Gitiles
Code Review Sign In
korap.ids-mannheim.de / ids-kl / derekovecs / 4750d87924557e3a152bc5db0f00a55969ce6b38 / . / templates / index.html.ep
blob: dda695d9e11af264cf0d39b55d442528084b41d7 [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 Kupietze871abd2018-01-25 16:18:27 +0100[diff] [blame]84 if(ui.newTab.index() == 2 && !collocatorTable_activated) {
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
218 if (collocatorData != null) {
219 maxHeat = Math.max.apply(Math,collocatorData.map(function(o){return Math.max.apply(Math,o.heat);}))
220 collocatorTable = $('#secondtable').DataTable({
221 data: collocatorData,
222 "sScrollY": "780px",
223 "bScrollCollapse": true,
224 "bPaginate": false,
225 "bJQueryUI": true,
226 "dom": '<"top">rt<"bottom"flp><"clear">',
227 "columns": [
228 { "data": "rank", type: "allnumeric" },
229 { "data": "pos", width: "7%", sClass: "dt-center mono compact", render: function ( data, type, row ) {return bitvec2window(data, row.heat, row.word) }},
230 { "data": "max", render: function ( data, type, row ) {return data.toFixed(3) }},
231 { "data": "average", render: function ( data, type, row ) {return data.toFixed(3) }},
232 { "data": "prob", type: "scientific", render: function ( data, type, row ) {return data.toExponential(3) } },
233 { "data": "cprob", type: "scientific", render: function ( data, type, row ) {return data.toExponential(3) } },
234 { "data": "overall", type: "scientific", render: function ( data, type, row ) {return data.toExponential(3) } },
235 { "data": "word", sClass: "collocator" },
236 { "data": "rank", type: "allnumeric" }
237 ],
238 "columnDefs": [
239 { className: "dt-right", "targets": [0,2,3,4,5,6] },
240 { className: "dt-center", "targets": [ 1] },
241 { "searchable": false,
242 "orderable": false,
243 "targets": [0, 8]
244 },
245 { "type": "scientific", targets: [2,3,4,5,6] },
246 { "orderSequence": [ "desc" ], "targets": [ 2, 3, 4, 5, 6 ] },
247 { "orderSequence": [ "asc", "desc" ], "targets": [ 1, 7 ] },
248 { "targets": [8], "visible": false }
249 ],
250 "oLanguage": {
251 "sSearch": "Filter: "
252 },
253 "order": [[ 4, 'desc' ]],
254 } );
255 $.ajaxSetup({
256 type: 'POST',
257 timeout: 30000,
258 error: function(xhr) {
259 $('#display_error')
260 .html('Error: ' + xhr.status + ' ' + xhr.statusText);
261 }
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]262 });
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]263
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]264
265 if($('#sprofiles').length) {
266 similarProfileTable = $('#sprofiles').DataTable({
267 ajax: {
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]268 method: "GET",
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]269 url: '/derekovecs/getSimilarProfiles',
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]270 dataType: 'json',
271 dataSrc: "",
272 timeout: 30000,
273 data: { w: paraResults[0][0].rank }
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]274 },
275 "initComplete":function(settings, json){
276 $('td.paradigmator a').on('mousedown', function(e) {
277 if (e.which === 2) {
278 e.preventDefault();
279 queryKorAPalternatives(paraResults[0][0].word, this.childNodes["0"].textContent);
280 return false;
281 }
282 });
283 },
284 "sScrollY": "780px",
285 "bScrollCollapse": true,
286 "bPaginate": false,
287 "bJQueryUI": true,
288 "dom": '<"top">rt<"bottom"flp><"clear">',
289 "columns": [
290 { "data": "v", render: function ( data, type, row ) {return data.toFixed(3) }},
291 { "data": "w", sClass: "paradigmator", render: function ( data, type, row ) {urlprefix.set("word", data); return '<a href="?' + urlprefix + '">' + data + '</a>' } }
292 ],
293 "columnDefs": [
294 { className: "dt-right", "targets": [0] },
295 ],
296 "oLanguage": {
297 "sSearch": "Filter: "
298 },
299 "order": [[ 0, 'desc' ]],
300 });
301 }
302 // var filterQuot = /(^quot?=[A-Z])|(quot$)/g;
303 var filterQuot = /^quot/;
Marc Kupietz4784fae2019-02-04 12:32:12 +0100[diff] [blame]304 var ccResult;
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]305 var baseURL = window.location.pathname.replace(/[/]$/, '')
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]306 classicCollocatorTable = $('#classicoloctable').DataTable({
307 ajax: {
308 method: "GET",
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]309 url: baseURL + '/getClassicCollocators',
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]310 dataType: 'json',
