocatorData,
267 "sScrollY": "780px",
268 "bScrollCollapse": true,
269 "bPaginate": false,
270 "bJQueryUI": true,
271 "dom": '<"top">rt<"bottom"flp><"clear">',
272 "columns": [
273 { "data": "rank", type: "allnumeric" },
274 { "data": "pos", width: "7%", sClass: "dt-center mono compact", render: function ( data, type, row ) {return bitvec2window(data, row.heat, row.word) }},
275 { "data": "max", render: function ( data, type, row ) {return data.toFixed(3) }},
276 { "data": "average", render: function ( data, type, row ) {return data.toFixed(3) }},
277 { "data": "prob", type: "scientific", render: function ( data, type, row ) {return data.toExponential(3) } },
278 { "data": "cprob", type: "scientific", render: function ( data, type, row ) {return data.toExponential(3) } },
279 { "data": "overall", type: "scientific", render: function ( data, type, row ) {return data.toExponential(3) } },
280 { "data": "word", sClass: "collocator" },
281 { "data": "rank", type: "allnumeric" }
282 ],
283 "columnDefs": [
284 { className: "dt-right", "targets": [0,2,3,4,5,6] },
285 { className: "dt-center", "targets": [ 1] },
286 { "searchable": false,
287 "orderable": false,
288 "targets": [0, 8]
289 },
290 { "type": "scientific", targets: [2,3,4,5,6] },
291 { "orderSequence": [ "desc" ], "targets": [ 2, 3, 4, 5, 6 ] },
292 { "orderSequence": [ "asc", "desc" ], "targets": [ 1, 7 ] },
293 { "targets": [8], "visible": false }
294 ],
295 "oLanguage": {
296 "sSearch": "Filter: "
297 },
298 "order": [[ 4, 'desc' ]],
299 } );
300 $.ajaxSetup({
301 type: 'POST',
302 timeout: 30000,
303 error: function(xhr) {
304 $('#display_error')
305 .html('Error: ' + xhr.status + ' ' + xhr.statusText);
306 }
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]307 });
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]308
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]309
310 if($('#sprofiles').length) {
311 similarProfileTable = $('#sprofiles').DataTable({
312 ajax: {
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]313 method: "GET",
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]314 url: '/derekovecs/getSimilarProfiles',
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]315 dataType: 'json',
316 dataSrc: "",
317 timeout: 30000,
318 data: { w: paraResults[0][0].rank }
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]319 },
320 "initComplete":function(settings, json){
321 $('td.paradigmator a').on('mousedown', function(e) {
322 if (e.which === 2) {
323 e.preventDefault();
324 queryKorAPalternatives(paraResults[0][0].word, this.childNodes["0"].textContent);
325 return false;
326 }
327 });
328 },
329 "sScrollY": "780px",
330 "bScrollCollapse": true,
331 "bPaginate": false,
332 "bJQueryUI": true,
333 "dom": '<"top">rt<"bottom"flp><"clear">',
334 "columns": [
335 { "data": "v", render: function ( data, type, row ) {return data.toFixed(3) }},
336 { "data": "w", sClass: "paradigmator", render: function ( data, type, row ) {urlprefix.set("word", data); return '<a href="?' + urlprefix + '">' + data + '</a>' } }
337 ],
338 "columnDefs": [
339 { className: "dt-right", "targets": [0] },
340 ],
341 "oLanguage": {
342 "sSearch": "Filter: "
343 },
344 "order": [[ 0, 'desc' ]],
345 });
346 }
Marc Kupietzd7760b42019-02-21 09:01:44 +0100[diff] [blame^]347
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]348 // var filterQuot = /(^quot?=[A-Z])|(quot$)/g;
349 var filterQuot = /^quot/;
Marc Kupietz4784fae2019-02-04 12:32:12 +0100[diff] [blame]350 var ccResult;
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]351 var baseURL = window.location.pathname.replace(/[/]$/, '')
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]352 classicCollocatorTable = $('#classicoloctable').DataTable({
353 ajax: {
354 method: "GET",
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]355 url: baseURL + '/getClassicCollocators',
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]356 dataType: 'json',
Marc Kupietz4784fae2019-02-04 12:32:12 +0100[diff] [blame]357 dataSrc: function (result) {
358 ccResult = result;
359 $("#rawfreq_tt").attr("title", "Raw frequencies of the co-occurence.\n"
360 + "f(" + ccResult.w1+")=" + ccResult.f1.toLocaleString("en-GB") + ";\n"
361 + "N=" + ccResult.N.toLocaleString("en-GB"));
362 return result.collocates;
363 },
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]364 timeout: 30000,
365 data: { w: paraResults[0][0].rank }
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]366 },
Marc Kupietzad783722018-01-13 17:45:21 +0100[diff] [blame]367 "initComplete":function(settings, json){
368 $("td.collocator").click(function(){
369 queryKorAPCII(this.textContent + " /w5 " + CIIsearchWords);
370 });
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]371 classicCollocatorTable.columns(".detail").visible(false);
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]372 $("#ccd").css('width', '680px');
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]373 classicCollocatorTable.columns.adjust().draw();
Marc Kupietzad783722018-01-13 17:45:21 +0100[diff] [blame]374 },
