prehistoric-systems/omnivore
1
0
Fork 0
mirror of https://github.com/omnivore-app/omnivore.git synced 2026-03-11 08:54:26 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Make puppeteer-parse a module

Browse source
This commit is contained in:
Hongbo Wu 2022-06-29 11:17:43 +08:00
parent 08aec0b73d
commit cb858484c6
8 changed files with 125 additions and 1073 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

635
packages/content-fetch/fetch-content.js
View file

@ -1,635 +0,0 @@
/* eslint-disable no-undef */
/* eslint-disable no-empty */
/* eslint-disable @typescript-eslint/explicit-function-return-type */
/* eslint-disable @typescript-eslint/no-var-requires */
/* eslint-disable @typescript-eslint/no-require-imports */
require('dotenv').config();
const Url = require('url');
const puppeteer = require('puppeteer-core');
const axios = require('axios');
const jwt = require('jsonwebtoken');
const { promisify } = require('util');
const { parseHTML } = require('linkedom');
const { preHandleContent } = require('@omnivore/content-handler');
const signToken = promisify(jwt.sign);
const MOBILE_USER_AGENT = 'Mozilla/5.0 (Linux; Android 6.0.1; Nexus 5X Build/MMB29P) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/93.0.4577.62 Mobile Safari/537.36 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)'
const DESKTOP_USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 11_6_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4372.0 Safari/537.36'
const BOT_DESKTOP_USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 11_6_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4372.0 Safari/537.36'
const NON_BOT_DESKTOP_USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 11_6_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4372.0 Safari/537.36'
const NON_BOT_HOSTS = ['bloomberg.com', 'forbes.com']
const NON_SCRIPT_HOSTS= ['medium.com', 'fastcompany.com'];
const ALLOWED_CONTENT_TYPES = ['text/html', 'application/octet-stream', 'text/plain', 'application/pdf'];
// Add stealth plugin to hide puppeteer usage
// const StealthPlugin = require('puppeteer-extra-plugin-stealth');
// puppeteer.use(StealthPlugin());
const userAgentForUrl = (url) => {
try {
const u = new URL(url);
for (const host of NON_BOT_HOSTS) {
if (u.hostname.endsWith(host)) {
return NON_BOT_DESKTOP_USER_AGENT;
}
}
} catch (e) {
console.log('error getting user agent for url', url, e)
}
return DESKTOP_USER_AGENT
};
const fetchContentWithScrapingBee = async (url) => {
const response = await axios.get('https://app.scrapingbee.com/api/v1', {
params: {
'api_key': process.env.SCRAPINGBEE_API_KEY,
'url': url,
'render_js': 'false',
'premium_proxy': 'true',
'country_code':'us'
}
})
const dom = parseHTML(response.data).document;
return { title: dom.title, domContent: dom.documentElement.outerHTML, url: url }
}
const enableJavascriptForUrl = (url) => {
try {
const u = new URL(url);
for (const host of NON_SCRIPT_HOSTS) {
if (u.hostname.endsWith(host)) {
return false;
}
}
} catch (e) {
console.log('error getting hostname for url', url, e)
}
return true
};
// launch Puppeteer
const getBrowserPromise = (async () => {
console.log("starting with proxy url", process.env.PROXY_URL)
return puppeteer.launch({
args: [
'--allow-running-insecure-content',
'--autoplay-policy=user-gesture-required',
'--disable-component-update',
'--disable-domain-reliability',
'--disable-features=AudioServiceOutOfProcess,IsolateOrigins,site-per-process',
'--disable-print-preview',
'--disable-setuid-sandbox',
'--disable-site-isolation-trials',
'--disable-speech-api',
'--disable-web-security',
'--disk-cache-size=33554432',
'--enable-features=SharedArrayBuffer',
'--hide-scrollbars',
'--ignore-gpu-blocklist',
'--in-process-gpu',
'--mute-audio',
'--no-default-browser-check',
'--no-pings',
'--no-sandbox',
'--no-zygote',
'--use-gl=swiftshader',
'--window-size=1920,1080',
].filter((item) => !!item),
defaultViewport: { height: 1080, width: 1920 },
executablePath: process.env.CHROMIUM_PATH,
headless: !!process.env.LAUNCH_HEADLESS,
timeout: 120000, // 2 minutes
});
})();
let logRecord, functionStartTime;
const uploadToSignedUrl = async ({ id, uploadSignedUrl }, contentType, contentObjUrl) => {
const stream = await axios.get(contentObjUrl, { responseType: 'stream' });
return await axios.put(uploadSignedUrl, stream.data, {
headers: {
'Content-Type': contentType,
},
maxBodyLength: 1000000000,
maxContentLength: 100000000,
})
};
