WIP: Export PDF with puppeteer

app.json

@@ -154,6 +154,9 @@
         },
         {
             "url": "https://github.com/heroku/heroku-buildpack-nodejs"
+        },
+        {
+            "url": "https://github.com/jontewks/puppeteer-heroku-buildpack"
         }
     ]
 }

lib/config/default.js

@@ -1,6 +1,7 @@
 'use strict'
 
 const os = require('os')
+const uuid = require('uuid/v4')
 
 module.exports = {
   domain: '',
@@ -186,5 +187,6 @@ module.exports = {
   //    2nd appearance: "31-good-morning-my-friend---do-you-have-5-1"
   //    3rd appearance: "31-good-morning-my-friend---do-you-have-5-2"
   linkifyHeaderStyle: 'keep-case',
-  autoVersionCheck: true
+  autoVersionCheck: true,
+  codimdSignKey: uuid()
 }

lib/note/index.js

@@ -1,5 +1,7 @@
 'use strict'
 
+const jwt = require('jsonwebtoken')
+
 const config = require('../config')
 const logger = require('../logger')
 
@@ -77,6 +79,24 @@ async function showNote (req, res) {
   return responseCodiMD(res, note)
 }
 
+async function authenticateUser (req) {
+  const authHeader = req.header('Authorization')
+  const token = authHeader && authHeader.replace('Bearer ', '')
+
+  if (token) {
+    const { userId } = jwt.verify(token, config.codimdSignKey)
+    const user = await User.findOne({ id: userId })
+
+    if (user) {
+      return [true, user]
+    } else {
+      return [false]
+    }
+  } else {
+    return [req.isAuthenticated(), req.user]
+  }
+}
+
 function canViewNote (note, isLogin, userId) {
   if (note.permission === 'private') {
     return note.ownerId === userId
@@ -87,6 +107,26 @@ function canViewNote (note, isLogin, userId) {
   return true
 }
 
+async function findNote (req, res, next) {
+  const noteId = req.params.noteId
+
+  const note = await getNoteById(noteId)
+
+  if (!note) {
+    return errorNotFound(req, res)
+  }
+
+  const [isAuthenticated, user] = await authenticateUser(req)
+
+  if (!canViewNote(note, isAuthenticated, user ? user.id : null)) {
+    return errorForbidden(req, res)
+  }
+
+  res.locals.note = note
+
+  next()
+}
+
 async function showPublishNote (req, res) {
   const shortid = req.params.shortid
 
@@ -141,18 +181,9 @@ async function showPublishNote (req, res) {
 }
 
 async function noteActions (req, res) {
+  const { note } = res.locals
   const noteId = req.params.noteId
 
-  const note = await getNoteById(noteId)
-
-  if (!note) {
-    return errorNotFound(req, res)
-  }
-
-  if (!canViewNote(note, req.isAuthenticated(), req.user ? req.user.id : null)) {
-    return errorForbidden(req, res)
-  }
-
   const action = req.params.action
   switch (action) {
     case 'publish':
@@ -191,3 +222,5 @@ async function noteActions (req, res) {
 exports.showNote = showNote
 exports.showPublishNote = showPublishNote
 exports.noteActions = noteActions
+exports.actionPDF = actionPDF
+exports.findNote = findNote

lib/note/noteActions.js

@@ -1,16 +1,16 @@
 'use strict'
 
 const fs = require('fs')
-const path = require('path')
-const markdownpdf = require('markdown-pdf')
+const jwt = require('jsonwebtoken')
+const puppeteer = require('puppeteer')
 const shortId = require('shortid')
 const querystring = require('querystring')
 const moment = require('moment')
 const { Pandoc } = require('@hackmd/pandoc.js')
 
 const config = require('../config')
 const logger = require('../logger')
-const { Note, Revision } = require('../models')
+const { Note, Revision, User } = require('../models')
 const { errorInternalError, errorNotFound } = require('../response')
 
 function actionPublish (req, res, note) {
@@ -64,43 +64,67 @@ function actionInfo (req, res, note) {
   res.send(data)
 }
 