Marc Kupietz9c1a3d22019-01-23 10:04:18 +0100[diff] [blame]375 "createdRow": function (row, data, rowIndex) {
376 $.each($('td.collocator', row), function (colIndex) {
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]377 $(this).attr('title', "f("+data.word+")="+data.f2.toLocaleString("en-GB") + " f1: "+ccResult.f1+ " total: "+ccResult.N);
Marc Kupietz9c1a3d22019-01-23 10:04:18 +0100[diff] [blame]378 });
379 },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]380 "sScrollY": "780px",
381 "bScrollCollapse": true,
382 "bPaginate": false,
383 "bJQueryUI": true,
384 "dom": '<"top">rt<"bottom"flp><"clear">',
385 "columns": [
386 // { "data": "pos", width: "7%", sClass: "dt-center mono compact", render: function ( data, type, row ) {return bitvec2window(data, row.heat, row.word) }},
Marc Kupietz1d96a082019-02-18 09:29:06 +0100[diff] [blame]387 { "data": "llr", render: function ( data, type, row ) {return toFixed(data, 0) }},
388 { "data": "pmi", render: function ( data, type, row ) {return toFixed(data, 2) }},
389 { "data": "md", render: function ( data, type, row ) {return toFixed(data, 2) }},
390 { "data": "lfmd", render: function ( data, type, row ) {return toFixed(data, 2) }},
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]391 // { "data": "fpmi", type: "scientific", render: function ( data, type, row ) {return data.toExponential(2) } },
Marc Kupietz1d96a082019-02-18 09:29:06 +0100[diff] [blame]392 { "data": "npmi", render: function ( data, type, row ) {return toFixed(data, 3) }},
393 { "data": "dice", render: function ( data, type, row ) {return toExponential(data, 2) }},
394 { "data": "ld", render: function ( data, type, row ) {return toFixed(data, 2) }},
395 { "data": "af", render: function ( data, type, row ) {return toFixed(data, 2) }},
396 { "data": "delta", render: function ( data, type, row ) {return toFixed(data, 2) }},
Marc Kupietz29f5bbd2019-01-24 15:12:59 +0100[diff] [blame]397 { "data": "afwin", width: "auto", sClass: "dt-center mono compact", render: function ( data, type, row ) {return bitmask2window(data, row.win) }},
Marc Kupietz1d96a082019-02-18 09:29:06 +0100[diff] [blame]398 { "data": "llfmd", render: function ( data, type, row ) {return toFixed(data, 1) }},
399 { "data": "rlfmd", render: function ( data, type, row ) {return toFixed(data, 1) }},
400 { "data": "lnpmi", render: function ( data, type, row ) {return toFixed(data, 2) }},
401 { "data": "rnpmi", render: function ( data, type, row ) {return toFixed(data, 2) }},
Marc Kupietzb4b53ca2018-01-25 08:57:11 +0100[diff] [blame]402 { "data": "f" },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]403 { "data": "word", sClass: "collocator" }
404 ],
405 "columnDefs": [
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]406 { className: "dt-right", "targets": [0,1,2,3,4,5,6,7,8,10,11,12,13,14] },
407 { className: "dt-right detail", "targets": [5,10,11,12,13] },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]408 { "searchable": false,
409 "orderable": false,
410 "targets": []
411 },
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]412 { "orderSequence": [ "desc" ], "targets": [0,1,2,3,4,5,6,7,10,11,12,13,14] },
413 { "orderSequence": [ "asc", "desc" ], "targets": [8,9,15] },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]414 ],
415 "oLanguage": {
416 "sSearch": "Filter: "
417 },
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]418 "order": [[ 7, 'desc' ]],
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]419 });
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]420
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]421 $('#show-details').change(function (e) {
422 var columns = classicCollocatorTable.columns(".detail");
423 if(this.checked) {
424 columns.visible(true);
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]425 $("#ccd").css('width', 'auto');
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]426 } else {
427 columns.visible(false);
Marc Kupietz08bfd402019-01-23 10:05:17 +0100[diff] [blame]428 $("#ccd").css('width', '680px');
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]429 }
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]430 classicCollocatorTable.columns.adjust().draw();
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]431 } );
432
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]433 $("td.collocator").click(function(){
434 queryKorAPCII(this.textContent + " /w5 " + CIIsearchWords);
Marc Kupietzad783722018-01-13 17:45:21 +0100[diff] [blame]435 });
436
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]437 collocatorTable.on( 'order.dt search.dt', function () {
438 collocatorTable.column(0, {order:'applied'}).nodes().each( function (cell, i) {
439 cell.innerHTML = i+1;
440 } );
441 }).draw();
442 }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]443
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]444 if (localStorage && !window.location.hash) { // let's not crash if some user has IE7
445 var index = parseInt(localStorage['tab']||'0');
446 $("#tabs").tabs({ active: index });
447 }
448 $("#tabs").css("visibility", "visible"); // now we can show the tabs
449 });
