Code tree reorganized; older implemenatations act as a start point

1 files changed, 294 insertions, 109 deletions

diff --git a/html/search.html b/html/search.html
index 2a12941..5c5e64b 100644
--- a/html/search.html
+++ b/html/search.html
@@ -4,9 +4,9 @@
         <meta charset="utf-8">
 
         <style>
-body { font-family: 'Cantarell', Helvetica, Arial, sans-serif; }
+body { font-family: 'Cantarell', Helvetica, Arial, sans-serif; margin: 2em; }
 
-.twitter-typeahead { width: 87% ;}
+.twitter-typeahead { width: 87%; }
 .typeahead, .tt-query, .tt-hint {
     width: 100%; height: 30px;
     padding: 8px 12px; outline: none;
@@ -26,9 +26,20 @@ body { font-family: 'Cantarell', Helvetica, Arial, sans-serif; }
     box-shadow: 0 5px 10px rgba(0,0,0,.2);
 }
 .tt-suggestion { padding: 3px 20px; line-height: 24px; font-size: 18px; }
+.tt-suggestion:hover { cursor: pointer; }
 .tt-suggestion span { padding-left: 16px; font-size: 14px; color: #777; float: right; }
 .tt-cursor { background: #ddd; }
 .tt-highlight { font-weight: normal; color: #777; }
+.tt-hint { color: #9598; }
+
+#selections { width: 87%; padding-top: 40px; }
+#selections div { padding: 4px 40px; line-height: 24px; font-size: 18px; color: #666; }
+#selections div span { padding-left: 16px; font-size: 14px; color: #999; float: right; }
+.-info- { font-size: 12px !important; color: #959 !important;
+    font-family: 'JetBrains Mono NL', Consolas, Monaco, monospace, fixed !important;
+    line-height: 14px !important; border-bottom: 1px solid #ddda;
+}
+.-info- b { font-weight: 900; }
         </style>
 
         <!-- js labraries -->
@@ -42,55 +53,73 @@ body { font-family: 'Cantarell', Helvetica, Arial, sans-serif; }
             <input class="typeahead" id="tagsInput" type="text" placeholder="try me!">
         </div>
 
+        <div id="selections">
+        </div>
+
     </body>
     <script>
 
-// arrays for kb-format (utf/EL-Gr to ascii translation)
-var ORiGiN = 'ςερτυθιοπασδφγηξκλζχψωβνμΕΡΤΥΘΙΟΠΑΣΔΦΓΗΞΚΛΖΧΨΩΒΝΜάέήίόύώϊϋΆΈΉΊΌΎΏΪΫQWERTYUIOPASDFGHJKLZXCVBNM'.split('');
-var kbKeyZ = 'sertyuiopasdfghjklzxcvbnmertyuiopasdfghjklzxcvbnmaehioyviyaehioyviyqwertyuiopasdfghjklzxcvbnm'.split('');
 
-// convert string to kb-format
-function kb_trans(s) {
-    var charArr = s.split('')
-    var i
+// PUBLIC VARIABLES ////////////////////////////////////////////////////////////
+
+var _kwlinks;       // keyword links (word-connections; imported via ajax-get)
+var _products;      // all products (imported via ajax-get)
+
+var trackSearch = [];   // searching analytics
+
+// setup options
+var maxResults = 24;    // limit suggestions
+var blendProds = 4;     // minimum final-produncts to blend with next-word suggestions
+var allowFuzzy = false; // enable|disable fuzzy search
+var bi1stScore = .65;   // bigram minimum match score for 1st word
+var bi2ndScore = .85;   // bigram minimum match score for 2nd word
+
+
+var ignoredKeys_kb = [];    // keywords to ignore (in kb-format)
+['μας με σε για του της των από στο στον &'].split(' ').forEach(w => { ignoredKbs.push(kb_trans(w)); });
+
+
+// SUPLAMENARY FUNCTIONS ///////////////////////////////////////////////////////
+
+// arrays for kb-format (utf/EL-Gr to ascii translation)
+var ORiGiNal = 'ςερτυθιοπασδφγηξκλζχψωβνμΕΡΤΥΘΙΟΠΑΣΔΦΓΗΞΚΛΖΧΨΩΒΝΜάέήίόύώϊϋΆΈΉΊΌΎΏΪΫQWERTYUIOPASDFGHJKLZXCVBNMqwertyuiopasdfghjklzxcvbnm0123456789-'.split('');
+var kbKeyZed = 'sertyuiopasdfghjklzxcvbnmertyuiopasdfghjklzxcvbnmaehioyviyaehioyviyqwertyuiopasdfghjklzxcvbnmqwertyuiopasdfghjklzxcvbnm0123456789-'.split('');
+const map = new Map();
+for (var i=0; i<ORiGiNal.length; i++) map.set(ORiGiNal[i], kbKeyZed[i]);
+// function kb_trans
+// translates string to keyboard-latin keys;
+// [map]'s implementation is 40x faster than [for]'s
+function kb_trans(str) {
+    str = str.replace('\'','');
     var out = ''
-    charArr.forEach( el => {
-        i = 0
-        exist = -1
-        ORiGiN.forEach( ori => {
-            if (ori == el) {
-                exist = i;
-            }
-            i++;
-        })
-        out += (exist == -1) ? el : kbKeyZ[exist];
-    });
+    for (var i=0 ; i< str.length; i++)  out += map.get(str[i]);
     return out;
 }
 