const getUploadIdAndSignedUrl = async (userId, url, articleSavingRequestId) => {
const auth = await signToken({ uid: userId }, process.env.JWT_SECRET);
const data = JSON.stringify({
query: `mutation UploadFileRequest($input: UploadFileRequestInput!) {
uploadFileRequest(input:$input) {
... on UploadFileRequestError {
errorCodes
}
... on UploadFileRequestSuccess {
id
uploadSignedUrl
}
}
}`,
variables: {
input: {
url,
contentType: 'application/pdf',
clientRequestId: articleSavingRequestId,
}
}
});
const response = await axios.post(`${process.env.REST_BACKEND_ENDPOINT}/graphql`, data,
{
headers: {
Cookie: `auth=${auth};`,
'Content-Type': 'application/json',
},
});
return response.data.data.uploadFileRequest;
};
const uploadPdf = async (url, userId, articleSavingRequestId) => {
validateUrlString(url);
const uploadResult = await getUploadIdAndSignedUrl(userId, url, articleSavingRequestId);
await uploadToSignedUrl(uploadResult, 'application/pdf', url);
return uploadResult.id;
};
const sendCreateArticleMutation = async (userId, input) => {
const data = JSON.stringify({
query: `mutation CreateArticle ($input: CreateArticleInput!){
createArticle(input:$input){
... on CreateArticleSuccess{
createdArticle{
id
}
}
... on CreateArticleError{
errorCodes
}
}
}`,
variables: {
input: Object.assign({}, input , { source: 'puppeteer-parse' }),
},
});
const auth = await signToken({ uid: userId }, process.env.JWT_SECRET);
const response = await axios.post(`${process.env.REST_BACKEND_ENDPOINT}/graphql`, data,
{
headers: {
Cookie: `auth=${auth};`,
'Content-Type': 'application/json',
},
});
return response.data.data.createArticle;
};
const saveUploadedPdf = async (userId, url, uploadFileId, articleSavingRequestId) => {
return sendCreateArticleMutation(userId, {
url: encodeURI(url),
articleSavingRequestId,
uploadFileId: uploadFileId,
},
);
};
async function fetchContent(req, res) {
functionStartTime = Date.now();
let url = getUrl(req);
const userId = (req.query ? req.query.userId : undefined) || (req.body ? req.body.userId : undefined);
const articleSavingRequestId = (req.query ? req.query.saveRequestId : undefined) || (req.body ? req.body.saveRequestId : undefined);
console.log('user id', userId, 'url', url)
logRecord = {
url,
userId,
articleSavingRequestId,
labels: {
source: 'parseContent',
},
};
console.log(`Article parsing request`, logRecord);
if (!url) {
logRecord.urlIsInvalid = true;
console.log(`Valid URL to parse not specified`, logRecord);
return res.sendStatus(400);
}
// pre handle url with custom handlers
let title, content, contentType;
try {
const result = await preHandleContent(url);
if (result && result.url) {
url = result.url
validateUrlString(url);
}
if (result && result.title) { title = result.title }
if (result && result.content) { content = result.content }
if (result && result.contentType) { contentType = result.contentType }
} catch (e) {
console.log('error with handler: ', e);
}
let context, page, finalUrl;
try {
if ((!content || !title) && contentType !== 'application/pdf') {
const result = await retrievePage(url)
if (result && result.context) { context = result.context }
if (result && result.page) { page = result.page }
if (result && result.finalUrl) { finalUrl = result.finalUrl }
if (result && result.contentType) { contentType = result.contentType }
} else {
finalUrl = url
}
if (contentType === 'application/pdf') {
const uploadedFileId = await uploadPdf(finalUrl, userId, articleSavingRequestId);
const l = await saveUploadedPdf(userId, finalUrl, uploadedFileId, articleSavingRequestId);
} else {
if (!content || !title) {
const result = await retrieveHtml(page);
if (result.isBlocked) {
const sbResult = await fetchContentWithScrapingBee(url)
title = sbResult.title
content = sbResult.domContent
} else {
title = result.title;
content = result.domContent;
}
} else {
console.log('using prefetched content and title');
}
logRecord.fetchContentTime = Date.now() - functionStartTime;
const apiResponse = await sendCreateArticleMutation(userId, {
url: finalUrl,
articleSavingRequestId,
preparedDocument: {
document: content,
pageInfo: {
title,
canonicalUrl: finalUrl,
},
},
skipParsing: !content,
});
logRecord.totalTime = Date.now() - functionStartTime;
logRecord.result = apiResponse.createArticle;
}
} catch (e) {
logRecord.error = e.message;
console.log(`Error while retrieving page`, logRecord);
// fallback to scrapingbee
const sbResult = await fetchContentWithScrapingBee(url);
const sbUrl = finalUrl || sbResult.url;