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]450
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]451 $(function(){
452 $("#dropdownoptions").dialog({
453 title: "Options",
454 autoOpen: false,
455 modal: false,
456 draggable: false,
457 height: "auto",
458 width: "auto",
459 resizable: false,
460 buttons: {
461 "Cancel": function() {
462 $( this ).dialog( "close" );
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]463 },
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]464 "Apply": function() {
465 window.open($(location).attr('pathname')+'?'+$('form').serialize(), "_self");
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]466 }
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]467 }
468 });
469 });
470
471 $(function(){
472 $("#showoptions").click(function(){
473 $("#dropdownoptions").dialog("open");
474 var target = $(this);
475 $("#dropdownoptions").dialog("widget").position({
476 my: 'left bottom',
477 at: 'left bottom',
478 of: target
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]479 });
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]480 });
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]481 });
482
483 $( function() {
484 $( "#no_iterations" ).spinner({
485 spin: function( event, ui ) {
486 if ( ui.value < 1000 ) {
487 $( this ).spinner( "value", 1000 );
488 return false;
489 } else if ( ui.value > 10000 ) {
490 $( this ).spinner( "value", 10000 );
491 return false;
492 }
493 }
494 });
495 } );
496
497 $( function() {
498 $( "#neighbours" ).spinner({
499 spin: function( event, ui ) {
500 if ( ui.value < 0 ) {
501 $( this ).spinner( "value", 0 );
502 return false;
503 } else if ( ui.value > 200 ) {
504 $( this ).spinner( "value", 200 );
505 return false;
506 }
507 }
508 });
509 } );
510
511 $( function() {
512 $( "#cutoff" ).spinner({
513 spin: function( event, ui ) {
514 if ( ui.value < 100000 ) {
515 $( this ).spinner( "value", 100000 );
516 return false;
517 } else if ( ui.value > 2000000 ) {
518 $( this ).spinner( "value", 2000000 );
519 return false;
520 }
521 }
522 });
523 } );
524
525 $( function() {
526 $( "#tabs" ).tabs().addClass('tabs-min');
527 } );
528
529 $( function() {
530 $( ".controlgroup-vertical" ).controlgroup({
531 "direction": "vertical"
532 });
533 } );
534
535 $(function() {
536 $( document ).tooltip({
537 content: function() {
538 return $(this).attr('title');
539 }}
540 )
541 });
542
543 $(function () {
544 $(document).tooltip({
545 content: function () {
546 return $(this).prop('title');
547 },
548 show: null,
549 close: function (event, ui) {
550 ui.tooltip.hover(
551 function () {
552 $(this).stop(true).fadeTo(400, 1);
553 },
554 function () {
555 $(this).fadeOut("400", function () {
556 $(this).remove();
557 })
558 });
559 }
560 });
561 });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]562 </script>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]563 <script src="//d3js.org/d3.v3.min.js" charset="utf-8"></script>
564 <script src="/derekovecs/js/tsne.js"></script>
565 <script src="/derekovecs/js/som.js"></script>
566 <script src="/derekovecs/js/labeler.js"></script>
Marc Kupietzcdd51302018-03-19 16:36:54 +0100[diff] [blame]567 <script src="/derekovecs/js/derekovcs.js"></script>
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]568 <script>
569
570 var opt = {epsilon: <%= $epsilon %>, perplexity: <%= $perplexity %>},
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]571 mapWidth = 800, // width map
572 mapHeight = 800,
573 jitterRadius = 7;
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]574
575 var T = new tsnejs.tSNE(opt); // create a tSNE instance
576
577 var Y;
578
579 var data;
580 var labeler;
581
582 function applyJitter() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]583 svg.selectAll('.tsnet')
584 .data(labels)
585 .transition()
586 .duration(50)
587 .attr("transform", function(d, i) {
588 T.Y[i][0] = (d.x - mapWidth/2 - tx)/ss/20;
589 T.Y[i][1] = (d.y - mapHeight/2 - ty)/ss/20;
590 return "translate(" +
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]591 (d.x) + "," +
592 (d.y) + ")";
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]593 });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]594 }
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]595
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]596 function updateEmbedding() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]597 var Y = T.getSolution();
598 svg.selectAll('.tsnet')
599 .data(data.words)
600 .attr("transform", function(d, i) {
601 return "translate(" +
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]602 ((Y[i][0]*20*ss + tx) + mapWidth/2) + "," +
603 ((Y[i][1]*20*ss + ty) + mapHeight/2) + ")"; });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]604 }
605
606 var svg;
607 var labels = [];
608 var anchor_array = [];
609 var text;
610
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]611 function getMergedClass(i) {
Marc Kupietzd7760b42019-02-21 09:01:44 +0100[diff] [blame^]612 if(typeof data !== 'undefined' && i > data.mergedEnd) {
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]613 return " merged"
