diff --git a/browser-extension/background.js b/browser-extension/background.js index 31120cb..fd21c04 100644 --- a/browser-extension/background.js +++ b/browser-extension/background.js @@ -26,6 +26,12 @@ chrome.runtime.onInstalled.addListener(() => { title: 'Clip selection to rNotes', contexts: ['selection'], }); + + chrome.contextMenus.create({ + id: 'unlock-article', + title: 'Unlock & Clip article to rNotes', + contexts: ['page', 'link'], + }); }); // --- Helpers --- @@ -132,6 +138,31 @@ async function uploadImage(imageUrl) { return response.json(); } +async function unlockArticle(url) { + const token = await getToken(); + if (!token) { + showNotification('rNotes Error', 'Not signed in. Open extension settings to sign in.'); + return null; + } + + const settings = await getSettings(); + const response = await fetch(`${settings.host}/api/articles/unlock`, { + method: 'POST', + headers: { + 'Content-Type': 'application/json', + 'Authorization': `Bearer ${token}`, + }, + body: JSON.stringify({ url }), + }); + + if (!response.ok) { + const text = await response.text(); + throw new Error(`Unlock failed: ${response.status} ${text}`); + } + + return response.json(); +} + // --- Context Menu Handler --- chrome.contextMenus.onClicked.addListener(async (info, tab) => { @@ -197,6 +228,28 @@ chrome.contextMenus.onClicked.addListener(async (info, tab) => { break; } + case 'unlock-article': { + const targetUrl = info.linkUrl || tab.url; + showNotification('Unlocking Article', `Finding readable version of ${new URL(targetUrl).hostname}...`); + + const result = await unlockArticle(targetUrl); + if (result && result.success && result.archiveUrl) { + // Create a CLIP note with the archive URL + await createNote({ + title: tab.title || 'Unlocked Article', + content: `

Unlocked via ${result.strategy}

Original: ${targetUrl}

Archive: ${result.archiveUrl}

`, + type: 'CLIP', + url: targetUrl, + }); + showNotification('Article Unlocked', `Readable version found via ${result.strategy}`); + // Open the unlocked article in a new tab + chrome.tabs.create({ url: result.archiveUrl }); + } else { + showNotification('Unlock Failed', result?.error || 'No archived version found'); + } + break; + } + case 'clip-selection': { // Get selection HTML let content = ''; diff --git a/browser-extension/popup.html b/browser-extension/popup.html index d0db5ac..69b33a8 100644 --- a/browser-extension/popup.html +++ b/browser-extension/popup.html @@ -133,6 +133,14 @@ color: #e5e5e5; border: 1px solid #404040; } + .btn-unlock { + background: #172554; + color: #93c5fd; + border: 1px solid #1e40af; + } + .btn-unlock svg { + flex-shrink: 0; + } .status { margin: 0 14px 10px; @@ -212,6 +220,16 @@ +
+ +
+
diff --git a/browser-extension/popup.js b/browser-extension/popup.js index c0fa35d..b73d7eb 100644 --- a/browser-extension/popup.js +++ b/browser-extension/popup.js @@ -152,6 +152,7 @@ async function init() { // Enable buttons document.getElementById('clipPageBtn').disabled = false; + document.getElementById('unlockBtn').disabled = false; // Load notebooks await populateNotebooks(); @@ -255,6 +256,49 @@ document.getElementById('clipSelectionBtn').addEventListener('click', async () = } }); +document.getElementById('unlockBtn').addEventListener('click', async () => { + const btn = document.getElementById('unlockBtn'); + btn.disabled = true; + showStatus('Unlocking article...', 'loading'); + + try { + const token = await getToken(); + const settings = await getSettings(); + + const response = await fetch(`${settings.host}/api/articles/unlock`, { + method: 'POST', + headers: { + 'Content-Type': 'application/json', + 'Authorization': `Bearer ${token}`, + }, + body: JSON.stringify({ url: currentTab.url }), + }); + + const result = await response.json(); + + if (result.success && result.archiveUrl) { + // Also save as a note + await createNote({ + title: currentTab.title || 'Unlocked Article', + content: `