const content = sbResult.domContent;
logRecord.fetchContentTime = Date.now() - functionStartTime;
const apiResponse = await sendCreateArticleMutation(userId, {
url: sbUrl,
articleSavingRequestId,
preparedDocument: {
document: content,
pageInfo: {
title: sbResult.title,
canonicalUrl: sbUrl,
},
},
skipParsing: !content,
});
logRecord.totalTime = Date.now() - functionStartTime;
logRecord.result = apiResponse.createArticle;
} finally {
if (context) {
await context.close();
}
console.log(`parse-page`, logRecord);
}
return res.sendStatus(200);
}
function validateUrlString(url) {
const u = new URL(url);
// Make sure the URL is http or https
if (u.protocol !== 'http:' && u.protocol !== 'https:') {
throw new Error('Invalid URL protocol check failed')
}
// Make sure the domain is not localhost
if (u.hostname === 'localhost' || u.hostname === '0.0.0.0') {
throw new Error('Invalid URL is localhost')
}
// Make sure the domain is not a private IP
if (/^(10|172\.16|192\.168)\..*/.test(u.hostname)) {
throw new Error('Invalid URL is private ip')
}
}
function getUrl(req) {
console.log('body', req.body)
const urlStr = (req.query ? req.query.url : undefined) || (req.body ? req.body.url : undefined);
if (!urlStr) {
throw new Error('No URL specified');
}
validateUrlString(urlStr);
const parsed = Url.parse(urlStr);
return parsed.href;
}
async function blockResources(client) {
const blockedResources = [
// Assets
// '*/favicon.ico',
// '.css',
// '.jpg',
// '.jpeg',
// '.png',
// '.svg',
// '.woff',
// Analytics and other fluff
'*.optimizely.com',
'everesttech.net',
'userzoom.com',
'doubleclick.net',
'googleadservices.com',
'adservice.google.com/*',
'connect.facebook.com',
'connect.facebook.net',
'sp.analytics.yahoo.com',
]
await client.send('Network.setBlockedURLs', { urls: blockedResources });
}
async function retrievePage(url) {
validateUrlString(url);
const browser = await getBrowserPromise;
logRecord.timing = { ...logRecord.timing, browserOpened: Date.now() - functionStartTime };
const context = await browser.createIncognitoBrowserContext();
const page = await context.newPage()
if (!enableJavascriptForUrl(url)) {
await page.setJavaScriptEnabled(false);
}
await page.setUserAgent(userAgentForUrl(url));
const client = await page.target().createCDPSession();
// intercept request when response headers was received
await client.send('Network.setRequestInterception', {
patterns: [
{
urlPattern: '*',
resourceType: 'Document',
interceptionStage: 'HeadersReceived',
},
],
});
const path = require('path');
const download_path = path.resolve('./download_dir/');
await client.send('Page.setDownloadBehavior', {
behavior: 'allow',
userDataDir: './',
downloadPath: download_path,
})
client.on('Network.requestIntercepted', async e => {
const headers = e.responseHeaders || {};
const [contentType] = (headers['content-type'] || headers['Content-Type'] || '')
.toLowerCase()
.split(';');
const obj = { interceptionId: e.interceptionId };
if (e.responseStatusCode >= 200 && e.responseStatusCode < 300) {
// We only check content-type on success responses
// as it doesn't matter what the content type is for things
// like redirects
if (contentType && !ALLOWED_CONTENT_TYPES.includes(contentType)) {
obj['errorReason'] = 'BlockedByClient';
}
}
try {
await client.send('Network.continueInterceptedRequest', obj);
// eslint-disable-next-line no-empty
} catch {}
});
await blockResources(client);
/*
* Disallow MathJax from running in Puppeteer and modifying the document,
* we shall instead run it in our frontend application to transform any
* mathjax content when present.
*/
await page.setRequestInterception(true);
let requestCount = 0;
page.on('request', request => {
if (['font', 'image', 'media'].includes(request.resourceType())) {
request.abort();
return;
}
if (requestCount++ > 100) {
request.abort();
return;
}
if (
request.resourceType() === 'script' &&
request.url().toLowerCase().indexOf('mathjax') > -1
) {
request.abort();
return
}
request.continue();
});
// Puppeteer fails during download of PDf files,
// so record the failure and use those items
let lastPdfUrl = undefined;
page.on('response', response => {
if (response.headers()['content-type'] === 'application/pdf') {
lastPdfUrl = response.url();
}
});
try {
const response = await page.goto(url, { timeout: 8 * 1000, waitUntil: ['networkidle2'] });
const finalUrl = response.url();
const contentType = response.headers()['content-type'];
logRecord.finalUrl = response.url();