614 } else {
615 return "";
616 }
617 }
618
619 function getRankTooltip(i) {
620 if(data.mergedEnd) {
621 if(data.ranks[i] < data.mergedEnd) {
622 return "rank: "+i +" "+"freq. rank: "+(data.ranks[i]).toString().replace(/\B(?=(\d{3})+(?!\d))/g, ",");
623 } else {
624 return "rank: "+i +" "+"freq. rank: "+(data.ranks[i]-data.mergedEnd).toString().replace(/\B(?=(\d{3})+(?!\d))/g, ",") + " (merged vocab)";
625 }
626 } else {
627 return "rank: "+i +" "+"freq. rank: "+data.ranks[i].toString().replace(/\B(?=(\d{3})+(?!\d))/g, ",");
628 }
629 }
630
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]631 function drawEmbedding() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]632 var urlprefix = new URLSearchParams(window.location.search);
633 urlprefix.delete("word");
634 urlprefix.append("word","");
Marc Kupietz8f9c86a2017-12-04 17:17:13 +0100[diff] [blame]635
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]636 $("#embed").empty();
637 var div = d3.select("#embed");
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]638
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]639 // get min and max in each column of Y
640 var Y = T.Y;
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]641
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]642 svg = div.append("svg") // svg is global
643 .attr("width", mapWidth)
644 .attr("height", mapHeight);
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]645
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]646 var g = svg.selectAll(".b")
647 .data(data.words)
648 .enter().append("g")
649 .attr("class", "tsnet");
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]650
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]651 g.append("a")
652 .attr("xlink:href", function(word) {
653 return "?"+urlprefix+word; })
654 .attr("class", function(d, i) {
655 var res="";
656 if(data.marked[i]) {
657 res="marked ";
658 }
659 if(data.target.indexOf(" "+d+" ") >= 0) {
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]660 res += "target";
661 }
662 if(data.mergedEnd && data.ranks[i] >= data.mergedEnd) {
663 return res+" merged";
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]664 } else {
665 return res;
666 }
667 })
668 .attr("title", function(d, i) {
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]669 return getRankTooltip(i);
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]670 })
671 .append("text")
672 .attr("text-anchor", "top")
673 .attr("font-size", 12)
674 .text(function(d) { return d; });
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]675
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]676 g.append("svg:title")
677 .text(function(d, i) {
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]678 return getRankTooltip(i);
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]679 });
680
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]681 var zoomListener = d3.behavior.zoom()
682 .scaleExtent([0.1, 10])
683 .center([0,0])
684 .on("zoom", zoomHandler);
685 zoomListener(svg);
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]686 }
687
688 var tx=0, ty=0;
689 var ss=1;
690 var iter_id=-1;
691
692 function zoomHandler() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]693 tx = d3.event.translate[0];
694 ty = d3.event.translate[1];
695 ss = d3.event.scale;
696 updateEmbedding();
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]697 }
698
699 var stepnum = 0;
700
701 function stopStep() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]702 clearInterval(iter_id);
703 text = svg.selectAll("text");
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]704
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]705 // jitter function needs different data and co-ordinate representation
706 labels = d3.range(data.words.length).map(function(i) {
707 var x = (T.Y[i][0]*20*ss + tx) + mapWidth/2;
708 var y = (T.Y[i][1]*20*ss + ty) + mapHeight/2;
709 anchor_array.push({x: x, y: y, r: jitterRadius});
710 return {
711 x: x,
712 y: y,
713 name: data.words[i]
714 };
715 });
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]716
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]717 // get the actual label bounding boxes for the jitter function
718 var index = 0;
719 text.each(function() {
720 labels[index].width = this.getBBox().width;
721 labels[index].height = this.getBBox().height;
722 index += 1;
723 });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]724
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]725
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]726 // setTimeout(updateEmbedding, 1);
727 // setTimeout(
728 labeler = d3.labeler()
729 .label(labels)
730 .anchor(anchor_array)
731 .width(mapWidth)
732 .height(mapHeight)
733 .update(applyJitter);
734 // .start(1000);
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]735
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]736 iter_id = setInterval(jitterStep, 1);
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]737 }
738
739 var jitter_i=0;
740
741 function jitterStep() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]742 if(jitter_i++ > 100) {