Marc Kupietz4784fae2019-02-04 12:32:12 +0100[diff] [blame]311 dataSrc: function (result) {
312 ccResult = result;
313 $("#rawfreq_tt").attr("title", "Raw frequencies of the co-occurence.\n"
314 + "f(" + ccResult.w1+")=" + ccResult.f1.toLocaleString("en-GB") + ";\n"
315 + "N=" + ccResult.N.toLocaleString("en-GB"));
316 return result.collocates;
317 },
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]318 timeout: 30000,
319 data: { w: paraResults[0][0].rank }
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]320 },
Marc Kupietzad783722018-01-13 17:45:21 +0100[diff] [blame]321 "initComplete":function(settings, json){
322 $("td.collocator").click(function(){
323 queryKorAPCII(this.textContent + " /w5 " + CIIsearchWords);
324 });
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]325 classicCollocatorTable.columns(".detail").visible(false);
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]326 $("#ccd").css('width', '680px');
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]327 classicCollocatorTable.columns.adjust().draw();
Marc Kupietzad783722018-01-13 17:45:21 +0100[diff] [blame]328 },
Marc Kupietz9c1a3d22019-01-23 10:04:18 +0100[diff] [blame]329 "createdRow": function (row, data, rowIndex) {
330 $.each($('td.collocator', row), function (colIndex) {
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]331 $(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]332 });
333 },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]334 "sScrollY": "780px",
335 "bScrollCollapse": true,
336 "bPaginate": false,
337 "bJQueryUI": true,
338 "dom": '<"top">rt<"bottom"flp><"clear">',
339 "columns": [
340 // { "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]341 { "data": "llr", render: function ( data, type, row ) {return toFixed(data, 0) }},
342 { "data": "pmi", render: function ( data, type, row ) {return toFixed(data, 2) }},
343 { "data": "md", render: function ( data, type, row ) {return toFixed(data, 2) }},
344 { "data": "lfmd", render: function ( data, type, row ) {return toFixed(data, 2) }},
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]345 // { "data": "fpmi", type: "scientific", render: function ( data, type, row ) {return data.toExponential(2) } },
Marc Kupietz1d96a082019-02-18 09:29:06 +0100[diff] [blame]346 { "data": "npmi", render: function ( data, type, row ) {return toFixed(data, 3) }},
347 { "data": "dice", render: function ( data, type, row ) {return toExponential(data, 2) }},
348 { "data": "ld", render: function ( data, type, row ) {return toFixed(data, 2) }},
349 { "data": "af", render: function ( data, type, row ) {return toFixed(data, 2) }},
350 { "data": "delta", render: function ( data, type, row ) {return toFixed(data, 2) }},
Marc Kupietz29f5bbd2019-01-24 15:12:59 +0100[diff] [blame]351 { "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]352 { "data": "llfmd", render: function ( data, type, row ) {return toFixed(data, 1) }},
353 { "data": "rlfmd", render: function ( data, type, row ) {return toFixed(data, 1) }},
354 { "data": "lnpmi", render: function ( data, type, row ) {return toFixed(data, 2) }},
355 { "data": "rnpmi", render: function ( data, type, row ) {return toFixed(data, 2) }},
Marc Kupietzb4b53ca2018-01-25 08:57:11 +0100[diff] [blame]356 { "data": "f" },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]357 { "data": "word", sClass: "collocator" }
358 ],
359 "columnDefs": [
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]360 { className: "dt-right", "targets": [0,1,2,3,4,5,6,7,8,10,11,12,13,14] },
361 { className: "dt-right detail", "targets": [5,10,11,12,13] },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]362 { "searchable": false,
363 "orderable": false,
364 "targets": []
365 },
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]366 { "orderSequence": [ "desc" ], "targets": [0,1,2,3,4,5,6,7,10,11,12,13,14] },
367 { "orderSequence": [ "asc", "desc" ], "targets": [8,9,15] },
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]368 ],
369 "oLanguage": {
370 "sSearch": "Filter: "
371 },
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]372 "order": [[ 7, 'desc' ]],
Marc Kupietz78b434a2018-01-12 22:33:32 +0100[diff] [blame]373 });
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]374
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]375 $('#show-details').change(function (e) {
376 var columns = classicCollocatorTable.columns(".detail");
377 if(this.checked) {
