savvides · savvides · Aug 22, 2026
diff --git a/extension/content/extractor-core.cjs b/extension/content/extractor-core.cjs
diff --git a/extension/content/extractor.js b/extension/content/extractor.js
@@ -1,84 +1,16 @@
 /**
- * Injected content script to extract course content from Canvas, Google Docs, and web pages.
+ * Injected content script listener for extracting course content.
+ * Relies on shared/extractor-core.js being loaded in the same context.
  */
-function detectPageType(url, document) {
-  if (url.includes('instructure.com') || url.includes('/courses/')) {
-    if (document.querySelector('#assignment_show')) return 'Canvas Assignment';
-    if (document.querySelector('#syllabusContainer') || url.includes('/assignments/syllabus')) return 'Canvas Syllabus';
-    if (document.querySelector('#modules') || url.includes('/modules')) return 'Canvas Modules';
-    if (document.querySelector('#rubrics')) return 'Canvas Rubric';
-    return 'Canvas LMS Page';
-  }
-  if (url.includes('docs.google.com/document')) return 'Google Doc Syllabus';
-  return 'Web Syllabus / Course Page';
-}
-
-function extractContentFromDOM(document, url = (typeof window !== 'undefined' && window.location ? window.location.href : '')) {
-  const pageType = detectPageType(url, document);
-  let title = (document && document.title) || 'Course Document';
-  let content = '';
-
-  if (pageType.startsWith('Canvas')) {
-    const heading = document.querySelector('#assignment_show .title, .page-title, h1');
-    if (heading) title = (heading.innerText || heading.textContent || '').trim();
-
-    const mainBody = document.querySelector('#assignment_show .description.user_content, .show-content.user_content, #syllabusContainer, .module-item-title');
-    content = mainBody ? (mainBody.innerText || mainBody.textContent || '').trim() : (document.body ? (document.body.innerText || document.body.textContent || '').trim() : '');
-  } else if (pageType === 'Google Doc Syllabus') {
-    const kixApp = document.querySelector('.kix-appview-editor');
-    content = kixApp ? (kixApp.innerText || kixApp.textContent || '').trim() : (document.body ? (document.body.innerText || document.body.textContent || '').trim() : '');
-  } else {
-    // General web page extraction
-    const article = document.querySelector('main, article, [role="main"]');
-    content = article ? (article.innerText || article.textContent || '').trim() : (document.body ? (document.body.innerText || document.body.textContent || '').trim() : '');
-  }
-
-  return {
-    url,
-    title,
-    pageType,
-    content: content.slice(0, 15000),
-    wordCount: content.split(/\s+/).filter(Boolean).length
-  };
-}
-
-function extractPageContent() {
-  const url = typeof window !== 'undefined' && window.location ? window.location.href : '';
-  return extractContentFromDOM(document, url);
-}
-
-function detectCourseContext(url, docTitle) {
-  if (!url) return { isCourseRoot: false, courseId: null, origin: null };
-  try {
-    const parsedUrl = new URL(url);
-    const match = parsedUrl.pathname.match(/\/courses\/(\d+)(?:\/(?:modules)?)?\/?$/);
-    const anyCourseMatch = parsedUrl.pathname.match(/\/courses\/(\d+)/);
-    return {
-      isCourseRoot: !!match,
-      courseId: anyCourseMatch ? anyCourseMatch[1] : null,
-      origin: parsedUrl.origin
-    };
-  } catch (e) {
-    return { isCourseRoot: false, courseId: null, origin: null };
-  }
-}
 
 // Listen for messages from Side Panel
 if (typeof chrome !== 'undefined' && chrome.runtime && chrome.runtime.onMessage) {
   chrome.runtime.onMessage.addListener((request, sender, sendResponse) => {
     if (request.action === 'EXTRACT_CONTENT') {
-      const data = extractPageContent();
+      // extractPageContent is provided by shared/extractor-core.js
+      const data = typeof extractPageContent !== 'undefined' ? extractPageContent() : null;
       sendResponse(data);
     }
     return true;
   });
 }
-
-if (typeof module !== 'undefined' && module.exports) {
-  module.exports = {
-    detectPageType,
-    extractContentFromDOM,
-    extractPageContent,
-    detectCourseContext
-  };
-}
diff --git a/extension/manifest.json b/extension/manifest.json
@@ -26,7 +26,7 @@
         "*://docs.google.com/document/*",
         "<all_urls>"
       ],
-      "js": ["content/extractor.js"],
+      "js": ["shared/extractor-core.js", "content/extractor.js"],
       "run_at": "document_idle"
     }
   ],

diff --git a/extension/content/extractor-core.js → extension/shared/extractor-core.js b/extension/content/extractor-core.js → extension/shared/extractor-core.js
@@ -1,4 +1,7 @@
-export function detectPageType(url, document) {
+/**
+ * Core course content extraction logic for Canvas, Google Docs, and web pages.
+ */
+function detectPageType(url, document) {
   if (url.includes('instructure.com') || url.includes('/courses/')) {
     if (document.querySelector('#assignment_show')) return 'Canvas Assignment';
     if (document.querySelector('#syllabusContainer') || url.includes('/assignments/syllabus')) return 'Canvas Syllabus';
@@ -10,7 +13,7 @@ export function detectPageType(url, document) {
   return 'Web Syllabus / Course Page';
 }
 
