word counter

sob.js

@@ -0,0 +1,70 @@
+const Sob = {
+    stringToWords(str) { return str.match(/\b(\w+)\b/g) },
+
+    stringToChars(str) { return Array.from(str) },
+
+    stringToUTF16(str) {
+        const utf16 = [];
+        for (let i = 0; i < str.length; i += 1) {
+            utf16.push(str.charCodeAt(i));
+        }
+        return utf16;
+    },
+
+    stringToUTF8(str) {
+        const utf8 = [];
+        for (let i = 0; i < str.length; i += 1) {
+            let code = str.charCodeAt(i);
+            if (code < 0x80) {
+                utf8.push(code);
+            } else if (code < 0x800) {
+                utf8.push(0xc0 | (code >> 6));
+                utf8.push(0x80 | (code & 0x3f));
+            } else if (code < 0xd800 || code >= 0xe000) {
+                utf8.push(0xe0 | (code >> 12));
+                utf8.push(0x80 | ((code >> 6) & 0x3f));
+                utf8.push(0x80 | (code & 0x3f));
+            } else {
+                i += 1;
+                code = 0x10000 + (((code & 0x3ff) << 10) | (str.charCodeAt(i) & 0x3ff));
+                utf8.push(0xf0 | (code >>18));
+                utf8.push(0x80 | ((code>>12) & 0x3f));
+                utf8.push(0x80 | ((code>>6) & 0x3f));
+                utf8.push(0x80 | (code & 0x3f));
+            }
+        }
+        return utf8;
+    },
+
+    sanitiseHTML(html) {
+        return html
+                .replace("&", "&amp;")
+                .replace("<", "&lt;")
+                .replace(">", "&gt;")
+                .replace("\"", "&quot;")
+                .replace("'", "&#039;");
+    },
+
+    sanitiseEscapes(str) {
+        let chars = {
+            "\n" : "\\n", "\r" : "\\r",
+            "\f" : "\\f", "\t" : "\\t",
+            "\b" : "\\b",
+            "\"" : "\\\"", "\'" : "\\'",
+        };
+        return Sob.stringToChars(str).map(x => chars[x] ?? x).join("");
+    },
+
+    showList(xs, each=undefined) {
+        let first = true;
+        let sb = "";
+        for (const x of xs) {
+            if (!first) {
+                sb += ", ";
+            }
+            first = false;
+            sb += each == undefined ? x : each(x);
+        }
+        return sb;
+    },
+};

word-count.html

@@ -0,0 +1,6 @@
+<script src="./sob.js"></script>
+<script src="./char-count.js"></script>
+<textarea id="src" rows="10" style="width : 100%">paste text here</textarea>
+<hr>
+<button onclick="getCounts()">word count</button>
+<div id="dest"></div>

word-count.js

@@ -0,0 +1,31 @@
+function getTextContent() {
+    return document.getElementById("src").value;
+}
+
+function setResultHTML(html) {
+    document.getElementById("dest").innerHTML = html;
+}
+
+function getCounts() {
+    const text = getTextContent();
+    const words = Sob.stringToWords(text);
+    const chars = Sob.stringToChars(text);
+    const bytes = Sob.stringToUTF8(text);
+    let sb = "";
+    sb += "<p>"
+    sb += `<b>#words</b>: <code>${words.length}</code><br>`
+    sb += `<b>#characters</b>: <code>${chars.length}</code><br>`
+    sb += `<b>#bytes</b>: <code>${bytes.length}</code>`
+    sb += "</p>"
+    sb += "<details><summary>expand word list</summary><pre><code style=\"text-wrap : wrap\">"
+    sb += "[" + Sob.showList(words, x => `"${Sob.sanitiseEscapes(x)}"`) + "]";
+    sb += "</code></pre></details>"
+    sb += "<details><summary>expand character list</summary><pre><code style=\"text-wrap : wrap\">"
+    sb += "[" + Sob.showList(chars, x => `'${Sob.sanitiseEscapes(x)}'`) + "]";
+    sb += "</code></pre></details>"
+    sb += "<details><summary>expand byte list</summary><pre><code style=\"text-wrap : wrap\">"
+    sb += "[" + Sob.showList(bytes) + "]";
+    sb += "</code></pre></details>"
+    setResultHTML(sb);
+    console.log("got counts");
+}