logRecord.contentType = response.headers()['content-type'];
return { context, page, response, finalUrl, contentType };
} catch (error) {
if (lastPdfUrl) {
return { context, page, finalUrl: lastPdfUrl, contentType: 'application/pdf' };
}
await context.close();
throw error;
}
}
async function retrieveHtml(page) {
let domContent = '', title;
try {
title = await page.title();
logRecord.title = title;
const pageScrollingStart = Date.now();
/* scroll with a 5 second timeout */
await Promise.race([
new Promise(resolve => {
(async function () {
try {
await page.evaluate(`(async () => {
/* credit: https://github.com/puppeteer/puppeteer/issues/305 */
return new Promise((resolve, reject) => {
let scrollHeight = document.body.scrollHeight;
let totalHeight = 0;
let distance = 500;
let timer = setInterval(() => {
window.scrollBy(0, distance);
totalHeight += distance;
if(totalHeight >= scrollHeight){
clearInterval(timer);
resolve(true);
}
}, 10);
});
})()`);
} catch (e) {
logRecord.scrollError = true;
} finally {
resolve(true);
}
})();
}),
await page.waitForTimeout(1000),
]);
logRecord.timing = { ...logRecord.timing, pageScrolled: Date.now() - pageScrollingStart };
const iframes = {};
const urls = [];
const framesPromises = [];
const allowedUrls = /instagram\.com/gi;
for (const frame of page.mainFrame().childFrames()) {
if (frame.url() && allowedUrls.test(frame.url())) {
urls.push(frame.url());
framesPromises.push(frame.evaluate(el => el.innerHTML, await frame.$('body')));
}
}
(await Promise.all(framesPromises)).forEach((frame, index) => (iframes[urls[index]] = frame));
const domContentCapturingStart = Date.now();
// get document body with all hidden elements removed
domContent = await page.evaluate(iframes => {
const BI_SRC_REGEXP = /url\("(.+?)"\)/gi;
Array.from(document.body.getElementsByTagName('*')).forEach(el => {
const style = window.getComputedStyle(el);
try {
// Removing blurred images since they are mostly the copies of lazy loaded ones
if (['img', 'image'].includes(el.tagName.toLowerCase())) {
const filter = style.getPropertyValue('filter');
if (filter && filter.startsWith('blur')) {
el.parentNode && el.parentNode.removeChild(el);
}
}
} catch (err) {
// throw Error('error with element: ' + JSON.stringify(Array.from(document.body.getElementsByTagName('*'))))
}
// convert all nodes with background image to img nodes
if (!['', 'none'].includes(style.getPropertyValue('background-image'))) {
const filter = style.getPropertyValue('filter');
// avoiding image nodes with a blur effect creation
if (filter && filter.startsWith('blur')) {
el && el.parentNode && el.parentNode.removeChild(el);
} else {
const matchedSRC = BI_SRC_REGEXP.exec(style.getPropertyValue('background-image'));
// Using "g" flag with a regex we have to manually break down lastIndex to zero after every usage
// More details here: https://stackoverflow.com/questions/1520800/why-does-a-regexp-with-global-flag-give-wrong-results
BI_SRC_REGEXP.lastIndex = 0;
if (matchedSRC && matchedSRC[1] && !el.src) {
// Replacing element only of there are no content inside, b/c might remove important div with content.
// Article example: http://www.josiahzayner.com/2017/01/genetic-designer-part-i.html
// DIV with class "content-inner" has `url("https://resources.blogblog.com/blogblog/data/1kt/travel/bg_container.png")` background image.
if (el.innerHTML.length < 25) {
const img = document.createElement('img');
img.src = matchedSRC[1];
el && el.parentNode && el.parentNode.removeChild(el);
}
}
}
}
if (el.tagName === 'IFRAME') {
if (iframes[el.src]) {
const newNode = document.createElement('div');
newNode.className = 'omnivore-instagram-embed';
newNode.innerHTML = iframes[el.src];
el && el.parentNode && el.parentNode.replaceChild(newNode, el);
}
}
});
if (document.querySelector('[data-translate="managed_checking_msg"]') ||
document.getElementById('px-block-form-wrapper')) {
return 'IS_BLOCKED'
}
return document.documentElement.outerHTML;
}, iframes);
logRecord.puppeteerSuccess = true;
logRecord.timing = {
...logRecord.timing,
contenCaptured: Date.now() - domContentCapturingStart,
};
// [END puppeteer-block]
} catch (e) {
if (e.message.startsWith('net::ERR_BLOCKED_BY_CLIENT at ')) {
logRecord.blockedByClient = true;
} else {
logRecord.puppeteerSuccess = false;
logRecord.puppeteerError = {
message: e.message,
stack: e.stack,
};
}
}
if (domContent === 'IS_BLOCKED') {
return { isBlocked: true };
}
return { domContent, title };
}
module.exports = fetchContent;