378 columns.visible(true);
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]379 $("#ccd").css('width', 'auto');
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]380 } else {
381 columns.visible(false);
Marc Kupietz08bfd402019-01-23 10:05:17 +0100[diff] [blame]382 $("#ccd").css('width', '680px');
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]383 }
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]384 classicCollocatorTable.columns.adjust().draw();
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]385 } );
386
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]387 $("td.collocator").click(function(){
388 queryKorAPCII(this.textContent + " /w5 " + CIIsearchWords);
Marc Kupietzad783722018-01-13 17:45:21 +0100[diff] [blame]389 });
390
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]391 collocatorTable.on( 'order.dt search.dt', function () {
392 collocatorTable.column(0, {order:'applied'}).nodes().each( function (cell, i) {
393 cell.innerHTML = i+1;
394 } );
395 }).draw();
396 }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]397
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]398 if (localStorage && !window.location.hash) { // let's not crash if some user has IE7
399 var index = parseInt(localStorage['tab']||'0');
400 $("#tabs").tabs({ active: index });
401 }
402 $("#tabs").css("visibility", "visible"); // now we can show the tabs
403 });
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]404
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]405 $(function(){
406 $("#dropdownoptions").dialog({
407 title: "Options",
408 autoOpen: false,
409 modal: false,
410 draggable: false,
411 height: "auto",
412 width: "auto",
413 resizable: false,
414 buttons: {
415 "Cancel": function() {
416 $( this ).dialog( "close" );
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]417 },
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]418 "Apply": function() {
419 window.open($(location).attr('pathname')+'?'+$('form').serialize(), "_self");
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]420 }
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]421 }
422 });
423 });
424
425 $(function(){
426 $("#showoptions").click(function(){
427 $("#dropdownoptions").dialog("open");
428 var target = $(this);
429 $("#dropdownoptions").dialog("widget").position({
430 my: 'left bottom',
431 at: 'left bottom',
432 of: target
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]433 });
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]434 });
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]435 });
436
437 $( function() {
438 $( "#no_iterations" ).spinner({
439 spin: function( event, ui ) {
440 if ( ui.value < 1000 ) {
441 $( this ).spinner( "value", 1000 );
442 return false;
443 } else if ( ui.value > 10000 ) {
444 $( this ).spinner( "value", 10000 );
445 return false;
446 }
447 }
448 });
449 } );
450
451 $( function() {
452 $( "#neighbours" ).spinner({
453 spin: function( event, ui ) {
454 if ( ui.value < 0 ) {
455 $( this ).spinner( "value", 0 );
456 return false;
457 } else if ( ui.value > 200 ) {
458 $( this ).spinner( "value", 200 );
459 return false;
460 }
461 }
462 });
463 } );
464
465 $( function() {
466 $( "#cutoff" ).spinner({
467 spin: function( event, ui ) {
468 if ( ui.value < 100000 ) {
469 $( this ).spinner( "value", 100000 );
470 return false;
471 } else if ( ui.value > 2000000 ) {
472 $( this ).spinner( "value", 2000000 );
473 return false;
474 }
475 }
476 });
477 } );
478
479 $( function() {
480 $( "#tabs" ).tabs().addClass('tabs-min');
481 } );
482
483 $( function() {
484 $( ".controlgroup-vertical" ).controlgroup({
485 "direction": "vertical"
486 });
487 } );
488
489 $(function() {
490 $( document ).tooltip({
491 content: function() {
492 return $(this).attr('title');
493 }}
494 )
495 });
496
497 $(function () {
498 $(document).tooltip({
499 content: function () {
500 return $(this).prop('title');
501 },
502 show: null,
503 close: function (event, ui) {
504 ui.tooltip.hover(
505 function () {
506 $(this).stop(true).fadeTo(400, 1);
507 },
508 function () {
509 $(this).fadeOut("400", function () {
510 $(this).remove();
511 })
512 });
513 }
514 });
515 });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]516 </script>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]517 <script src="//d3js.org/d3.v3.min.js" charset="utf-8"></script>
518 <script src="/derekovecs/js/tsne.js"></script>
519 <script src="/derekovecs/js/som.js"></script>