743 clearInterval(iter_id);
744 } else {
745 labeler.start2(10);
746 applyJitter();
747 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]748 }
749
750 var last_cost=1000;
751
752 function step() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]753 var i = T.iter;
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]754
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]755 if(i > <%= $no_iterations %>) {
756 stopStep();
757 } else {
758 var cost = Math.round(T.step() * 100000) / 100000; // do a few steps
759 $("#cost").html("tsne iteration " + i + ", cost: " + cost.toFixed(5));
760 if(i % 250 == 0 && cost >= last_cost) {
761 stopStep();
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]762 } else {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]763 last_cost = cost;
764 updateEmbedding();
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]765 }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]766 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]767 }
768
769 function showMap(j) {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]770 data=j;
771 T.iter=0;
772 iter_id = -1;
773 last_cost=1000;
774 T.initDataRaw(data.vecs); // init embedding
775 drawEmbedding(); // draw initial embedding
Marc Kupietz78114532017-11-29 17:00:16 +0100[diff] [blame]776
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]777 if(iter_id >= 0) {
778 clearInterval(iter_id);
779 }
780 //T.debugGrad();
781 iter_id = setInterval(step, 1);
782 if(true) { // (<%= $show_som %>) {
783 makeSOM(j, <%= $no_iterations %>);
784 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]785 }
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]786 var queryword;
787
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]788 function showCollocatorSOM() {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]789 var baseURL = window.location.pathname.replace(/[/]$/, '')
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]790 if (collocatorTable) {
791 var ctableData = collocatorTable.rows().data();
792 var nwords = [],
793 nranks = [];
794 for (var i=0; i < ctableData.length && i < 100; i++) {
795 nranks.push(ctableData[i].rank);
796 nwords.push(ctableData[i].word);
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]797 }
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]798 $.post(baseURL+'/getVecsByRanks',
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]799 JSON.stringify(nranks),
800 function(data, status){
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]801 showMap({target: " "+urlParams.get('word')+" ", mergedEnd: <%= $mergedEnd %>, words: nwords, vecs: data, ranks: nranks, marked: Array(100).fill(false)} );
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]802 }, 'json');
803 }
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]804 }
805
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]806 function onload() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]807 queryword = document.getElementById('word');
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]808 }
809
810 function queryKorAP() {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]811 window.open('http://korap.ids-mannheim.de' + korapPath + '?q='+queryword.value, 'KorAP');
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]812 }
Marc Kupietz4dc270c2017-11-24 10:17:12 +0100[diff] [blame]813
814 function queryKorAPCII(query) {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]815 window.open('http://korap.ids-mannheim.de' +korapPath + '?ql=cosmas2&q='+query, 'KorAP');
Marc Kupietz4dc270c2017-11-24 10:17:12 +0100[diff] [blame]816 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]817 </script>
818 </head>
Marc Kupietz7b0e5de2019-01-31 12:49:32 +0100[diff] [blame]819 <body onload="onload()"> <div style="display:none;" id="pmi_ttt">Pointwise mutual information: $$\text{MI}=\text{MI}=\log_2\frac{p(w_1,w_2)}{p(w_1) p(w_2)}$$<p class="citation">Church, K. W. and Hanks, P. (1990): Word association norms, mutual information, and lexicography. Comput. Linguist. 16, 1 (March 1990), 22-29.</p></div>
820 <div style="display:none;" id="md_ttt">Pointwise mutual information squared [1], also called mutual dependency [2]: $$\text{MI}^2=\text{MD}=\log_2\frac{p^2(w_1,w_2)}{p(w_1) p(w_2)}$$<p class="citation">[1] Daille, B. (1994): <a href="http://www.bdaille.com/index.php?option=com_docman&task=doc_download&gid=8&Itemid=">Approche mixte pour l’extraction automatique de terminologie: statistiques lexicales et filtres linguistiques</a>. PhD thesis, Université Paris 7.</p><p class="citation">[2] Thanopoulos, A., Fakotakis, N., Kokkinakis, G. (2002): <a href="http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.11.8101&rep=rep1&type=pdf">Comparative evaluation of collocation extraction metrics</a>. In: Proc. of LREC 2002: 620–625.</p></div>
