Better regex parsing in annotation range

.changeset/loud-moons-grow.md

@@ -0,0 +1,5 @@
+---
+"codehike": patch
+---
+
+Better regex parsing in annotation range

packages/codehike/src/code/extract-annotations.test.ts

@@ -0,0 +1,60 @@
+import { expect, test } from "vitest"
+import { splitAnnotationsAndCode } from "./extract-annotations.js"
+
+async function t(comment: string) {
+  const code = `// ${comment} \nvar xyz = "https://example.com"`
+  const { annotations } = await splitAnnotationsAndCode(code, "javascript", "!")
+  return annotations[0]
+}
+
+test("extracts basic annotation name", async () => {
+  const annotation = await t("!foo bar")
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with parentheses range", async () => {
+  const annotation = await t("!foo(1) bar")
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with brackets range", async () => {
+  const annotation = await t("!foo[1] bar")
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with simple regex", async () => {
+  const annotation = await t("!foo[/x/] bar")
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with regex flags", async () => {
+  const annotation = await t("!foo[/x/gmi] bar")
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with regex containing brackets", async () => {
+  const annotation = await t(`!foo[/xyz[a-z]*/g] bar`)
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with regex containing parentheses", async () => {
+  const annotation = await t(`!foo(/(xyz[w]*)/g) bar`)
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with regex containing nested parentheses", async () => {
+  const annotation = await t(`!foo(/((xyz)[w]*)/g) bar`)
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with regex containing escaped slashes", async () => {
+  const annotation = await t(`!foo[/https?:\\/\\//g] bar`)
+  expect(annotation.name).toEqual("foo")
+})
+
+test("extracts name with complex regex pattern", async () => {
+  const code = `// !tooltip[/#\\[program\\]/] example \n#[program]`
+  const { annotations } = await splitAnnotationsAndCode(code, "rust", "!")
+  const annotation = annotations[0]
+  expect(annotation.name).toEqual("tooltip")
+})

packages/codehike/src/code/extract-annotations.tsx

@@ -26,12 +26,21 @@ async function extractCommentAnnotations(
   annotationPrefix = "!",
 ) {
   const extractor = (comment: string) => {
-    // const regex = /\s*(!?[\w-]+)?(\([^\)]*\)|\[[^\]]*\])?(.*)$/
+    const body = "(?:\\\\.|[^\\\\/])+"
+    const nestedBracketRegex = new RegExp(
+      `\\s*(${annotationPrefix}?[\\w-]+)?(\\[\\/${body}\\/[a-zA-Z]*\\])(.*)$`,
+    )
+    const nestedParenRegex = new RegExp(
+      `\\s*(${annotationPrefix}?[\\w-]+)?(\\(\\/${body}\\/[a-zA-Z]*\\))(.*)$`,
+    )
     const regex = new RegExp(
       `\\s*(${annotationPrefix}?[\\w-]+)?(\\([^\\)]*\\)|\\[[^\\]]*\\])?(.*)$`,
     )
 
-    const match = comment.match(regex)
+    const match =
+      comment.match(nestedBracketRegex) ||
+      comment.match(nestedParenRegex) ||
+      comment.match(regex)
     if (!match) {
       return null
     }