-// public data objects
-var products;
-var everyProduct;
-
-function loadData() {
-  const xhttp = new XMLHttpRequest();
-  xhttp.onload = function() {
-    products = JSON.parse(this.responseText);
-  }
-  xhttp.open("GET", "results/keywords-v3.json");
-  xhttp.send();
-}
-loadData();
-
-function loadData2() {
-  const xhttp = new XMLHttpRequest();
-  xhttp.onload = function() {
-    everyProduct = JSON.parse(this.responseText);
-  }
-  xhttp.open("GET", "results/products.json");
-  xhttp.send();
+
+
+// pure JS ajax GET request; return data as JSON
+// no fancy things like UTF8; if needed use base64
+// ---
+function ajax_get(url, callback) {
+    var xmlhttp = new XMLHttpRequest();
+    xmlhttp.onreadystatechange = function() {
+        if (xmlhttp.readyState == 4 && xmlhttp.status == 200) {
+
+            try {
+                var data = JSON.parse(xmlhttp.responseText);
+            } catch(err) {
+                console.log(err.message + " in " + xmlhttp.responseText);
+                return;
+            }
+            callback(data);
+        }
+    };
+
+    xmlhttp.open("GET", url, true);
+    xmlhttp.send();
 }
-loadData2();
+
 
 // bigram fuzzy match
 // --- credit: https://dirask.com/posts/JavaScript-check-words-similarity-fuzzy-compare-with-bigrams-paola1
@@ -98,11 +127,13 @@ const createBigram = word => {
     const input = word.toLowerCase();
     const vector = [];
     for (let i = 0; i < input.length; ++i) {
-      	vector.push(input.slice(i, i + 2));
+        vector.push(input.slice(i, i + 2));
     }
     return vector;
 };
 const checkSimilarity = (a, b) => {
+    if (!allowFuzzy)    return false;
+
     if (a.length > 0 && b.length > 0) {
         const aBigram = createBigram(a);
         const bBigram = createBigram(b);
@@ -112,27 +143,110 @@ const checkSimilarity = (a, b) => {
                 if (aBigram[x] === bBigram[y]) {
                     hits += 1;
                 }
-          	}
+            }
         }
         if (hits > 0) {
-          	const union = aBigram.length + bBigram.length;
-        	return (2.0 * hits) / union;
+            const union = aBigram.length + bBigram.length;
+          return (2.0 * hits) / union;
         }
     }
     return 0;
 };
-var biMMS = .6;		// bigram minimum match score
 