821 <div style="display:none;" id="lfmd_ttt">Pointwise mutual information cubed [1], also called log-frequency biased mutual dependency [2]: $$\text{MI}^3=\text{LFMD}=\log_2\frac{p^3(w_1,w_2)}{p(w_1) p(w_2)}$$<p class="citation">[1] Daille, B. (1994): <a href="http://www.bdaille.com/index.php?option=com_docman&task=doc_download&gid=8&Itemid=">Approche mixte pour l’extraction automatique de terminologie: statistiques lexicales et filtres linguistiques</a>. PhD thesis, Université Paris 7.</p><p class="citation">[2] Thanopoulos, A., Fakotakis, N., Kokkinakis, G. (2002): <a href="http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.11.8101&rep=rep1&type=pdf">Comparative evaluation of collocation extraction metrics</a>. In: Proc. of LREC 2002: 620–625.</p></div>
822 <div style="display:none;" id="npmi_ttt">Normalized pointwise mutual information: $$\frac{\log_2\frac{p(w_1,w_2)}{p(w_1)p(w_2)}}{-\log_2(p(w_1,w_2))}$$<p class="citation">Bouma, Gerlof (2009): <a href="https://svn.spraakdata.gu.se/repos/gerlof/pub/www/Docs/npmi-pfd.pdf">Normalized (pointwise) mutual information in collocation extraction</a>. In Proceedings of GSCL.</p></div>
823 <div style="display:none;" id="ll_ttt">Log-likelihood: $$2\sum_{ij}O_{ij}\log\frac{O_{ij}}{E_{ij}}$$<p class="citation">Dunning, T. (1993): Accurate methods for the statistics of surprise and coincidence. Comput. Linguist. 19, 1 (March 1993), 61-74.</p>
824 <p class="citation">Evert, Stefan (2004): <a href="http://purl.org/stefan.evert/PUB/Evert2004phd.pdf">The Statistics of Word Cooccurrences: Word Pairs and Collocations.</a> PhD dissertation, IMS, University of Stuttgart. Published in 2005, URN urn:nbn:de:bsz:93-opus-23714.</p></div>
825 <div style="display:none;" id="logdice_ttt">Log-Dice: $$14 + \log_2 \frac{2f_{1,2}}{f_1 + f_2}$$<p class="citation">Rychlý, Pavel (2008): <a href="http://www.fi.muni.cz/usr/sojka/download/raslan2008/13.pdf">A lexicographer-friendly association score.</a> In Proceedings of Recent Advances in Slavonic Natural Language Processing, RASLAN, 6–9, 2008</p></div>
826 <div style="display:none;" id="logdiceaf_ttt">Log-Dice using "auto-focus", i.e. the window that yields the highest score.</div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]827 <div id="header">
828 <div id="pagetitle">
Marc Kupietzc053d972019-01-10 10:41:51 +0100[diff] [blame]829 <h1>DeReKoVecs</h1>
830 <h2><%== $title %></h2>
Marc Kupietz4fcda0c2017-11-29 09:00:31 +0100[diff] [blame]831 </div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]832 <div id="options" class="widget">
833 <form id="queryform">
834 <input id="word" type="text" name="word" placeholder="Word(s) to be searched" value="<%= $word %>"
835 title="When looking for multiple words use spaces as separators to search around the average vector and | as separator to get the neighbours for each word."/>
836 <input id="SEARCH" type="button" value="SEARCH">
837 <input type="button" id="showoptions" name="showoptions" value="Options" />
838 </form>
839 <div id="dropdownoptions" style="display: none">
840 <form id="optionsform">
841 <div class="controlgroup-vertical">
842 <label for="cutoff">cut-off</label>
843 <input id="cutoff" type="text" name="cutoff" size="10" value="<%= $cutoff %>" title="Only consider the most frequent x word forms.">
844 <label for="dedupe">dedupe</label>
845 <input id="dedupe" type="checkbox" name="dedupe" value="1" <%= ($dedupe ? "checked" : "") %> title="radically filter out any near-duplicates">
846 % if($mergedEnd > 0) {
847 <label for="sbf">backw.</label>
848 <input id="sbf" type="checkbox" name="sbf" value="1" <%= ($searchBaseVocabFirst ? "checked" : "") %> title="If checkecked base vocabulary will be searched first. Otherwise merged vocabulray will be searched first.">
849 % }
850 <label for="neighbours">max. neighbours:</label>
851 <input id="neighbours" size="4" name="n" value="<%= $no_nbs %>">
852 <label for="no_iterations">max. iterations</label>
853 <input id="no_iterations" name="N" size="4" value="<%= $no_iterations %>">
854 <!-- <label for="dosom">SOM</label>
855 <input id="dosom" type="checkbox" name="som" value="1" <%= ($show_som ? "checked" : "") %>> -->
856 % if($collocators) {
857 <label for="sortby">window/sort</label>
858 <select id="sortby" name="sort">
859 <option value="0" <%= ($sort!=1 && $sort!=2? "selected":"") %>>auto focus</option>
860 <!-- <option value="1" <%= ($sort==1? "selected":"") %>>any single position</option>
861 <option value="2" <%= ($sort==2? "selected":"") %>>whole window</option> -->
862 </select>
863 % }
864 <input type="button" value="→ KorAP" onclick="queryKorAP();" title="query word with KorAP"/>
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]865 <input id="show-details" type="checkbox" name="show-details" value="1" >
866 <label for="show-details">
867 Show details
868 </label>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]869 </div>
870 </form>
871 </div>
872 </div>
873 </div>
874 <div id="topwrapper">
875 <div style="visibility: hidden;" id="tabs">
876 <ul>
Marc Kupietzd7760b42019-02-21 09:01:44 +0100[diff] [blame^]877 % if($distantWords) {
878 <li><a href="#tabs-0" title="Cos offsets of the words furthest away from their position in the reference corpus."">Offsets</a></li>
879 % }
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]880 <li><a href="#tabs-1">Semantics (TSNE-map)</a></li>