520 <script src="/derekovecs/js/labeler.js"></script>
Marc Kupietzcdd51302018-03-19 16:36:54 +0100[diff] [blame]521 <script src="/derekovecs/js/derekovcs.js"></script>
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]522 <script>
523
524 var opt = {epsilon: <%= $epsilon %>, perplexity: <%= $perplexity %>},
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]525 mapWidth = 800, // width map
526 mapHeight = 800,
527 jitterRadius = 7;
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]528
529 var T = new tsnejs.tSNE(opt); // create a tSNE instance
530
531 var Y;
532
533 var data;
534 var labeler;
535
536 function applyJitter() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]537 svg.selectAll('.tsnet')
538 .data(labels)
539 .transition()
540 .duration(50)
541 .attr("transform", function(d, i) {
542 T.Y[i][0] = (d.x - mapWidth/2 - tx)/ss/20;
543 T.Y[i][1] = (d.y - mapHeight/2 - ty)/ss/20;
544 return "translate(" +
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]545 (d.x) + "," +
546 (d.y) + ")";
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]547 });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]548 }
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]549
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]550 function updateEmbedding() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]551 var Y = T.getSolution();
552 svg.selectAll('.tsnet')
553 .data(data.words)
554 .attr("transform", function(d, i) {
555 return "translate(" +
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]556 ((Y[i][0]*20*ss + tx) + mapWidth/2) + "," +
557 ((Y[i][1]*20*ss + ty) + mapHeight/2) + ")"; });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]558 }
559
560 var svg;
561 var labels = [];
562 var anchor_array = [];
563 var text;
564
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]565 function getMergedClass(i) {
566 if(data.mergedEnd && i > data.mergedEnd) {
567 return " merged"
568 } else {
569 return "";
570 }
571 }
572
573 function getRankTooltip(i) {
574 if(data.mergedEnd) {
575 if(data.ranks[i] < data.mergedEnd) {
576 return "rank: "+i +" "+"freq. rank: "+(data.ranks[i]).toString().replace(/\B(?=(\d{3})+(?!\d))/g, ",");
577 } else {
578 return "rank: "+i +" "+"freq. rank: "+(data.ranks[i]-data.mergedEnd).toString().replace(/\B(?=(\d{3})+(?!\d))/g, ",") + " (merged vocab)";
579 }
580 } else {
581 return "rank: "+i +" "+"freq. rank: "+data.ranks[i].toString().replace(/\B(?=(\d{3})+(?!\d))/g, ",");
582 }
583 }
584
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]585 function drawEmbedding() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]586 var urlprefix = new URLSearchParams(window.location.search);
587 urlprefix.delete("word");
588 urlprefix.append("word","");
Marc Kupietz8f9c86a2017-12-04 17:17:13 +0100[diff] [blame]589
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]590 $("#embed").empty();
591 var div = d3.select("#embed");
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]592
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]593 // get min and max in each column of Y
594 var Y = T.Y;
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]595
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]596 svg = div.append("svg") // svg is global
597 .attr("width", mapWidth)
598 .attr("height", mapHeight);
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]599
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]600 var g = svg.selectAll(".b")
601 .data(data.words)
602 .enter().append("g")
603 .attr("class", "tsnet");
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]604
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]605 g.append("a")
606 .attr("xlink:href", function(word) {
607 return "?"+urlprefix+word; })
608 .attr("class", function(d, i) {
609 var res="";
610 if(data.marked[i]) {
611 res="marked ";
612 }
613 if(data.target.indexOf(" "+d+" ") >= 0) {
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]614 res += "target";
615 }
616 if(data.mergedEnd && data.ranks[i] >= data.mergedEnd) {
617 return res+" merged";
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]618 } else {
619 return res;
620 }
621 })
622 .attr("title", function(d, i) {
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]623 return getRankTooltip(i);
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]624 })
625 .append("text")
626 .attr("text-anchor", "top")
627 .attr("font-size", 12)
