Fix curly quotes in remark-capitalize-titles (#34)

* fix curly quotes and hypenated compounds

* fix lint issues

Author: essential-randomness

.changeset/curly-quote-capitalization.md

@@ -0,0 +1,9 @@
+---
+"@fujocoded/remark-capitalize-titles": patch
+---
+
+Capitalize words correctly when a heading contains curly quotes (as produced by
+`remark-smartypants`). Previously, the upstream `title` library only recognized
+straight quotes as punctuation, so a word following a curly `“` or `’` would
+stay lowercase. The plugin now converts curly quotes to straight quotes before
+title-casing, then restores the original curly characters in the output.

.changeset/hyphenated-compound-tails.md

@@ -0,0 +1,12 @@
+---
+"@fujocoded/remark-capitalize-titles": minor
+---
+
+Lowercase the second-and-later segments of a hyphenated compound during title
+casing, so output follows AP-style ("Three-way Merges", "Pre-commit Hooks",
+"Up-to-date") instead of capitalizing every segment ("Three-Way", "Pre-Commit",
+"Up-To-Date"). A segment is kept capitalized when either the full compound or
+the individual segment is listed in `special`.
+
+This is a breaking change for callers that expected every segment of a
+hyphenated word to be capitalized.

package.json

@@ -15,7 +15,9 @@
     "lint:fix": "oxlint --type-aware --fix",
     "lint:ci": "oxlint --type-aware --deny-warnings",
     "format": "oxfmt && prettier --write \"**/*.astro\"",
-    "format:check": "oxfmt --check && prettier --check \"**/*.astro\""
+    "format:check": "oxfmt --check && prettier --check \"**/*.astro\"",
+    "check": "npm run format:check && npm run lint:ci && npm run sherif && npm run typecheck && npm run build",
+    "fix": "npm run format && npm run lint:fix"
   },
   "devDependencies": {
     "@changesets/cli": "^2.29.8",

remark-capitalize-titles/index.ts

@@ -7,13 +7,59 @@ import type { MdxJsxFlowElement } from "mdast-util-mdx-jsx";
 
 import { DEFAULT_CAPITALIZATIONS as DEFAULT_CAPITALIZATIONS_ } from "./capitalizations.ts";
 
+// Astro's MDX integration runs remark-smartypants before user plugins, so
+// straight quotes arrive here as curly. The `title` library's regex only lists
+// straight quotes as punctuation, so curly quotes would otherwise prevent the
+// next word from being capitalized.
+const CURLY_TO_STRAIGHT: Record<string, string> = {
+  "\u201C": '"',
+  "\u201D": '"',
+  "\u2018": "'",
+  "\u2019": "'",
+};
+const CURLY_QUOTE_REGEX = /[\u201C\u201D\u2018\u2019]/g;
+
+// Matches a hyphenated compound like "Three-Way" or "Up-To-Date" so the
+// second-and-later segments can be lowercased (AP-style: "Three-way").
+const HYPHENATED_COMPOUND_REGEX = /[A-Za-z][A-Za-z']*(?:-[A-Za-z][A-Za-z']*)+/g;
+
+const lowercaseHyphenatedTails = (text: string, special: string[]) =>
+  text.replace(HYPHENATED_COMPOUND_REGEX, (match) => {
+    if (special.includes(match)) return match;
+    const parts = match.split("-");
+    return parts
+      .map((part, index) => {
+        if (index === 0) return part;
+        if (special.includes(part)) return part;
+        return part.charAt(0).toLowerCase() + part.slice(1);
+      })
+      .join("-");
+  });
+
 const title = (...params: Parameters<typeof libraryTitle>) => {
   const [text, options] = params;
-  const textChunks = text.split(")");
-  const intermediateTitle = textChunks
-    .map((title) => libraryTitle(title, options))
+  const curlyPositions: Array<[number, string]> = [];
+  const normalized = text.replace(
+    CURLY_QUOTE_REGEX,
+    (match, offset: number) => {
+      curlyPositions.push([offset, match]);
+      return CURLY_TO_STRAIGHT[match] ?? match;
+    },
+  );
+  const textChunks = normalized.split(")");
+  const titleCased = textChunks
+    .map((chunk) => libraryTitle(chunk, options))
     .join(")");
-  return intermediateTitle;
+  const intermediateTitle = lowercaseHyphenatedTails(
+    titleCased,
+    options?.special ?? [],
+  );
+  if (curlyPositions.length === 0) return intermediateTitle;
+  const chars = intermediateTitle.split("");
+  for (const [offset, original] of curlyPositions) {
+    chars[offset] = original;
+  }
+  return chars.join("");
 };
 
 type PluginArgs = { special: string[]; componentNames: string[] };

remark-capitalize-titles/package.json

@@ -38,6 +38,7 @@
   "scripts": {
     "build": "tsup index.ts --format cjs,esm --dts --clean",
     "validate": " npx publint",
+    "test": "vitest",
     "typecheck": "tsc --noEmit"
   },
   "dependencies": {
@@ -51,6 +52,7 @@
     "remark": "^15.0.1",
     "tsup": "^8.1.0",
     "typescript": "^5.5.2",
-    "unified": "^11.0.4"
+    "unified": "^11.0.4",
+    "vitest": "^3.0.5"
   }
 }