8
packages/puppeteer-parse/.env.example
View file

@ -1,8 +0,0 @@
# Should match with the JWT_SECRET that the api uses
JWT_SECRET=some_secret
# Address of the backend that is running locally
REST_BACKEND_ENDPOINT=http://localhost:4000/api
# set for local development
IS_LOCAL=true

21
packages/puppeteer-parse/.gcloudignore
View file

@ -1,21 +0,0 @@
# This file specifies files that are *not* uploaded to Google Cloud Platform
# using gcloud. It follows the same syntax as .gitignore, with the addition of
# "#!include" directives (which insert the entries of the given .gitignore-style
# file at that point).
#
# For more information, run:
# $ gcloud topic gcloudignore
#
.gcloudignore
# If you would like to upload your .git directory, .gitignore file or files
# from your .gitignore file, remove the corresponding line
# below:
.git
.gitignore
node_modules
.env*
.secrets*
Dockerfile*
previewImage.*
*.sa.json

1
packages/puppeteer-parse/.gitignore vendored
View file

@ -1 +0,0 @@
previewImage.*

113
packages/puppeteer-parse/Dockerfile
View file

@ -1,113 +0,0 @@
# FROM node:14-slim
# # Taken from pu
# # Install latest chrome dev package and fonts to support major charsets (Chinese, Japanese, Arabic, Hebrew, Thai and a few others)
# # Note: this installs the necessary libs to make the bundled version of Chromium that Puppeteer
# # installs, work.
# RUN apt-get update \
# && apt-get install -y wget gnupg \
# && wget -q -O - https://dl-ssl.google.com/linux/linux_signing_key.pub | apt-key add - \
# && sh -c 'echo "deb [arch=amd64] http://dl.google.com/linux/chrome/deb/ stable main" >> /etc/apt/sources.list.d/google.list' \
# && apt-get update \
# && apt-get install -y google-chrome-stable fonts-ipafont-gothic fonts-wqy-zenhei fonts-thai-tlwg fonts-kacst fonts-freefont-ttf libxss1 \
# --no-install-recommends \
# && rm -rf /var/lib/apt/lists/*
# ENV PUPPETEER_SKIP_CHROMIUM_DOWNLOAD true
# ENV CHROMIUM_PATH "/usr/bin/google-chrome-stable"
# ------------------------
# FROM --platform=linux/arm64 node:14.18
# RUN apt-get update \
# && apt-get install -y chromium \
# && apt-get install -y ca-certificates \
# fonts-liberation \
# libappindicator3-1 \
# libasound2 \
# libatk-bridge2.0-0 \
# libatk1.0-0 \
# libc6 \
# libcairo2 \
# libcups2 \
# libdbus-1-3 \
# libexpat1 \
# libfontconfig1 \
# libgbm1 \
# libgcc1 \
# libglib2.0-0 \
# libgtk-3-0 \
# libnspr4 \
# libnss3 \
# libpango-1.0-0 \
# libpangocairo-1.0-0 \
# libstdc++6 \
# libx11-6 \
# libx11-xcb1 \
# libxcb1 \
# libxcomposite1 \
# libxcursor1 \
# libxdamage1 \
# libxext6 \
# libxfixes3 \
# libxi6 \
# libxrandr2 \
# libxrender1 \
# libxss1 \
# libxtst6 \
# lsb-release \
# wget \
# xdg-utils
FROM node:14.18-alpine
# Installs latest Chromium (92) package.
RUN apk add --no-cache \
chromium \
nss \
freetype \
harfbuzz \
ca-certificates \
ttf-freefont \
nodejs \
yarn
# Add user so we don't need --no-sandbox.
RUN addgroup -S pptruser && adduser -S -g pptruser pptruser \
&& mkdir -p /home/pptruser/Downloads /app \
&& chown -R pptruser:pptruser /home/pptruser \
&& chown -R pptruser:pptruser /app
# Run everything after as non-privileged user.
WORKDIR /app
ENV CHROMIUM_PATH /usr/bin/chromium-browser
ENV LAUNCH_HEADLESS=true
ENV PORT 9090
COPY package.json .
COPY yarn.lock .
COPY tsconfig.json .
COPY .prettierrc .
COPY .eslintrc .
COPY /packages/puppeteer-parse/package.json ./packages/puppeteer-parse/package.json
COPY /packages/content-handler/package.json ./packages/content-handler/package.json
RUN yarn install --pure-lockfile
ADD /packages/puppeteer-parse ./packages/puppeteer-parse
ADD /packages/content-handler ./packages/content-handler
RUN yarn workspace @omnivore/content-handler build
# After building, fetch the production dependencies
RUN rm -rf /app/packages/puppeteer-parse/node_modules
RUN rm -rf /app/node_modules
RUN yarn install --pure-lockfile --production
EXPOSE 9090
# USER pptruser
ENTRYPOINT ["yarn", "workspace", "@omnivore/puppeteer-parse", "start"]

23
packages/puppeteer-parse/README.md
View file

@ -1,24 +1,3 @@
# Puppeteer parsing function handler
This workspace is used to provide the GCF for the app to hande requests for the article parsing via Puppeteer.
## Using locally
Copy .env.example file to .env file: `cp .env.example .env`
Run `yarn start` to start the Google Cloud Function locally (Works without hot reloading).
After this, you should be able to access the functon on [http://localhost:8080/puppeteer](http://localhost:8080/puppeteer)
## Deployment
To deploy the function use the following command:
`gcloud functions deploy puppeteer --runtime nodejs12 --trigger-http --memory 1GB --set-env-vars REST_BACKEND_ENDPOINT=<backend-address>,JWT_SECRET=<jwt-secret>`
where:
`<backend-address>` - address of the backend server (e.g "http://localhost:4000")
`<jwt-secret>` - JWT secret that the backend server is using (e.g "some_secret")
This workspace is used to provide the module for the app to hande requests for the article parsing via Puppeteer.