-// on document ready code /////////////////////////////////////////////////////
+
+// compact print of search-tracking
+// ---
+function echo_tracking() {
+    var actions = [];
+    var c;
+    var chs = 0;    // number of characters pressed;
+    var uis = 0;    // number of UI actions used (arrows, enters etc.)
+    var countingStarted = false;      // flag
+    trackSearch.forEach(e => {
+        switch(e.v) {
+            // use of ui actions
+            case 'ArrowDown' :  c = '↓';  uis++;  break;
+            case 'ArrowUp'   :  c = '↑';  uis++;  break;
+            case 'Enter'     :  c = '↲ ';  uis++;  break;
+            case 'ArrowLeft' :  c = '←';  uis++;  break;
+            case 'ArrowRight':  c = '→';  uis++;  break;
+            case ' '      :  c = '· ';  uis++;  break;
+            // ignored keys
+            case 'Alt'      :  c = 'Alt';   break;
+            case 'Control'  :  c = 'Ctrl';  break;
+            case 'Escape'  :  c = 'Esc';  break;
+            case 'Shift'  :  c = 'Shft';  break;
+            case 'Home'  :  c = 'Home';  break;
+            case 'End'  :  c = 'End';  break;
+            // backspace (user's typing errors)
+            case 'Backspace' :  c = 'BkSp';  break;
+            case 'Delete'  :  c = 'Del';  break;
+            // actual typed characters
+            default:
+                if (e.v.length == 1) {
+                    c = '<b><u>'+ e.v +'</u></b>';
+                    chs++;
+                }
+                else {  // some non important key; no counter increased
+                    c = e.v;    // just record the key
+                }
+        }
+        actions.push(c);
+    });
+
+    return actions.join(',') +' (<u>'+ chs +' chs</u>, '+ uis +' uis)';
+}
+
+
+// workline object to track status of async svents
+// ---
+var workline = {
+
+    trackerJL : 0,
+    set jsonLoaded(x) {
+        this.trackerJL = x;
+
+        // fire event on certain values
+        if (x == 2) {
+            console.log('All streams loaded');
+
+            // code to execute
+            // ...
+
+        }
+    },
+    get jsonLoaded() { return this.trackerJL; }
+
+};
+
+
+
+// LOAD DATA ///////////////////////////////////////////////////////////////////
+
+ajax_get('results/keywords-v3.json', function(data) {
+    _kwlinks = data;
+    workline.jsonLoaded++; console.log('keywords loaded');
+});
+
+ajax_get('results/products.json', function(data) {
+    _products = data;
+    workline.jsonLoaded++; console.log('products loaded')
+});
+
+
+
+
+
+// on document ready code ///////////////////////////////////////////////////////
 $(document).ready(function() { 
 
-	// suggestions engine ////////////////////////////////////////////////////
+    // suggestions engine //////////////////////////////////////////////////////
     // ---
-    var isuggest = function(qOrig, list) {
-        var results = [];   // suggestions to respond
-        var proList = [];   // list of products (for all suggestions)
-        var commonL = [];   // list of common products (for multiple suggestions)
-        var possibleNext = [];  // list of possible next suggestions
+    function suggestions_engine(qOrig) {
+        var results = [];       // suggestions to respond
+        var proList = [];       // list of products (for all suggestions)
+        var commonL = [];       // list of common products (for multiple suggestions)
+        var possibleNext = [];      // list of possible next suggestions
 
         var root, last;
 
@@ -140,34 +254,43 @@ $(document).ready(function() {
         var q = qOrig.replace('+',' ').replace('.',' ')
                 .replace('   ',' ')
                 .replace('  ',' ');
-        // split to words
-        var qAr = q.split(' ');
-        // console.log(qAr);
+        
+        var qAr = q.split(' ');     // split to words
 
         if (qAr.length == 1) {      // suggest 1st word ////////////////////////
-			var kbq = kb_trans(q)
-            // regex match all possible suggestions; (in kb-format)
-            substrRegex = new RegExp( kbq, 'i');    // match q anywhere
-            products.forEach( it => {
-                if ( (substrRegex.test(it.kb))
-					|| (checkSimilarity(it.kb, kbq) > biMMS) ) {
-					// console.log(it.kb, kbq, checkSimilarity(it.kb, kbq));
+        var kbq = kb_trans(q)
+
+            // ### (-) regex match all possible suggestions; (in kb-format)
+            // ### (-) substrRegex = new RegExp( kbq, 'i');    // match q anywhere
+            _kwlinks.forEach( it => {
+                // ### (-) if ( (substrRegex.test(it.kb))
+                if ( (it.kb.indexOf(kbq) !== -1)
+                   || (checkSimilarity(it.kb, kbq) > bi1stScore) ) {
+                    // a match found
                     results.push(it);
+                    // also get possible products
+                    // (but only if products are less max-results)
+                    if (proList.length < maxResults + 2) {
+                        it.c.forEach( wo => { proList = proList.concat(wo.p); });
+                        // keep unique products in the list
+                        proList = proList.filter((item, i, ar) => ar.indexOf(item) === i);
+                    }
                 }
             });
         }
 
-        if (qAr.length == 2) {	 // suggest 2nd word ///////////////////////////
+        if (qAr.length == 2) {   // suggest 2nd word ///////////////////////////
             root = qAr[0].trim();
             kbroot = kb_trans(root);
 