628 .text(function(d) { return d; });
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]629
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]630 g.append("svg:title")
631 .text(function(d, i) {
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]632 return getRankTooltip(i);
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]633 });
634
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]635 var zoomListener = d3.behavior.zoom()
636 .scaleExtent([0.1, 10])
637 .center([0,0])
638 .on("zoom", zoomHandler);
639 zoomListener(svg);
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]640 }
641
642 var tx=0, ty=0;
643 var ss=1;
644 var iter_id=-1;
645
646 function zoomHandler() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]647 tx = d3.event.translate[0];
648 ty = d3.event.translate[1];
649 ss = d3.event.scale;
650 updateEmbedding();
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]651 }
652
653 var stepnum = 0;
654
655 function stopStep() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]656 clearInterval(iter_id);
657 text = svg.selectAll("text");
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]658
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]659 // jitter function needs different data and co-ordinate representation
660 labels = d3.range(data.words.length).map(function(i) {
661 var x = (T.Y[i][0]*20*ss + tx) + mapWidth/2;
662 var y = (T.Y[i][1]*20*ss + ty) + mapHeight/2;
663 anchor_array.push({x: x, y: y, r: jitterRadius});
664 return {
665 x: x,
666 y: y,
667 name: data.words[i]
668 };
669 });
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]670
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]671 // get the actual label bounding boxes for the jitter function
672 var index = 0;
673 text.each(function() {
674 labels[index].width = this.getBBox().width;
675 labels[index].height = this.getBBox().height;
676 index += 1;
677 });
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]678
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]679
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]680 // setTimeout(updateEmbedding, 1);
681 // setTimeout(
682 labeler = d3.labeler()
683 .label(labels)
684 .anchor(anchor_array)
685 .width(mapWidth)
686 .height(mapHeight)
687 .update(applyJitter);
688 // .start(1000);
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]689
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]690 iter_id = setInterval(jitterStep, 1);
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]691 }
692
693 var jitter_i=0;
694
695 function jitterStep() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]696 if(jitter_i++ > 100) {
697 clearInterval(iter_id);
698 } else {
699 labeler.start2(10);
700 applyJitter();
701 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]702 }
703
704 var last_cost=1000;
705
706 function step() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]707 var i = T.iter;
Marc Kupietz34c08172017-11-29 17:08:47 +0100[diff] [blame]708
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]709 if(i > <%= $no_iterations %>) {
710 stopStep();
711 } else {
712 var cost = Math.round(T.step() * 100000) / 100000; // do a few steps
713 $("#cost").html("tsne iteration " + i + ", cost: " + cost.toFixed(5));
714 if(i % 250 == 0 && cost >= last_cost) {
715 stopStep();
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]716 } else {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]717 last_cost = cost;
718 updateEmbedding();
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]719 }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]720 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]721 }
722
723 function showMap(j) {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]724 data=j;
725 T.iter=0;
726 iter_id = -1;
727 last_cost=1000;
728 T.initDataRaw(data.vecs); // init embedding
729 drawEmbedding(); // draw initial embedding
Marc Kupietz78114532017-11-29 17:00:16 +0100[diff] [blame]730
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]731 if(iter_id >= 0) {
732 clearInterval(iter_id);
733 }
734 //T.debugGrad();
735 iter_id = setInterval(step, 1);
736 if(true) { // (<%= $show_som %>) {
737 makeSOM(j, <%= $no_iterations %>);
738 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]739 }
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]740 var queryword;
741
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]742 function showCollocatorSOM() {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]743 var baseURL = window.location.pathname.replace(/[/]$/, '')