393
packages/puppeteer-parse/index.js
View file

@ -5,6 +5,7 @@
/* eslint-disable @typescript-eslint/no-require-imports */
require('dotenv').config();
const Url = require('url');
const puppeteer = require('puppeteer-extra');
const axios = require('axios');
const jwt = require('jsonwebtoken');
const { promisify } = require('util');
@ -38,69 +39,16 @@ const DESKTOP_USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 11_6_0) Apple
const BOT_DESKTOP_USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 11_6_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4372.0 Safari/537.36'
const NON_BOT_DESKTOP_USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 11_6_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/89.0.4372.0 Safari/537.36'
const NON_BOT_HOSTS = ['bloomberg.com', 'forbes.com']
const NON_SCRIPT_HOSTS= ['medium.com', 'fastcompany.com'];
const filePath = `${os.tmpdir()}/previewImage.png`;
const path = require("path");
const ALLOWED_CONTENT_TYPES = ['text/html', 'application/octet-stream', 'text/plain', 'application/pdf'];
const { parseHTML } = require('linkedom');
const colors = {
emerg: 'inverse underline magenta',
alert: 'underline magenta',
crit: 'inverse underline red', // Any error that is forcing a shutdown of the service or application to prevent data loss.
error: 'underline red', // Any error which is fatal to the operation, but not the service or application
warning: 'underline yellow', // Anything that can potentially cause application oddities
notice: 'underline cyan', // Normal but significant condition
info: 'underline green', // Generally useful information to log
debug: 'underline gray',
};
const googleConfigs = {
level: 'info',
logName: 'logger',
levels: config.syslog.levels,
resource: {
labels: {
function_name: process.env.FUNCTION_TARGET,
project_id: process.env.GCP_PROJECT,
},
type: 'cloud_function',
},
};
function localConfig(id) {
return {
level: 'debug',
format: format.combine(
format.colorize({ all: true, colors }),
format(info =>
Object.assign(info, {
timestamp: DateTime.local().toLocaleString(DateTime.TIME_24_WITH_SECONDS),
}),
)(),
format.printf(info => {
// eslint-disable-next-line @typescript-eslint/no-unused-vars
const { timestamp, message, level, ...meta } = info;
return `[${id}@${info.timestamp}] ${info.message}${
Object.keys(meta).length ? '\n' + JSON.stringify(meta, null, 4) : ''
}`;
}),
),
};
}
function buildLoggerTransport(id, options) {
return process.env.IS_LOCAL
? new transports.Console(localConfig(id))
: new LoggingWinston({ ...googleConfigs, ...{ logName: id }, ...options });
}
function buildLogger(id, options) {
return loggers.get(id, {
levels: config.syslog.levels,
transports: [buildLoggerTransport(id, options)],
});
}
// Add stealth plugin to hide puppeteer usage
const StealthPlugin = require('puppeteer-extra-plugin-stealth');
puppeteer.use(StealthPlugin());
const userAgentForUrl = (url) => {
try {
@ -116,15 +64,38 @@ const userAgentForUrl = (url) => {
return DESKTOP_USER_AGENT
};
const fetchContentWithScrapingBee = async (url) => {
const response = await axios.get('https://app.scrapingbee.com/api/v1', {
params: {
'api_key': process.env.SCRAPINGBEE_API_KEY,
'url': url,
'render_js': 'false',
'premium_proxy': 'true',
'country_code':'us'
}
})
const dom = parseHTML(response.data).document;
return { title: dom.title, domContent: dom.documentElement.outerHTML, url: url }
}
const enableJavascriptForUrl = (url) => {
try {
const u = new URL(url);
for (const host of NON_SCRIPT_HOSTS) {
if (u.hostname.endsWith(host)) {
return false;
}
}
} catch (e) {
console.log('error getting hostname for url', url, e)
}
return true
};
// launch Puppeteer
const getBrowserPromise = (async () => {
// return puppeteer.launch({
// args: chromium.args,
// defaultViewport: chromium.defaultViewport,
// executablePath: process.env.CHROMIUM_PATH,
// headless: chromium.headless,
// ignoreHTTPSErrors: true,
// });
const getBrowserPromise = (async (proxyUrl, chromiumPath) => {
console.log("starting with proxy url", proxyUrl)
return puppeteer.launch({
args: [
'--allow-running-insecure-content',
@ -151,8 +122,8 @@ const getBrowserPromise = (async () => {
'--window-size=1920,1080',
].filter((item) => !!item),
defaultViewport: { height: 1080, width: 1920 },
executablePath: process.env.CHROMIUM_PATH,
headless: !!process.env.LAUNCH_HEADLESS,
executablePath: chromiumPath,
headless: true,
timeout: 120000, // 2 minutes
});
})();
@ -170,7 +141,7 @@ const uploadToSignedUrl = async ({ id, uploadSignedUrl }, contentType, contentOb
})
};
const getUploadIdAndSignedUrl = async (userId, url) => {
const getUploadIdAndSignedUrl = async (userId, url, articleSavingRequestId) => {
const auth = await signToken({ uid: userId }, process.env.JWT_SECRET);
const data = JSON.stringify({
query: `mutation UploadFileRequest($input: UploadFileRequestInput!) {
@ -188,17 +159,18 @@ const getUploadIdAndSignedUrl = async (userId, url) => {
input: {
url,
contentType: 'application/pdf',
clientRequestId: articleSavingRequestId,
}
}
});
const response = await axios.post(`${process.env.REST_BACKEND_ENDPOINT}/graphql`, data,
{
headers: {
Cookie: `auth=${auth};`,
'Content-Type': 'application/json',
},
});
{
headers: {
Cookie: `auth=${auth};`,
'Content-Type': 'application/json',
},
});
return response.data.data.uploadFileRequest;
};
@ -231,12 +203,12 @@ const sendCreateArticleMutation = async (userId, input) => {
const auth = await signToken({ uid: userId }, process.env.JWT_SECRET);
const response = await axios.post(`${process.env.REST_BACKEND_ENDPOINT}/graphql`, data,
{
headers: {
Cookie: `auth=${auth};`,
'Content-Type': 'application/json',
},
});
{
headers: {
Cookie: `auth=${auth};`,
'Content-Type': 'application/json',
},
});
return response.data.data.createArticle;
};