881 <li><a href="#tabs-2">Semantics (SOM)</a></li>
Marc Kupietzbb3da6a2019-01-22 15:21:38 +0100[diff] [blame]882 <li><a href="#tabs-3">Syntagmatic (collocates)</a></li>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]883 </ul>
Marc Kupietzd7760b42019-02-21 09:01:44 +0100[diff] [blame^]884 <div id="tabs-0" style="display: flex; padding: 5px; flex-flow: row wrap;">
885 <div id="vocabdist" style="width: 230px; margin-bottom: 15px;">
886 <table class="display compact nowrap" id="vocabdistt">
887 <thead>
888 <tr>
889 <th align="right">#</th><th align="right">cos</th><th align="left">word</th>
890 </tr>
891 </thead>
892 <tbody>
893 <tr>
894 <td align="right">
895 </td>
896 <td align="right">
897 </td>
898 <td></td>
899 </tr>
900 </tbody>
901 </table>
902 </div>
903 </div>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]904 <div id="tabs-1" style="display: flex; padding: 5px; flex-flow: row wrap;">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]905 % if($lists && (@$lists) > 0 && (@$lists)[0]) {
906 <div id="wrapper">
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]907 <div id="first" style="width: 230px; margin-bottom: 15px;">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]908 <table class="display compact nowrap" id="firsttable">
909 <thead>
910 <tr>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]911 <th align="right">#</th><th align="right">cos</th><th align="left">similars by w2v</th>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]912 </tr>
913 </thead>
914 <tbody>
915 <tr>
916 <td align="right">
917 </td>
918 <td align="right">
919 </td>
920 <td></td>
921 </tr>
922 </tbody>
923 </table>
924 </div>
Marc Kupietza55a09a2018-07-06 13:30:30 +0200[diff] [blame]925 % if(0 && $haveSProfiles) {
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]926 <div id="sprofilesdiv" style="width: 200px; padding-right: 10px;">
927 <table class="display compact nowrap" id="sprofiles">
928 <thead>
929 <tr>
930 <th align="right">cos</th><th align="left">similars by coll. profile</th>
931 </tr>
932 </thead>
933 <tbody>
934 <tr>
935 <td align="right">
936 </td>
937 <td></td>
938 </tr>
939 </tbody>
940 </table>
941 </div>
942 %}
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]943 <div id="second">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]944 <div id="embed">
945 </div>
946 <div id="cost">
947 </div>
948 </div>
949 </div>
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]950 % } elsif($word !~ /^\s*$/) {
951 <div id="wrapper">
952 <div id="not-found-dialog" title="Not found">
953 <p>ERROR: "<%= $word %>" not found in vocabluary.</p>
954 <p>If you are sure you have spelled the word as intended, you can try to increase the cutoff parameter in the options menu.</p>
955 </div>
956 <script>
957 $( function() {
958 $( "#not-found-dialog" ).dialog({
959 autoOpen: true,
960 modal: true,
961 draggable: false,
962 height: "auto",
963 width: "auto",
964 resizable: false,
965 buttons: {
966 "OK": function() {
967 $( this ).dialog( "close" );
968 },
969 "Apply": function() {
970 window.open($(location).attr('pathname')+'?'+$('form').serialize(), "_self");
971 }
972 }
973 });
974 });
975 </script>
976 </div>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]977 % }
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]978 </div>
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]979 <div id="tabs-2" style="display: flex; padding: 5px; flex-flow: row wrap;">
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]980 <div id="som2" style="width: 800;">
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]981 <div id="sominfo1"><span id="somcolor1"> </span> <span id="somword1"> </span> <span id="somcolor2"> </span> <span id="somword2"> </span> <span id="somcolor3"> </span></div>
982 <div id="sominfo" style="text-align: right">SOM iteration <span id="iterations">0</span></div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]983 </div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]984 </div>
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]985 <div id="tabs-3" style="display: flex; padding:5px; flex-flow: row wrap;">
986 <div style="margin-right: 20px; margin-bottom: 10px;" id="secondt">
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]987 <table class="display compact nowrap" id="secondtable">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]988 <thead>
989 <tr>
990 % if($collocators) {
991 <th>#</th>
Marc Kupietz4116b432017-12-06 14:15:32 +0100[diff] [blame]992 <th align="center" title="Activation of the respective collocator in the columns around the target normalized by its maximum (red). Columns selected by the auto-focus funtion (which window of all possible column-combinations maximizes ⊥(a/c)?) are marked with +. Click on the column postions to lauch a KorAP query with target word and collocator in the respective position.">w'</th>
993 <th align="right" title="Maximum activation of the collocator anywhere in the output layer.">max(a)</th>
994 <th title="Average raw activation of the collocator in the columns selected by auto-focus." align="right">⟨a⟩</th>