Unlocked via ${result.strategy}

Original: ${currentTab.url}

Archive: ${result.archiveUrl}

`, + type: 'CLIP', + url: currentTab.url, + }); + + showStatus(`Unlocked via ${result.strategy}! Opening...`, 'success'); + + // Open archive in new tab + chrome.tabs.create({ url: result.archiveUrl }); + } else { + showStatus(result.error || 'No archived version found', 'error'); + } + } catch (err) { + showStatus(`Error: ${err.message}`, 'error'); + } finally { + btn.disabled = false; + } +}); + document.getElementById('optionsLink').addEventListener('click', (e) => { e.preventDefault(); chrome.runtime.openOptionsPage(); diff --git a/prisma/schema.prisma b/prisma/schema.prisma index 4b9786e..ee83cbb 100644 --- a/prisma/schema.prisma +++ b/prisma/schema.prisma @@ -75,6 +75,7 @@ model Note { contentPlain String? @db.Text type NoteType @default(NOTE) url String? + archiveUrl String? language String? mimeType String? fileUrl String? diff --git a/src/app/api/articles/unlock/route.ts b/src/app/api/articles/unlock/route.ts new file mode 100644 index 0000000..289d143 --- /dev/null +++ b/src/app/api/articles/unlock/route.ts @@ -0,0 +1,61 @@ +import { NextRequest, NextResponse } from 'next/server'; +import { prisma } from '@/lib/prisma'; +import { requireAuth, isAuthed } from '@/lib/auth'; +import { unlockArticle } from '@/lib/article-unlock'; + +/** + * POST /api/articles/unlock + * + * Attempts to find an archived/readable version of a paywalled article. + * + * Body: { url: string, noteId?: string } + * - url: The article URL to unlock + * - noteId: (optional) If provided, updates the note's archiveUrl on success + * + * Returns: { success, strategy, archiveUrl, error? } + */ +export async function POST(request: NextRequest) { + try { + const auth = await requireAuth(request); + if (!isAuthed(auth)) return auth; + + const body = await request.json(); + const { url, noteId } = body; + + if (!url || typeof url !== 'string') { + return NextResponse.json({ error: 'URL is required' }, { status: 400 }); + } + + // Validate URL format + try { + new URL(url); + } catch { + return NextResponse.json({ error: 'Invalid URL format' }, { status: 400 }); + } + + const result = await unlockArticle(url); + + // If successful and noteId provided, update the note's archiveUrl + if (result.success && result.archiveUrl && noteId) { + const existing = await prisma.note.findUnique({ + where: { id: noteId }, + select: { authorId: true }, + }); + + if (existing && (!existing.authorId || existing.authorId === auth.user.id)) { + await prisma.note.update({ + where: { id: noteId }, + data: { archiveUrl: result.archiveUrl }, + }); + } + } + + return NextResponse.json(result); + } catch (error) { + console.error('Article unlock error:', error); + return NextResponse.json( + { success: false, strategy: 'none', error: 'Internal server error' }, + { status: 500 } + ); + } +} diff --git a/src/app/api/notes/[id]/route.ts b/src/app/api/notes/[id]/route.ts index 9c9617d..80a3b66 100644 --- a/src/app/api/notes/[id]/route.ts +++ b/src/app/api/notes/[id]/route.ts @@ -50,7 +50,7 @@ export async function PUT( } const body = await request.json(); - const { title, content, type, url, language, isPinned, notebookId, tags } = body; + const { title, content, type, url, archiveUrl, language, isPinned, notebookId, tags } = body; const data: Record = {}; if (title !== undefined) data.title = title.trim(); @@ -60,6 +60,7 @@ export async function PUT( } if (type !== undefined) data.type = type; if (url !== undefined) data.url = url || null; + if (archiveUrl !== undefined) data.archiveUrl = archiveUrl || null; if (language !== undefined) data.language = language || null; if (isPinned !== undefined) data.isPinned = isPinned; if (notebookId !== undefined) data.notebookId = notebookId || null; diff --git a/src/app/api/notes/route.ts b/src/app/api/notes/route.ts index c7f8f7c..0a6f78d 100644 --- a/src/app/api/notes/route.ts +++ b/src/app/api/notes/route.ts @@ -43,7 +43,7 @@ export async function POST(request: NextRequest) { if (!isAuthed(auth)) return auth; const { user } = auth; const body = await request.json(); - const { title, content, type, notebookId, url, language, tags, fileUrl, mimeType, fileSize, duration } = body; + const { title, content, type, notebookId, url, archiveUrl, language, tags, fileUrl, mimeType, fileSize, duration } = body; if (!title?.trim()) { return NextResponse.json({ error: 'Title is required' }, { status: 400 }); @@ -75,6 +75,7 @@ export async function POST(request: NextRequest) { notebookId: notebookId || null, authorId: user.id, url: url || null, + archiveUrl: archiveUrl || null, language: language || null, fileUrl: fileUrl || null, mimeType: mimeType || null, diff --git a/src/app/notes/[id]/page.tsx b/src/app/notes/[id]/page.tsx index 9cf6de4..24097f2 100644 --- a/src/app/notes/[id]/page.tsx +++ b/src/app/notes/[id]/page.tsx @@ -25,6 +25,7 @@ interface NoteData { contentPlain: string | null; type: string; url: string | null; + archiveUrl: string | null; language: string | null; fileUrl: string | null; mimeType: string | null; @@ -49,6 +50,8 @@ export default function NoteDetailPage() { const [saving, setSaving] = useState(false); const [diarizing, setDiarizing] = useState(false); const [speakers, setSpeakers] = useState<{ speaker: string; start: number; end: number }[] | null>(null); + const [unlocking, setUnlocking] = useState(false); + const [unlockError, setUnlockError] = useState(null); useEffect(() => { fetch(`/api/notes/${params.id}`) @@ -135,6 +138,30 @@ export default function NoteDetailPage() { } }; + const handleUnlock = async () => { + if (!note?.url || unlocking) return; + setUnlocking(true); + setUnlockError(null); + try { + const res = await authFetch('/api/articles/unlock', { + method: 'POST', + headers: { 'Content-Type': 'application/json' }, + body: JSON.stringify({ url: note.url, noteId: note.id }), + }); + const result = await res.json(); + if (result.success && result.archiveUrl) { + setNote({ ...note, archiveUrl: result.archiveUrl }); + } else { + setUnlockError(result.error || 'No archived version found'); + } + } catch (error) { + setUnlockError('Failed to unlock article'); + console.error('Unlock error:', error); + } finally { + setUnlocking(false); + } + }; + if (loading) { return (
@@ -241,16 +268,63 @@ export default function NoteDetailPage() {
- {/* URL */} + {/* URL + Unlock */} {note.url && ( - - {note.url} - +
+ + {note.url} + + + {note.archiveUrl ? ( +
+ + + + + View Unlocked Article + + {note.archiveUrl} +
+ ) : ( +
+ + {unlockError && ( + {unlockError} + )} +
+ )} +
)} {/* Uploaded file/image */} diff --git a/src/components/NoteCard.tsx b/src/components/NoteCard.tsx index d2e7f7c..1997865 100644 --- a/src/components/NoteCard.tsx +++ b/src/components/NoteCard.tsx @@ -22,9 +22,10 @@ interface NoteCardProps { updatedAt: string; tags: { id: string; name: string; color: string | null }[]; url?: string | null; + archiveUrl?: string | null; } -export function NoteCard({ id, title, type, contentPlain, isPinned, updatedAt, tags, url }: NoteCardProps) { +export function NoteCard({ id, title, type, contentPlain, isPinned, updatedAt, tags, url, archiveUrl }: NoteCardProps) { const snippet = (contentPlain || '').slice(0, 120); return ( @@ -41,6 +42,11 @@ export function NoteCard({ id, title, type, contentPlain, isPinned, updatedAt, t ★ )} + {archiveUrl && ( + + unlocked + + )} {new Date(updatedAt).toLocaleDateString()} diff --git a/src/lib/article-unlock.ts b/src/lib/article-unlock.ts new file mode 100644 index 0000000..df7569f --- /dev/null +++ b/src/lib/article-unlock.ts @@ -0,0 +1,232 @@ +/** + * Article Unlock — multi-strategy approach to get readable versions of + * paywalled or permissioned articles. + * + * Strategies (tried in order): + * 1. Wayback Machine — check for existing snapshot, or request a new one + * 2. Google Web Cache — fast lookup, often has full text + * 3. archive.ph — check for existing snapshots (read-only, no submission) + */ + +export interface UnlockResult { + success: boolean; + strategy: string; + archiveUrl?: string; + content?: string; + error?: string; +} + +const BROWSER_UA = + 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36'; + +// --------------------------------------------------------------------------- +// Strategy 1: Internet Archive Wayback Machine +// --------------------------------------------------------------------------- + +async function tryWaybackMachine(url: string): Promise { + // First check if a snapshot already exists + try { + const checkUrl = `https://archive.org/wayback/available?url=${encodeURIComponent(url)}`; + const res = await fetch(checkUrl, { + headers: { 'User-Agent': BROWSER_UA }, + signal: AbortSignal.timeout(10000), + }); + + if (res.ok) { + const data = await res.json(); + const snapshot = data?.archived_snapshots?.closest; + if (snapshot?.available && snapshot?.url) { + return { + success: true, + strategy: 'wayback', + archiveUrl: snapshot.url.replace('http://', 'https://'), + }; + } + } + } catch { + // availability check failed, try Save Page Now + } + + // No existing snapshot — request one via Save Page Now (SPN) + try { + const saveRes = await fetch('https://web.archive.org/save', { + method: 'POST', + headers: { + 'Content-Type': 'application/x-www-form-urlencoded', + 'User-Agent': BROWSER_UA, + Accept: 'application/json', + }, + body: `url=${encodeURIComponent(url)}&capture_all=1`, + signal: AbortSignal.timeout(30000), + }); + + if (saveRes.ok) { + const data = await saveRes.json(); + // SPN returns a job_id — we can construct the URL + if (data.url) { + return { + success: true, + strategy: 'wayback-save', + archiveUrl: data.url, + }; + } + if (data.job_id) { + // Poll for completion (up to 30s) + const archiveUrl = await pollWaybackJob(data.job_id); + if (archiveUrl) { + return { success: true, strategy: 'wayback-save', archiveUrl }; + } + } + } + + // Sometimes SPN redirects to the archived page directly + if (saveRes.status === 302 || saveRes.status === 301) { + const location = saveRes.headers.get('location'); + if (location) { + return { success: true, strategy: 'wayback-save', archiveUrl: location }; + } + } + } catch { + // SPN failed + } + + return null; +} + +async function pollWaybackJob(jobId: string): Promise { + for (let i = 0; i < 6; i++) { + await new Promise((r) => setTimeout(r, 5000)); + try { + const res = await fetch(`https://web.archive.org/save/status/${jobId}`, { + headers: { Accept: 'application/json', 'User-Agent': BROWSER_UA }, + signal: AbortSignal.timeout(10000), + }); + if (res.ok) { + const data = await res.json(); + if (data.status === 'success' && data.original_url && data.timestamp) { + return `https://web.archive.org/web/${data.timestamp}/${data.original_url}`; + } + if (data.status === 'error') return null; + } + } catch { + // keep polling + } + } + return null; +} + +// --------------------------------------------------------------------------- +// Strategy 2: Google Web Cache +// --------------------------------------------------------------------------- + +async function tryGoogleCache(url: string): Promise { + const cacheUrl = `https://webcache.googleusercontent.com/search?q=cache:${encodeURIComponent(url)}`; + try { + const res = await fetch(cacheUrl, { + headers: { 'User-Agent': BROWSER_UA }, + redirect: 'follow', + signal: AbortSignal.timeout(10000), + }); + + if (res.ok) { + // Google cache returns the full page — verify it's not an error page + const text = await res.text(); + if (text.length > 1000 && !text.includes('did not match any documents')) { + return { + success: true, + strategy: 'google-cache', + archiveUrl: cacheUrl, + content: text, + }; + } + } + } catch { + // Google cache not available + } + + return null; +} + +// --------------------------------------------------------------------------- +// Strategy 3: archive.ph (read-only — check for existing snapshots) +// --------------------------------------------------------------------------- + +async function tryArchivePh(url: string): Promise { + // Only check if an archive already exists — do NOT submit new pages + // (archive.ph has no API and aggressive anti-bot + security concerns) + const checkUrl = `https://archive.ph/newest/${encodeURIComponent(url)}`; + try { + const res = await fetch(checkUrl, { + headers: { 'User-Agent': BROWSER_UA }, + redirect: 'manual', // archive.ph redirects to the snapshot + signal: AbortSignal.timeout(10000), + }); + + // A 302 redirect means a snapshot exists + if (res.status === 302 || res.status === 301) { + const location = res.headers.get('location'); + if (location && location.includes('archive.ph/') && !location.includes('/submit')) { + return { + success: true, + strategy: 'archive-ph', + archiveUrl: location, + }; + } + } + + // A 200 with content also means it found one + if (res.ok) { + const finalUrl = res.url; + if (finalUrl && finalUrl !== checkUrl && finalUrl.includes('archive.ph/')) { + return { + success: true, + strategy: 'archive-ph', + archiveUrl: finalUrl, + }; + } + } + } catch { + // archive.ph not reachable + } + + return null; +} + +// --------------------------------------------------------------------------- +// Main unlock function +// --------------------------------------------------------------------------- + +export async function unlockArticle(url: string): Promise { + // Validate URL + try { + new URL(url); + } catch { + return { success: false, strategy: 'none', error: 'Invalid URL' }; + } + + // Try strategies in order of reliability + const strategies = [ + { name: 'Wayback Machine', fn: tryWaybackMachine }, + { name: 'Google Cache', fn: tryGoogleCache }, + { name: 'archive.ph', fn: tryArchivePh }, + ]; + + const errors: string[] = []; + + for (const { name, fn } of strategies) { + try { + const result = await fn(url); + if (result?.success) { + return result; + } + } catch (err) { + errors.push(`${name}: ${err instanceof Error ? err.message : 'unknown error'}`); + } + } + + return { + success: false, + strategy: 'none', + error: `No archived version found. Tried: ${strategies.map((s) => s.name).join(', ')}`, + }; +}