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]744 if (collocatorTable) {
745 var ctableData = collocatorTable.rows().data();
746 var nwords = [],
747 nranks = [];
748 for (var i=0; i < ctableData.length && i < 100; i++) {
749 nranks.push(ctableData[i].rank);
750 nwords.push(ctableData[i].word);
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]751 }
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]752 $.post(baseURL+'/getVecsByRanks',
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]753 JSON.stringify(nranks),
754 function(data, status){
Marc Kupietz2e2e4db2018-12-21 15:07:37 +0100[diff] [blame]755 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]756 }, 'json');
757 }
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]758 }
759
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]760 function onload() {
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]761 queryword = document.getElementById('word');
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]762 }
763
764 function queryKorAP() {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]765 window.open('http://korap.ids-mannheim.de' + korapPath + '?q='+queryword.value, 'KorAP');
Marc Kupietz39179ab2017-07-04 16:28:06 +0200[diff] [blame]766 }
Marc Kupietz4dc270c2017-11-24 10:17:12 +0100[diff] [blame]767
768 function queryKorAPCII(query) {
Marc Kupietz86b50292019-02-17 21:03:59 +0100[diff] [blame]769 window.open('http://korap.ids-mannheim.de' +korapPath + '?ql=cosmas2&q='+query, 'KorAP');
Marc Kupietz4dc270c2017-11-24 10:17:12 +0100[diff] [blame]770 }
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]771 </script>
772 </head>
Marc Kupietz7b0e5de2019-01-31 12:49:32 +0100[diff] [blame]773 <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>
774 <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>
775 <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>
776 <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>
777 <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>
778 <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>
779 <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>
780 <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]781 <div id="header">
782 <div id="pagetitle">
Marc Kupietzc053d972019-01-10 10:41:51 +0100[diff] [blame]783 <h1>DeReKoVecs</h1>
784 <h2><%== $title %></h2>
Marc Kupietz4fcda0c2017-11-29 09:00:31 +0100[diff] [blame]785 </div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]786 <div id="options" class="widget">
787 <form id="queryform">
788 <input id="word" type="text" name="word" placeholder="Word(s) to be searched" value="<%= $word %>"
789 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."/>
790 <input id="SEARCH" type="button" value="SEARCH">
791 <input type="button" id="showoptions" name="showoptions" value="Options" />
792 </form>
793 <div id="dropdownoptions" style="display: none">
794 <form id="optionsform">
795 <div class="controlgroup-vertical">
796 <label for="cutoff">cut-off</label>
797 <input id="cutoff" type="text" name="cutoff" size="10" value="<%= $cutoff %>" title="Only consider the most frequent x word forms.">
798 <label for="dedupe">dedupe</label>
799 <input id="dedupe" type="checkbox" name="dedupe" value="1" <%= ($dedupe ? "checked" : "") %> title="radically filter out any near-duplicates">
800 % if($mergedEnd > 0) {
801 <label for="sbf">backw.</label>
802 <input id="sbf" type="checkbox" name="sbf" value="1" <%= ($searchBaseVocabFirst ? "checked" : "") %> title="If checkecked base vocabulary will be searched first. Otherwise merged vocabulray will be searched first.">
803 % }
804 <label for="neighbours">max. neighbours:</label>
805 <input id="neighbours" size="4" name="n" value="<%= $no_nbs %>">
806 <label for="no_iterations">max. iterations</label>
807 <input id="no_iterations" name="N" size="4" value="<%= $no_iterations %>">
808 <!-- <label for="dosom">SOM</label>
809 <input id="dosom" type="checkbox" name="som" value="1" <%= ($show_som ? "checked" : "") %>> -->
810 % if($collocators) {
811 <label for="sortby">window/sort</label>
812 <select id="sortby" name="sort">
813 <option value="0" <%= ($sort!=1 && $sort!=2? "selected":"") %>>auto focus</option>
814 <!-- <option value="1" <%= ($sort==1? "selected":"") %>>any single position</option>
815 <option value="2" <%= ($sort==2? "selected":"") %>>whole window</option> -->