995 <th title="Sum of activations over the selected colunns normalized by the total activation sum of the selected columns." align="right">Σa/Σw'</th>
996 <th title="Co-norm of the column-normalized activations over the colunns selected by the auto-focus." align="right">⊥(a/c)</th>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]997 <th title="Sum of the activations over the whole window normalized by the total window sum (no auto-focus)." align="right">Σa/Σw</th>
Marc Kupietzbb3da6a2019-01-22 15:21:38 +0100[diff] [blame]998 <th align="left">collocate (by w2v)</th>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]999 % }
1000 </tr>
1001 </thead>
1002 <tbody>
1003 <tr>
1004 <td align="right">
1005 </td>
1006 <td align="right">
1007 </td>
1008 <td align="right">
1009 </td>
1010 <td align="right">
1011 </td>
1012 <td align="right">
1013 </td>
1014 <td align="right">
1015 </td>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]1016 </tr>
1017 </tbody>
1018 </table>
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]1019 </div>
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]1020 <div id="ccd" style="">
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]1021 <table class="display compact nowrap" id="classicoloctable">
1022 <thead>
Marc Kupietzcd136e22018-01-29 09:18:12 +0100[diff] [blame]1023 % if($collocators) {
1024 <tr>
Marc Kupietzf19814b2019-01-21 16:50:19 +0100[diff] [blame]1025 <th id="ll_tt">LL</th>
Marc Kupietzfab14a72019-01-22 16:20:29 +0100[diff] [blame]1026 <th id="mi_tt">MI</th>
1027 <th id="md_tt">MI²</th>
Marc Kupietzf19814b2019-01-21 16:50:19 +0100[diff] [blame]1028 <th id="lfmd_tt">MI³</th>
Marc Kupietzfab14a72019-01-22 16:20:29 +0100[diff] [blame]1029 <th id="npmi_tt">nPMI</th>
1030 <th id="dice_tt">dice</th>
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]1031 <th id="logdice_tt">LD</th>
1032 <th id="logdiceaf_tt">LDaf</th>
Marc Kupietz001bffd2019-02-21 08:52:41 +0100[diff] [blame]1033 <th id="delta_tt" title="Delta to log-Dice score in reference corpus. ⚠: If the collocate is not within the top 200 of the reference corpus, a reference value of min(lD)-0.1 is assumed.">Δ</th>
Marc Kupietz7b0e5de2019-01-31 12:49:32 +0100[diff] [blame]1034 <th id="af_win" title="Positions around the target word that are selected by the auto-focus function are marked with ◾. Positions at which the collocate appears at least once are marked with ◽.">af-window</th>
Marc Kupietz5a7f9ac2018-01-30 11:22:44 +0100[diff] [blame]1035 <th title="PMI³ restricted to left neighbour">l-PMI³</th>
1036 <th title="PMI³ restricted to right neighbour">r-PMI³</th>
1037 <th title="nPMI restricted to left neighbour">l-nPMI</th>
1038 <th title="nPMI restricted to right neighbour">r-nPMI</th>
Marc Kupietz4784fae2019-02-04 12:32:12 +0100[diff] [blame]1039 <th id="rawfreq_tt" title="raw frequency sum of collocation within window">raw</th>
Marc Kupietzbb3da6a2019-01-22 15:21:38 +0100[diff] [blame]1040 <th>collocate (by CA)</th>
Marc Kupietzcd136e22018-01-29 09:18:12 +0100[diff] [blame]1041 </tr>
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]1042 % }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]1043 </thead>
1044 <tbody>
1045 <tr>
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]1046 <td align="right"></td>
1047 <td align="right"></td>
1048 <td align="right"></td>
Marc Kupietzb4b53ca2018-01-25 08:57:11 +0100[diff] [blame]1049 <td align="right"></td>
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]1050 <td align="right"></td>
1051 <td align="right"></td>
1052 <td align="right"></td>
1053 <td align="right"></td>
1054 <td align="right"></td>
1055 <td align="left"> </td>
1056 <td align="right"></td>
1057 <td align="right"></td>
1058 <td align="right"></td>
1059 <td align="right"></td>
1060 <td align="left"> </td>
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]1061 </tr>
1062 </tbody>
1063 </table>
1064 </div>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]1065 <!--
1066 <div style="clear:both" ></div>
1067 <div style="float: right; overflow: hidden" id="extra"><button onClick="showCollocatorSOM()"> </button></div>
1068 -->
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]1069 </div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]1070 </div>
1071 </div> <!-- topwrapper -->
1072 <div style="clear: both;"></div>
1073 </div>
Marc Kupietzebea4702018-01-12 09:16:09 +0100[diff] [blame]1074 % if($training_args && (@$lists)[0]) {
1075 <div class="info">
1076 % if($training_args =~ /-type\s*3/) {
1077 Calculations are based on a word embedding model trained with an extension of <a href="https://github.com/wlin12/wang2vec/">wang2vec</a> using the following parameters: <%= $training_args %>
1078 % } else {
1079 Calculations are based on a word embedding model trained with <a href="https://code.google.com/p/word2vec/">word2vec</a> using the following parameters: <%= $training_args %>
1080 % }
1081 </div>
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]1082 % }
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]1083 </body>
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]1084</html>