Fix/324 apostrophe word count

src/model/writing-session-tracker.ts

@@ -30,20 +30,26 @@ const WORD_COUNT_REGEX = (() => {
       .source;
   return new RegExp(
     [
-      "(?:[0-9]+(?:(?:,|\\.)[0-9]+)*|[\\-" + spaceDelimitedChars + "])+",
+      "(?:[0-9]+(?:(?:,|\\.)[0-9]+)*|[\\-'" + spaceDelimitedChars + "])+",
       nonSpaceDelimitedWords,
     ].join("|"),
     "g"
   );
 })();
 
-function countWords(
+export function countWords(
   text: string,
   removeMarkdown = true,
   removeComments = true
 ): number {
   let textToCount = text;
 
+  if (removeComments) {
+    textToCount = textToCount
+      .replace(/<!--.*?-->/gs, "") // HTML comments
+      .replace(/%%.*?%%/gs, ""); // Markdown comments
+  }
+
   if (removeMarkdown) {
     textToCount = textToCount
       .replace(/`\$?=[^`]+`/g, "") // inline dataview
@@ -52,12 +58,6 @@ function countWords(
       .replace(/\*|_|\[\[|\]\]|\||==|~~|---|#|> |`/g, ""); // Markdown Syntax
   }
 
-  if (removeComments) {
-    textToCount = textToCount
-      .replace(/<!--.*?-->/gs, "") // HTML comments
-      .replace(/%%.*?%%/gs, ""); // Markdown comments
-  }
-
   return (textToCount.match(WORD_COUNT_REGEX) || []).length;
 }
 

test/__mocks__/obsidian.ts

@@ -0,0 +1,18 @@
+export const debounce = <T extends (...args: unknown[]) => unknown>(fn: T, _wait: number, _immediate?: boolean): T => {
+  return fn;
+};
+
+export abstract class TAbstractFile {
+  path: string = "";
+  vault: unknown = null;
+}
+
+export class TFile extends TAbstractFile {
+  stat = { ctime: 0, mtime: 0, size: 0 };
+  basename = "";
+  extension = "";
+  name = "";
+  parent = null;
+}
+
+export class Vault {}

test/model/writing-session-tracker.test.ts

@@ -0,0 +1,130 @@
+import { describe, expect, it } from "vitest";
+import { countWords } from "src/model/writing-session-tracker";
+
+describe("countWords", () => {
+  it("counts basic space-delimited words", () => {
+    expect(countWords("Hello world")).toBe(2);
+    expect(countWords("one two three four five")).toBe(5);
+    expect(countWords("a")).toBe(1);
+  });
+
+  it("handles multiple spaces", () => {
+    expect(countWords("Hello    world")).toBe(2);
+    expect(countWords("  leading")).toBe(1);
+    expect(countWords("trailing  ")).toBe(1);
+  });
+
+  it("handles newlines", () => {
+    expect(countWords("Hello\nworld")).toBe(2);
+    expect(countWords("line1\nline2\nline3")).toBe(3);
+  });
+
+  it("handles empty string", () => {
+    expect(countWords("")).toBe(0);
+  });
+
+  it("counts numbers as words", () => {
+    expect(countWords("12345")).toBe(1);
+    expect(countWords("hello 123 world")).toBe(3);
+  });
+
+  it("counts comma-separated numbers as one word", () => {
+    expect(countWords("1,000")).toBe(1);
+    expect(countWords("1,000,000")).toBe(1);
+  });
+
+  it("counts decimal numbers as one word", () => {
+    expect(countWords("1,000.50")).toBe(1);
+    expect(countWords("3.14")).toBe(1);
+  });
+
+  it("strips bold markers", () => {
+    expect(countWords("**bold**")).toBe(1);
+    expect(countWords("**bold** and **more**")).toBe(3);
+  });
+
+  it("strips italic markers", () => {
+    expect(countWords("*italic*")).toBe(1);
+    expect(countWords("_italic_")).toBe(1);
+  });
+
+  it("strips wiki links", () => {
+    expect(countWords("[[link]]")).toBe(1);
+    expect(countWords("[[File|Alias]]")).toBe(1);
+  });
+
+  it("strips external links", () => {
+    expect(countWords("[text](url)")).toBe(1);
+  });
+
+  it("strips heading markers", () => {
+    expect(countWords("# Header")).toBe(1);
+    expect(countWords("## Subheader")).toBe(1);
+  });
+
+  it("removes YAML frontmatter", () => {
+    expect(countWords("---\ntitle: x\n---\n\nBody text")).toBe(2);
+  });
+
+  it("strips HTML comments", () => {
+    expect(countWords("Hello <!-- comment --> world")).toBe(2);
+  });
+
+  it("strips markdown comments", () => {
+    expect(countWords("Hello %% comment %% world")).toBe(2);
+  });
+
+  it("counts hyphenated words as one word", () => {
+    expect(countWords("well-known")).toBe(1);
+    expect(countWords("up-to-date")).toBe(1);
+  });
+
+  it("counts CJK characters individually", () => {
+    expect(countWords("日本語")).toBe(3);
+    expect(countWords("東京")).toBe(2);
+  });
+
+  it("handles markdown code inlines", () => {
+    // NB: backticks are stripped by the markdown regex, so `code`
+    // becomes `code` and counts as a word. Existing behavior.
+    expect(countWords("text `code` text")).toBe(3);
+  });
+
+  it("strips inline dataview", () => {
+    expect(countWords("text `=this.field` text")).toBe(2);
+  });
+
+  it("handles real-world paragraph text", () => {
+    const text = "The quick brown fox jumps over the lazy dog.";
+    expect(countWords(text)).toBe(9);
+  });
+
+  describe("flag combinations with mixed content", () => {
+    const mixed = "Hello <!-- html --> **bold** %% md %% world";
+
+    it("strips comments and markdown when both flags active", () => {
+      expect(countWords(mixed, true, true), "comments removed first, then ** stripped — clean word count").toBe(3);
+    });
+
+    it("strips only markdown, preserves %% comments", () => {
+      expect(countWords(mixed, true, false), "md strips ** and > but %% is left intact so it acts as word-delimiter for md").toBe(6);
+    });
+
+    it("strips only comments, preserves markdown syntax", () => {
+      expect(countWords(mixed, false, true), "* is not in the word character class, so **bold** becomes bold").toBe(3);
+    });
+
+    it("preserves all syntax when processing disabled", () => {
+      expect(countWords(mixed, false, false), "raw text — all HTML, markdown, and %% tokens counted as-is").toBe(7);
+    });
+  });
+
+  it("counts contractions with apostrophes as a single word", () => {
+    expect.soft(countWords("Can't"), '"Can\'t" should be 1 word').toBe(1);
+    expect.soft(countWords("don't"), '"don\'t" should be 1 word').toBe(1);
+    expect.soft(countWords("it's"), '"it\'s" should be 1 word').toBe(1);
+    expect.soft(countWords("won't"), '"won\'t" should be 1 word').toBe(1);
+    expect.soft(countWords("o'clock"), '"o\'clock" should be 1 word').toBe(1);
+    expect.soft(countWords("won't don't can't"), 'phrase with 3 contractions should be 3 words').toBe(3);
+  });
+});

vitest.config.ts

@@ -1,6 +1,12 @@
 import { defineConfig } from "vitest/config";
+import path from "path";
 
 export default defineConfig({
+  resolve: {
+    alias: {
+      obsidian: path.resolve(__dirname, "test/__mocks__/obsidian.ts"),
+    },
+  },
   test: {
     coverage: {
       reporter: ["text", "html"],