remark-capitalize-titles/tests/index.test.ts

@@ -0,0 +1,151 @@
+import { describe, expect, test } from "vitest";
+import { remark } from "remark";
+import type { Compatible } from "vfile";
+import remarkCapitalizeTitles from "../index.ts";
+
+const processMarkdown = async (value: Compatible) => {
+  const file = await remark().use(remarkCapitalizeTitles).process(value);
+  return file.toString().slice(0, -1);
+};
+
+describe("Handles the basics", () => {
+  test("title-cases a simple heading", async () => {
+    expect(
+      await processMarkdown("# cloning: not just for mad scientists"),
+    ).toBe("# Cloning: Not Just for Mad Scientists");
+  });
+
+  test("leaves non-heading text untouched", async () => {
+    expect(
+      await processMarkdown(
+        "git mindwiped: fully discarding changes\n\n# git mindwiped: fully discarding changes",
+      ),
+    ).toBe(
+      "git mindwiped: fully discarding changes\n\n# Git Mindwiped: Fully Discarding Changes",
+    );
+  });
+});
+
+describe("Preserves special cases", () => {
+  test("preserves GitHub, FujoCoded, LLC, and NPM", async () => {
+    expect(
+      await processMarkdown(
+        "## merging with github's interface: pull requests",
+      ),
+    ).toBe("## Merging with GitHub's Interface: Pull Requests");
+    expect(await processMarkdown("### an intro to fujocoded llc")).toBe(
+      "### An Intro to FujoCoded LLC",
+    );
+    expect(await processMarkdown("## next up: building with npm")).toBe(
+      "## Next Up: Building with NPM",
+    );
+  });
+
+  test("preserves SHA", async () => {
+    expect(await processMarkdown("#### sha: your commit's unique name")).toBe(
+      "#### SHA: Your Commit's Unique Name",
+    );
+  });
+
+  test("preserves TL;DR", async () => {
+    expect(await processMarkdown("# tl;dr: why this matters")).toBe(
+      "# TL;DR: Why This Matters",
+    );
+  });
+});
+
+describe("Respects tricky punctuations", () => {
+  test("handles curly quotes (as produced by smartypants)", async () => {
+    expect(
+      await processMarkdown(
+        "#### “answer me, darling~”: git & github’s connection check!",
+      ),
+    ).toBe("#### “Answer Me, Darling~”: Git & GitHub’s Connection Check!");
+  });
+
+  test("handles apostrophes inside a word", async () => {
+    expect(
+      await processMarkdown("## git'ing good: more commit scenarios"),
+    ).toBe("## Git'ing Good: More Commit Scenarios");
+  });
+
+  test("handles a trailing question mark with an inner apostrophe", async () => {
+    expect(await processMarkdown("## i'm ready to practice, now what?")).toBe(
+      "## I'm Ready to Practice, Now What?",
+    );
+  });
+
+  test("handles repeated punctuation (???)", async () => {
+    expect(await processMarkdown("### step ???: git advanced")).toBe(
+      "### Step ???: Git Advanced",
+    );
+  });
+
+  test("handles a leading ellipsis", async () => {
+    expect(await processMarkdown("### ...and more!")).toBe("### ...and More!");
+  });
+
+  test("handles parenthesized possessives", async () => {
+    expect(
+      await processMarkdown("### traveling through (your code's) history"),
+    ).toBe("### Traveling Through (Your Code's) History");
+  });
+});
+
+describe("Handles inline code spans", () => {
+  test("handles an inline code span inside a heading", async () => {
+    expect(
+      await processMarkdown(
+        "### the flavors of `git reset`: soft, hard, or mixed",
+      ),
+    ).toBe("### The Flavors of `git reset`: Soft, Hard, or Mixed");
+  });
+
+  test("handles multiple inline code spans with separators", async () => {
+    expect(await processMarkdown("## git & github's `push`/`pull` dance")).toBe(
+      "## Git & GitHub's `push`/`pull` Dance",
+    );
+  });
+
+  test("handles punctuation immediately following an inline code span", async () => {
+    expect(
+      await processMarkdown("## multiverse collapse: prepare to `merge`!"),
+    ).toBe("## Multiverse Collapse: Prepare to `merge`!");
+  });
+
+  test("handles a comma-separated list of inline code spans with a hyphenated term", async () => {
+    expect(
+      await processMarkdown(
+        "#### the jokes write themselves: `ours`, `theirs`, and three-way merges",
+      ),
+    ).toBe(
+      "#### The Jokes Write Themselves: `ours`, `theirs`, and Three-way Merges",
+    );
+  });
+});
+
+describe("Handles hyphenated compound words", () => {
+  test("keeps the second word lowercase in a hyphenated name with a possessive", async () => {
+    expect(
+      await processMarkdown("### our toy project: boba-tan's sexyman shrine"),
+    ).toBe("### Our Toy Project: Boba-tan's Sexyman Shrine");
+  });
+
+  test("handles hyphenated words alongside a slash separator", async () => {
+    expect(await processMarkdown("## push/pull: git's memory-sync dance")).toBe(
+      "## Push/pull: Git's Memory-sync Dance",
+    );
+  });
+
+  test("handles common hyphenated prefixes", async () => {
+    expect(
+      await processMarkdown("### pre-commit hooks for post-merge cleanups"),
+    ).toBe("### Pre-commit Hooks for Post-merge Cleanups");
+  });
+
+  test("handles hyphenated phrases with small words inside", async () => {
+    expect(await processMarkdown("## up-to-date and ready-to-merge")).toBe(
+      "## Up-to-date and Ready-to-merge",
+    );
+  });
+});