@ -249,24 +221,8 @@ const saveUploadedPdf = async (userId, url, uploadFileId, articleSavingRequestId
);
};
/**
* Cloud Function entry point, HTTP trigger.
* Loads the requested URL via Puppeteer, captures page content and sends it to backend
*
* @param {Object} req Cloud Function request context.
* @param {Object} res Cloud Function response context.
*/
exports.puppeteer = Sentry.GCPFunction.wrapHttpFunction(async (req, res) => {
async function fetchContent(req, res) {
functionStartTime = Date.now();
// Grabbing execution and trace ids to attach logs to the appropriate function call
const execution_id = req.get('function-execution-id');
const traceId = (req.get('x-cloud-trace-context') || '').split('/')[0];
const logger = buildLogger('cloudfunctions.googleapis.com%2Fcloud-functions', {
trace: `projects/${process.env.GCLOUD_PROJECT}/traces/${traceId}`,
labels: {
execution_id: execution_id,
},
});
let url = getUrl(req);
const userId = req.body.userId || req.query.userId;
@ -281,7 +237,7 @@ exports.puppeteer = Sentry.GCPFunction.wrapHttpFunction(async (req, res) => {
},
};
logger.info(`Article parsing request`, logRecord);
console.log(`Article parsing request`, logRecord);
if (!url) {
logRecord.urlIsInvalid = true;
@ -306,16 +262,15 @@ exports.puppeteer = Sentry.GCPFunction.wrapHttpFunction(async (req, res) => {
let context, page, finalUrl;
try {
if ((!content || !title) && contentType !== 'application/pdf') {
const result = await retrievePage(url)
if (result && result.context) { context = result.context }
if (result && result.page) { page = result.page }
if (result && result.finalUrl) { finalUrl = result.finalUrl }
if (result && result.contentType) { contentType = result.contentType }
console.log('context, page, finalUrl, contentType', context, page, finalUrl, contentType);
} else {
finalUrl = url
}
if ((!content || !title) && contentType !== 'application/pdf') {
const result = await retrievePage(url)
if (result && result.context) { context = result.context }
if (result && result.page) { page = result.page }
if (result && result.finalUrl) { finalUrl = result.finalUrl }
if (result && result.contentType) { contentType = result.contentType }
} else {
finalUrl = url
}
if (contentType === 'application/pdf') {
const uploadedFileId = await uploadPdf(finalUrl, userId, articleSavingRequestId);
@ -323,14 +278,20 @@ exports.puppeteer = Sentry.GCPFunction.wrapHttpFunction(async (req, res) => {
} else {
if (!content || !title) {
const result = await retrieveHtml(page);
title = result.title;
content = result.domContent;
if (result.isBlocked) {
const sbResult = await fetchContentWithScrapingBee(url)
title = sbResult.title
content = sbResult.domContent
} else {
title = result.title;
content = result.domContent;
}
} else {
console.log('using prefetched content and title');
console.log(content);
}
logRecord.contentFetchTime = Date.now() - functionStartTime;
logRecord.fetchContentTime = Date.now() - functionStartTime;
const apiResponse = await sendCreateArticleMutation(userId, {
url: finalUrl,
@ -347,12 +308,12 @@ exports.puppeteer = Sentry.GCPFunction.wrapHttpFunction(async (req, res) => {
logRecord.totalTime = Date.now() - functionStartTime;
logRecord.result = apiResponse.createArticle;
logger.info(`parse-page`, logRecord);
console.log(`parse-page`, logRecord);
}
} catch (e) {
console.log('error', e)
logRecord.error = e.message;
logger.error(`Error while retrieving page`, logRecord);
console.log(`Error while retrieving page`, logRecord);
return res.sendStatus(503);
} finally {
if (context) {
@ -361,136 +322,7 @@ exports.puppeteer = Sentry.GCPFunction.wrapHttpFunction(async (req, res) => {
}
return res.sendStatus(200);
});
/**
* Cloud Function entry point, HTTP trigger.
* Loads the requested URL via Puppeteer and captures a screenshot of the provided element
*
* @param {Object} req Cloud Function request context.
* Inlcudes:
* * url - URL address of the page to open
* @param {Object} res Cloud Function response context.
*/
exports.preview = Sentry.GCPFunction.wrapHttpFunction(async (req, res) => {
functionStartTime = Date.now();
// Grabbing execution and trace ids to attach logs to the appropriate function call
const execution_id = req.get('function-execution-id');
const traceId = (req.get('x-cloud-trace-context') || '').split('/')[0];
const logger = buildLogger('cloudfunctions.googleapis.com%2Fcloud-functions', {
trace: `projects/${process.env.GCLOUD_PROJECT}/traces/${traceId}`,
labels: {
execution_id: execution_id,
},
});
if (!process.env.PREVIEW_IMAGE_BUCKET) {
logger.error(`PREVIEW_IMAGE_BUCKET not set`)
return res.sendStatus(500);
}
const url = getUrl(req);
console.log('preview request url', url);
logRecord = {
url,
query: req.query,
origin: req.get('Origin'),
labels: {
source: 'publicImagePreview',
},
};
logger.info(`Public preview image generation request`, logRecord);
if (!url) {
logRecord.urlIsInvalid = true;
logger.error(`Valid URL to parse is not specified`, logRecord);
return res.sendStatus(400);
}
const { origin } = new URL(url);
if (!ALLOWED_ORIGINS.some(o => o === origin)) {
logRecord.forbiddenOrigin = true;
logger.error(`This origin is not allowed: ${origin}`, logRecord);
return res.sendStatus(400);
}
const browser = await getBrowserPromise;
logRecord.timing = { ...logRecord.timing, browserOpened: Date.now() - functionStartTime };
const page = await browser.newPage();
const pageLoadingStart = Date.now();
const modifiedUrl = new URL(url);
modifiedUrl.searchParams.append('fontSize', 24);
modifiedUrl.searchParams.append('adjustAspectRatio', 1.91);
try {
await page.goto(modifiedUrl);
logRecord.timing = { ...logRecord.timing, pageLoaded: Date.now() - pageLoadingStart };
} catch (error) {
console.log('error going to page: ', modifiedUrl)
console.log(error)
throw error
}