-            substrRegex = new RegExp( kb_trans(qAr[1]), 'i');
+            // ### (-) substrRegex = new RegExp( kb_trans(qAr[1]), 'i');
 
-            products.forEach( it => { 	// loop through suggestions
-                if (it.kb == kbroot ) {		// match 1st suggestion
-                    it.c.forEach ( wo => {		// regex match linked words
-                        if ( (substrRegex.test(wo.kb))
-							|| (checkSimilarity(wo.kb, kbroot) > biMMS) ) {
+            _kwlinks.forEach( it => {   // loop through suggestions
+                if (it.kb == kbroot ) {    // match 1st suggestion
+                    it.c.forEach ( wo => {    // regex match linked words
+                        // ### (-) if ( (substrRegex.test(wo.kb))
+                        if ( (wo.kb.indexOf(qAr[1]) !== -1)
+                           || (checkSimilarity(wo.kb, kbroot) > bi2ndScore) ) {
                             results.push({
                                 w: root +' '+ wo.w,
                                 f: 100
@@ -180,21 +303,22 @@ $(document).ready(function() {
         }
 
         if (qAr.length > 2) {
-            root = qAr.shift();     // get out the first item of qAr
+            root = qAr.shift();     // get out the first item of tempQAr
             last = qAr.pop();       // get out the lase item of qAr
                 // now qAr includes only the items after root and before last;
                 // so qAr includes all already selected suggestions (but root)
             
+            // prepare/cache kb-formats for any key we may need
+            // ---
             var kbqAr = []; // array of selected suggestions in kb-format
             qAr.forEach( w => { kbqAr.push(kb_trans(w)); })
-
-            // kb-translate the root/last keys
+            // kb-translate the root/last keys too
             kbroot = kb_trans(root);
             kblast = kb_trans(last);
             
-            substrRegex = new RegExp( kblast, 'i'); // construct regex for mathing
+            // ### (-) substrRegex = new RegExp( kblast, 'i'); // construct regex for mathing
 
-            products.forEach( it => {
+            _kwlinks.forEach( it => {
                 if (it.kb == kbroot ) {     // find root
 
                     // calculate list of common items/products (commonL)
@@ -202,18 +326,19 @@ $(document).ready(function() {
                     // ---
                     is1stOcc = true;       // 1st occurance flag
                     it.c.forEach ( swo => {
-                        if (kbqAr.includes( swo.kb )) {
+                        if (kbqAr.indexOf( swo.kb ) !== -1) {
                             // swo is one of the already selected suggestions
                             // so...
                             // update the commonL(ist)
                             if (is1stOcc) {
-                                commonL = swo.p;
+                                commonL = swo.p;    // init list (on 1st occurance)
                                 is1stOcc = false;
                             }
                             else {
-                                // list ot common products
+                                // caclulate list of common products
                                 // = intersection of (so-far) commonL and swo.p
-                                commonL = commonL.filter(value => swo.p.includes(value));
+                                // commonL = commonL.filter(value => swo.p.includes(value));
+                                commonL = commonL.filter(function(n) { return swo.p.indexOf(n) !== -1; });
                             }
                         }
                         else {  // if swo is not already selected
@@ -221,13 +346,12 @@ $(document).ready(function() {
                             possibleNext.push(swo);
                         }
                     });
-                    // console.log('commonL:', commonL)
 
                     possibleNext.forEach( poss => {     // for tthe possible next suggestions
                         // if word matches regex
                         // and list of word's products has commons with commonL
                         // then it is a valid next suggestion
-                        if (substrRegex.test(poss.kb)) {
+                        if (poss.kb.indexOf(kblast) !== -1 ) {
                             // check intersection of commonL and suggestion's product-lists
                             tempL = commonL.filter(value => poss.p.includes(value));
                             if (tempL.length) {
@@ -245,55 +369,97 @@ $(document).ready(function() {
             });
         }
 
+        // calculate unique products
+        let unique = proList.filter((item, i, ar) => ar.indexOf(item) === i);
+            // credit: https://stackoverflow.com/questions/11246758/
 