+async function printPDF (noteUrl, headers = {}) {
+  const browser = await puppeteer.launch({
+    headless: true,
+    args: ['--disable-dev-shm-usage']
+  })
+
+  const page = await browser.newPage()
+  await page.setExtraHTTPHeaders(headers)
+  await page.goto(noteUrl, { waitUntil: 'networkidle0' })
+  const pdf = await page.pdf({ format: 'A4' })
+
+  await browser.close()
+  return pdf
+}
+
 function actionPDF (req, res, note) {
-  const url = config.serverURL || 'http://' + req.get('host')
-  const body = note.content
-  const extracted = Note.extractMeta(body)
-  let content = extracted.markdown
-  const title = Note.decodeTitle(note.title)
+  const noteId = req.params.noteId
 
-  const highlightCssPath = path.join(config.appRootPath, '/node_modules/highlight.js/styles/github-gist.css')
+  if (req.method === 'POST') {
+    const token = req.user && jwt.sign({ userId: req.user.id.toString() }, config.codimdSignKey, { expiresIn: 5 * 60 })
 
-  if (!fs.existsSync(config.tmpPath)) {
-    fs.mkdirSync(config.tmpPath)
-  }
-  const pdfPath = config.tmpPath + '/' + Date.now() + '.pdf'
-  content = content.replace(/\]\(\//g, '](' + url + '/')
-  const markdownpdfOptions = {
-    highlightCssPath: highlightCssPath
-  }
-  markdownpdf(markdownpdfOptions).from.string(content).to(pdfPath, function () {
-    if (!fs.existsSync(pdfPath)) {
-      logger.error('PDF seems to not be generated as expected. File doesn\'t exist: ' + pdfPath)
-      return errorInternalError(req, res)
+    const noteURL = `${config.serverURL}/${noteId}/pdf`
+
+    return printPDF(noteURL, {
+      Authorization: `Bearer ${token}`
+    }).then(pdf => {
+      res.set({ 'Content-Type': 'application/pdf', 'Content-Length': pdf.length })
+      res.send(pdf)
+    })
+  } else {
+    const body = note.content
+    const extracted = Note.extractMeta(body)
+    const markdown = extracted.markdown
+    const meta = Note.parseMeta(extracted.meta)
+    const createTime = note.createdAt
+    const updateTime = note.lastchangeAt
+    const title = Note.generateWebTitle(meta.title || Note.decodeTitle(note.title))
+
+    const data = {
+      title: title,
+      description: meta.description || (markdown ? Note.generateDescription(markdown) : null),
+      viewcount: note.viewcount,
+      createtime: createTime,
+      updatetime: updateTime,
+      body: body,
+      owner: note.owner ? note.owner.id : null,
+      ownerprofile: note.owner ? User.getProfile(note.owner) : null,
+      lastchangeuser: note.lastchangeuser ? note.lastchangeuser.id : null,
+      lastchangeuserprofile: note.lastchangeuser ? User.getProfile(note.lastchangeuser) : null,
+      robots: meta.robots || false, // default allow robots
+      GA: meta.GA,
+      disqus: meta.disqus,
+      cspNonce: res.locals.nonce
     }
-    const stream = fs.createReadStream(pdfPath)
-    let filename = title
-    // Be careful of special characters
-    filename = encodeURIComponent(filename)
-    // Ideally this should strip them
-    res.setHeader('Content-disposition', 'attachment; filename="' + filename + '.pdf"')
-    res.setHeader('Cache-Control', 'private')
-    res.setHeader('Content-Type', 'application/pdf; charset=UTF-8')
-    res.setHeader('X-Robots-Tag', 'noindex, nofollow') // prevent crawling
-    stream.on('end', () => {
-      stream.close()
-      fs.unlinkSync(pdfPath)
+
+    res.set({
+      'Cache-Control': 'private' // only cache by client
     })
-    stream.pipe(res)
-  })
+
+    return res.render('pretty.ejs', data)
+  }
 }
 
 const outputFormats = {

lib/routes.js

@@ -74,7 +74,8 @@ appRouter.get('/p/:shortid/:action', response.publishSlideActions)
 // get note by id
 appRouter.get('/:noteId', wrap(noteController.showNote))
 // note actions
-appRouter.get('/:noteId/:action', noteController.noteActions)
+appRouter.get('/:noteId/:action', noteController.findNote, noteController.noteActions)
+appRouter.post('/:noteId/:action', noteController.findNote, noteController.noteActions)
 // note actions with action id
 appRouter.get('/:noteId/:action/:actionId', noteController.noteActions)