816 </select>
817 % }
818 <input type="button" value="→ KorAP" onclick="queryKorAP();" title="query word with KorAP"/>
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]819 <input id="show-details" type="checkbox" name="show-details" value="1" >
820 <label for="show-details">
821 Show details
822 </label>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]823 </div>
824 </form>
825 </div>
826 </div>
827 </div>
828 <div id="topwrapper">
829 <div style="visibility: hidden;" id="tabs">
830 <ul>
831 <li><a href="#tabs-1">Semantics (TSNE-map)</a></li>
832 <li><a href="#tabs-2">Semantics (SOM)</a></li>
Marc Kupietzbb3da6a2019-01-22 15:21:38 +0100[diff] [blame]833 <li><a href="#tabs-3">Syntagmatic (collocates)</a></li>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]834 </ul>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]835 <div id="tabs-1" style="display: flex; padding: 5px; flex-flow: row wrap;">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]836 % if($lists && (@$lists) > 0 && (@$lists)[0]) {
837 <div id="wrapper">
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]838 <div id="first" style="width: 230px; margin-bottom: 15px;">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]839 <table class="display compact nowrap" id="firsttable">
840 <thead>
841 <tr>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]842 <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]843 </tr>
844 </thead>
845 <tbody>
846 <tr>
847 <td align="right">
848 </td>
849 <td align="right">
850 </td>
851 <td></td>
852 </tr>
853 </tbody>
854 </table>
855 </div>
Marc Kupietza55a09a2018-07-06 13:30:30 +0200[diff] [blame]856 % if(0 && $haveSProfiles) {
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]857 <div id="sprofilesdiv" style="width: 200px; padding-right: 10px;">
858 <table class="display compact nowrap" id="sprofiles">
859 <thead>
860 <tr>
861 <th align="right">cos</th><th align="left">similars by coll. profile</th>
862 </tr>
863 </thead>
864 <tbody>
865 <tr>
866 <td align="right">
867 </td>
868 <td></td>
869 </tr>
870 </tbody>
871 </table>
872 </div>
873 %}
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]874 <div id="second">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]875 <div id="embed">
876 </div>
877 <div id="cost">
878 </div>
879 </div>
880 </div>
Marc Kupietz81aeed22019-02-17 21:22:45 +0100[diff] [blame]881 % } elsif($word !~ /^\s*$/) {
882 <div id="wrapper">
883 <div id="not-found-dialog" title="Not found">
884 <p>ERROR: "<%= $word %>" not found in vocabluary.</p>
885 <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>
886 </div>
887 <script>
888 $( function() {
889 $( "#not-found-dialog" ).dialog({
890 autoOpen: true,
891 modal: true,
892 draggable: false,
893 height: "auto",
894 width: "auto",
895 resizable: false,
896 buttons: {
897 "OK": function() {
898 $( this ).dialog( "close" );
899 },
900 "Apply": function() {
901 window.open($(location).attr('pathname')+'?'+$('form').serialize(), "_self");
902 }
903 }
904 });
905 });
906 </script>
907 </div>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]908 % }
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]909 </div>
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]910 <div id="tabs-2" style="display: flex; padding: 5px; flex-flow: row wrap;">
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]911 <div id="som2" style="width: 800;">
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]912 <div id="sominfo1"><span id="somcolor1"> </span> <span id="somword1"> </span> <span id="somcolor2"> </span> <span id="somword2"> </span> <span id="somcolor3"> </span></div>
913 <div id="sominfo" style="text-align: right">SOM iteration <span id="iterations">0</span></div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]914 </div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]915 </div>
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]916 <div id="tabs-3" style="display: flex; padding:5px; flex-flow: row wrap;">
917 <div style="margin-right: 20px; margin-bottom: 10px;" id="secondt">
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]918 <table class="display compact nowrap" id="secondtable">
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]919 <thead>
920 <tr>
921 % if($collocators) {
922 <th>#</th>
Marc Kupietz4116b432017-12-06 14:15:32 +0100[diff] [blame]923 <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>