-        if ((qAr.length != 1) && (proList.length < 13)) {
-            // get unique product ids
-            let unique = proList.filter((item, i, ar) => ar.indexOf(item) === i);
-                // credit: https://stackoverflow.com/questions/11246758/
-            
+        // console.log(q, qAr, qAr.length, proList.length);
+
+        if (unique.length < (maxResults +1)) {
             results = [];
             unique.forEach( pr => {
-                everyProduct.forEach( pi => {
+                _products.forEach( pi => {
                     if (pi.id == pr)
                         results.push(pi);
                 })         
             });
         }
 
+        return results;
+   }
+
+    // request suggestions procedure
+    // args...
+    // qOrig: original query string
+    // list: artay structure to host results
+    // ---
+    var isuggest = function(qOrig, list) {
+        var results = suggestions_engine(qOrig);
+
+        // if no results...
+        // request again after removing last (key)word
+        if (results.length == 0) {
+            var qAr = qOrig.trim().split(' ');
+            qAr.pop();  // remove last word
+            results = suggestions_engine(qAr.join(' '));
+        }
+
         list(results);
     }
 
-	// setup suggestions search/input control
-	// ---
+    // setup suggestions search/input control
+    // ---
     const $tagsInput = $('#tagsInput')
     $tagsInput.typeahead(
         {
             hint: true,
             highlight: true,
-            minLength: 0
+            minLength: 1
         },
         {
-            limit: 12,
-            name: 'products',
+            limit: maxResults,
+            name: 'kwlinks',
             displayKey: 'w',
             source: isuggest,
-			templates: {
-				suggestion: function(data) {
-					console.log(data.w);
-					if (data.id)
-						return '<div>'+ data.w + '<span>' + data.id + '</span></div>';
-					return '<div>'+ data.w +'</div>';
-				}
-			}
+            templates: {
+                suggestion: function(data) {
+                    // console.log(data.w);
+                    if (data.id)
+                        return '<div>'+ data.w + '<span>' + data.id + '</span></div>';
+                    return '<div>'+ data.w +'<span>+</span></div>';
+                }
+            }
         }
     )
     .bind("typeahead:selected", function(obj, datum, name) {
-        $('.typeahead').typeahead('val','').trigger('blur');
-        $('.typeahead').typeahead('val', datum.w +' ')
+        // console.log(datum);
+        if (datum.hasOwnProperty('id')) {
+        // final product selected; do whatever ...
+        // ex. add to selection list
+        $('#selections').append('<div>'+ datum.w + '<span>' + datum.id + '</span></div>');
+        
+        // then reset search control
+        $('.typeahead').typeahead('val','').trigger('blur')
             .trigger("query");
-		// give some time to the engine to calculate results
-		// then fire focus again...
-        setTimeout(() => {  $('.typeahead').focus(); }, 100);  
+        setTimeout(() => { $('.typeahead').focus(); }, 100);
+
+        // finaly save tracking info;
+        $('#selections').append('<div class="-info-">'+ echo_tracking() +'</div>');
+        trackSearch.length = 0;       // ... and reset info to be ready for nextsearch
+        }
+        else {
+            $('.typeahead').typeahead('val','').trigger('blur');
+            $('.typeahead').typeahead('val', datum.w +' ')
+                .trigger("query");
+            // give some time to the engine to calculate results
+            // then fire focus again...
+            setTimeout(() => { $('.typeahead').focus(); }, 100);
+        }
+        trackSearch.push({
+            e: 'key',
+            v: 'Enter',
+            i: $('#tagsInput').val()
+        });
     })
     .bind("typeahead:cursorchange", function(obj, data) {
         // console.log(obj, data);
@@ -301,8 +467,27 @@ $(document).ready(function() {
         // console.log('triggered cursorchange /'+dt);
     });
 
+    // TRACK user search attempt ///////////////////////////////////////////////
+    $('.typeahead').on('keyup', function(e) {
+        trackSearch.push({
+            e: 'key',
+            v: e.key,
+            i: $('#tagsInput').val()
+        });
+    });
+
+});
+
+
+
+
+// also check (for suearching benchmarks)
+// ---
+// https://www.measurethat.net/Benchmarks/Show/13675/0/regextest-vs-stringincludes-vs-stringmatch
+// https://stackoverflow.com/questions/5296268/fastest-way-to-check-a-string-contain-another-substring-in-javascript
+// https://stackoverflow.com/questions/40387106/string-startwith-vs-regex
+// https://www.measurethat.net/Benchmarks/Show/4797/1/js-regex-vs-startswith-vs-indexof
 
 
-});       
-        </script>
-</html>
+    </script>
+</html>
\ No newline at end of file