-export function extractContentFromDOM(document, url = (typeof window !== 'undefined' && window.location ? window.location.href : '')) {
+function extractContentFromDOM(document, url = (typeof window !== 'undefined' && window.location ? window.location.href : '')) {
   const pageType = detectPageType(url, document);
   let title = (document && document.title) || 'Course Document';
   let content = '';
@@ -39,12 +42,12 @@ export function extractContentFromDOM(document, url = (typeof window !== 'undefi
   };
 }
 
-export function extractPageContent() {
+function extractPageContent() {
   const url = typeof window !== 'undefined' && window.location ? window.location.href : '';
   return extractContentFromDOM(document, url);
 }
 
-export function detectCourseContext(url, docTitle) {
+function detectCourseContext(url, docTitle) {
   if (!url) return { isCourseRoot: false, courseId: null, origin: null };
   try {
     const parsedUrl = new URL(url);
@@ -60,3 +63,11 @@ export function detectCourseContext(url, docTitle) {
   }
 }
 
+if (typeof module !== 'undefined' && module.exports) {
+  module.exports = {
+    detectPageType,
+    extractContentFromDOM,
+    extractPageContent,
+    detectCourseContext
+  };
+}
diff --git a/extension/sidepanel/index.html b/extension/sidepanel/index.html
@@ -110,7 +110,7 @@ <h4>Privacy & FERPA Commitment</h4>
     </section>
   </main>
 
-
+  <script src="../shared/extractor-core.js"></script>
   <script type="module" src="sidepanel.js"></script>
 </body>
 </html>
diff --git a/extension/sidepanel/sidepanel.js b/extension/sidepanel/sidepanel.js
@@ -1,8 +1,9 @@
 import { getSettings, saveSettings, getDossier, addToDossier, removeFromDossier, clearDossier } from '../shared/storage.js';
 import { renderAuditHTML, renderDossierListHTML } from './renderer-helper.js';
-import { detectCourseContext } from '../content/extractor-core.js';
 import { compileSingleAuditToMarkdown, compileDossierToMarkdown } from '../shared/dossier-compiler.js';
 
+// detectCourseContext is now provided globally via ../shared/extractor-core.js
+
 let activePayload = null;
 let activeCourseContext = null;
 let activeAuditResult = null;

diff --git a/test/test-doctor.sh b/test/test-doctor.sh
@@ -40,7 +40,7 @@ DOCTOR_CMD="$MOCK_IDSTACK_DIR/bin/idstack-doctor"
 # mock we drop in. `claude` is absent unless a test installs it.
 MOCK_BIN="$TEST_ROOT/bin"
 mkdir -p "$MOCK_BIN"
-for _c in bash sh env command dirname echo grep head python3 readlink sed tr awk cat ls rm mkdir ln chmod mktemp printf sort wc basename find touch cp; do
+for _c in bash sh env command dirname echo grep head python3 pyenv cut readlink sed tr awk cat ls rm mkdir ln chmod mktemp printf sort wc basename find touch cp; do
   _p=$(command -v "$_c" 2>/dev/null) && ln -sf "$_p" "$MOCK_BIN/$_c"
 done
 export PATH="$MOCK_BIN"

diff --git a/test/test-extractor.js b/test/test-extractor.js
@@ -136,7 +136,7 @@ try {
   JSDOM = SimpleJSDOM;
 }
 
-const { detectPageType, extractContentFromDOM, extractPageContent } = require('../extension/content/extractor-core.cjs');
+const { detectPageType, extractContentFromDOM, extractPageContent, detectCourseContext } = require('../extension/shared/extractor-core.js');
 
 // Test 1: Canvas Assignment fixture
 const canvasAssignmentHTML = `
@@ -297,6 +297,7 @@ global.chrome = {
 };
 global.window = domAssignment.window;
 global.document = domAssignment.window.document;
+global.extractPageContent = extractPageContent;
 
 delete require.cache[require.resolve('../extension/content/extractor.js')];
 require('../extension/content/extractor.js');
@@ -314,9 +315,9 @@ assert.strictEqual(responseData.title, 'Enzymes Lab Analysis');
 delete global.chrome;
 delete global.window;
 delete global.document;
+delete global.extractPageContent;
 
 // Test 11: Canvas Course Root Detection
-const { detectCourseContext } = require('../extension/content/extractor-core.cjs');
 const rootCtx = detectCourseContext('https://canvas.instructure.com/courses/987654', 'Biology 101');
 assert.strictEqual(rootCtx.isCourseRoot, true);
 assert.strictEqual(rootCtx.courseId, '987654');