924 <th align="right" title="Maximum activation of the collocator anywhere in the output layer.">max(a)</th>
925 <th title="Average raw activation of the collocator in the columns selected by auto-focus." align="right">⟨a⟩</th>
926 <th title="Sum of activations over the selected colunns normalized by the total activation sum of the selected columns." align="right">Σa/Σw'</th>
927 <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]928 <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]929 <th align="left">collocate (by w2v)</th>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]930 % }
931 </tr>
932 </thead>
933 <tbody>
934 <tr>
935 <td align="right">
936 </td>
937 <td align="right">
938 </td>
939 <td align="right">
940 </td>
941 <td align="right">
942 </td>
943 <td align="right">
944 </td>
945 <td align="right">
946 </td>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]947 </tr>
948 </tbody>
949 </table>
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]950 </div>
Marc Kupietzdeb8f622018-03-12 09:41:00 +0100[diff] [blame]951 <div id="ccd" style="">
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]952 <table class="display compact nowrap" id="classicoloctable">
953 <thead>
Marc Kupietzcd136e22018-01-29 09:18:12 +0100[diff] [blame]954 % if($collocators) {
955 <tr>
Marc Kupietzf19814b2019-01-21 16:50:19 +0100[diff] [blame]956 <th id="ll_tt">LL</th>
Marc Kupietzfab14a72019-01-22 16:20:29 +0100[diff] [blame]957 <th id="mi_tt">MI</th>
958 <th id="md_tt">MI²</th>
Marc Kupietzf19814b2019-01-21 16:50:19 +0100[diff] [blame]959 <th id="lfmd_tt">MI³</th>
Marc Kupietzfab14a72019-01-22 16:20:29 +0100[diff] [blame]960 <th id="npmi_tt">nPMI</th>
961 <th id="dice_tt">dice</th>
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]962 <th id="logdice_tt">LD</th>
963 <th id="logdiceaf_tt">LDaf</th>
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]964 <th id="delta_tt" title="Delta to log-Dice score in reference corpus.">Δ</th>
Marc Kupietz7b0e5de2019-01-31 12:49:32 +0100[diff] [blame]965 <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]966 <th title="PMI³ restricted to left neighbour">l-PMI³</th>
967 <th title="PMI³ restricted to right neighbour">r-PMI³</th>
968 <th title="nPMI restricted to left neighbour">l-nPMI</th>
969 <th title="nPMI restricted to right neighbour">r-nPMI</th>
Marc Kupietz4784fae2019-02-04 12:32:12 +0100[diff] [blame]970 <th id="rawfreq_tt" title="raw frequency sum of collocation within window">raw</th>
Marc Kupietzbb3da6a2019-01-22 15:21:38 +0100[diff] [blame]971 <th>collocate (by CA)</th>
Marc Kupietzcd136e22018-01-29 09:18:12 +0100[diff] [blame]972 </tr>
Marc Kupietz1acb3172018-02-17 09:41:26 +0100[diff] [blame]973 % }
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]974 </thead>
975 <tbody>
976 <tr>
Marc Kupietz63b268f2019-01-22 22:34:29 +0100[diff] [blame]977 <td align="right"></td>
978 <td align="right"></td>
979 <td align="right"></td>
Marc Kupietzb4b53ca2018-01-25 08:57:11 +0100[diff] [blame]980 <td align="right"></td>
Marc Kupietz78d7c1a2019-02-17 22:31:22 +0100[diff] [blame]981 <td align="right"></td>
982 <td align="right"></td>
983 <td align="right"></td>
984 <td align="right"></td>
985 <td align="right"></td>
986 <td align="left"> </td>
987 <td align="right"></td>
988 <td align="right"></td>
989 <td align="right"></td>
990 <td align="right"></td>
991 <td align="left"> </td>
Marc Kupietze6a7a732018-01-12 09:21:08 +0100[diff] [blame]992 </tr>
993 </tbody>
994 </table>
995 </div>
Marc Kupietz384c9132018-03-19 16:45:24 +0100[diff] [blame]996 <!--
997 <div style="clear:both" ></div>
998 <div style="float: right; overflow: hidden" id="extra"><button onClick="showCollocatorSOM()"> </button></div>
999 -->
Marc Kupietz66bfd952017-12-11 09:59:45 +0100[diff] [blame]1000 </div>
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]1001 </div>
1002 </div> <!-- topwrapper -->
1003 <div style="clear: both;"></div>
1004 </div>
Marc Kupietzebea4702018-01-12 09:16:09 +0100[diff] [blame]1005 % if($training_args && (@$lists)[0]) {
1006 <div class="info">
1007 % if($training_args =~ /-type\s*3/) {
1008 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 %>
1009 % } else {
1010 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 %>
1011 % }
1012 </div>
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]1013 % }
Marc Kupietz58270662017-12-04 12:10:06 +0100[diff] [blame]1014 </body>
Marc Kupietz83305222016-04-28 09:57:22 +0200[diff] [blame]1015</html>