// We lookup the destination path from our own page content and avoid trusting any passed query params
// selector - CSS selector of the element to get screenshot of
const selector = decodeURIComponent(
await page.$eval(
"head > meta[name='omnivore:preview_image_selector']",
element => element.content,
),
);
if (!selector) {
logRecord.selectorIsInvalid = true;
logger.error(`Valid element selector is not specified`, logRecord);
await page.close();
return res.sendStatus(400);
}
logRecord.selector = selector;
// destination - destination pathname for the image to save with
const destination = decodeURIComponent(
await page.$eval(
"head > meta[name='omnivore:preview_image_destination']",
element => element.content,
),
);
if (!destination) {
logRecord.destinationIsInvalid = true;
logger.error(`Valid file destination is not specified`, logRecord);
await page.close();
return res.sendStatus(400);
}
logRecord.destination = destination;
const screenshotTakingStart = Date.now();
try {
await page.waitForSelector(selector, { timeout: 3000 }); // wait for the selector to load
} catch (error) {
logRecord.elementNotFound = true;
logger.error(`Element is not presented on the page`, logRecord);
await page.close();
return res.sendStatus(400);
}
const element = await page.$(selector);
await element.screenshot({ path: filePath }); // take screenshot of the element in puppeteer
logRecord.timing = { ...logRecord.timing, screenshotTaken: Date.now() - screenshotTakingStart };
await page.close();
try {
const [file] = await previewBucket.upload(filePath, {
destination,
metadata: logRecord,
});
logRecord.file = file.metadata;
} catch (e) {
console.log('error uploading to bucket, this is non-fatal', e)
}
logger.info(`preview-image`, logRecord);
return res.redirect(`${process.env.PREVIEW_IMAGE_CDN_ORIGIN}/${destination}`);
});
}
function validateUrlString(url) {
const u = new URL(url);
@ -509,16 +341,15 @@ function validateUrlString(url) {
}
function getUrl(req) {
if (req.query.url || req.body.url) {
const urlStr = req.query.url || req.body.url;
validateUrlString(urlStr);
const url = Url.parse(urlStr);
return url.href;
const urlStr = (req.query ? req.query.url : undefined) || (req.body ? req.body.url : undefined);
if (!urlStr) {
throw new Error('No URL specified');
}
try {
return Url.parse(JSON.parse(req.body).url).href;
} catch (e) {}
validateUrlString(urlStr);
const parsed = Url.parse(urlStr);
return parsed.href;
}
async function blockResources(client) {
@ -554,7 +385,11 @@ async function retrievePage(url) {
logRecord.timing = { ...logRecord.timing, browserOpened: Date.now() - functionStartTime };
const context = await browser.createIncognitoBrowserContext();
const page = await context.newPage();
const page = await context.newPage()
if (!enableJavascriptForUrl(url)) {
await page.setJavaScriptEnabled(false);
}
await page.setUserAgent(userAgentForUrl(url));
const client = await page.target().createCDPSession();
@ -570,16 +405,15 @@ async function retrievePage(url) {
],
});
const path = require('path');
const download_path = path.resolve('./download_dir/');
await client.send('Page.setDownloadBehavior', {
behavior: 'allow',
userDataDir: './',
downloadPath: download_path,
behavior: 'allow',
userDataDir: './',
downloadPath: download_path,
})
client.on('Network.requestIntercepted', async e => {
client.on('Network.requestIntercepted', async (e) => {
const headers = e.responseHeaders || {};
const [contentType] = (headers['content-type'] || headers['Content-Type'] || '')
@ -605,10 +439,10 @@ async function retrievePage(url) {
await blockResources(client);
/*
* Disallow MathJax from running in Puppeteer and modifying the document,
* we shall instead run it in our frontend application to transform any
* mathjax content when present.
*/
* Disallow MathJax from running in Puppeteer and modifying the document,
* we shall instead run it in our frontend application to transform any
* mathjax content when present.
*/
await page.setRequestInterception(true);
let requestCount = 0;
page.on('request', request => {
@ -630,7 +464,6 @@ async function retrievePage(url) {
request.continue();
});
// Puppeteer fails during download of PDf files,
// so record the failure and use those items
let lastPdfUrl = undefined;
@ -719,12 +552,16 @@ async function retrieveHtml(page) {
Array.from(document.body.getElementsByTagName('*')).forEach(el => {
const style = window.getComputedStyle(el);
// Removing blurred images since they are mostly the copies of lazy loaded ones
if (['img', 'image'].includes(el.tagName.toLowerCase())) {
const filter = style.getPropertyValue('filter');
if (filter && filter.startsWith('blur')) {
el.parentNode && el.parentNode.removeChild(el);
try {
// Removing blurred images since they are mostly the copies of lazy loaded ones
if (['img', 'image'].includes(el.tagName.toLowerCase())) {
const filter = style.getPropertyValue('filter');
if (filter && filter.startsWith('blur')) {
el.parentNode && el.parentNode.removeChild(el);
}
}
} catch (err) {
// throw Error('error with element: ' + JSON.stringify(Array.from(document.body.getElementsByTagName('*'))))
}
// convert all nodes with background image to img nodes
@ -761,6 +598,12 @@ async function retrieveHtml(page) {
}
}
});
if (document.querySelector('[data-translate="managed_checking_msg"]') ||
document.getElementById('px-block-form-wrapper')) {
return 'IS_BLOCKED'
}
return document.documentElement.outerHTML;
}, iframes);
logRecord.puppeteerSuccess = true;
@ -781,5 +624,15 @@ async function retrieveHtml(page) {
};
}
}
if (domContent === 'IS_BLOCKED') {
return { isBlocked: true };
}
return { domContent, title };
}
module.exports = {
fetchContent,
getBrowserPromise,
getUrl,
};

4
packages/puppeteer-parse/package.json
View file

@ -1,7 +1,7 @@
{
"name": "@omnivore/puppeteer-parse",
"version": "1.0.0",
"description": "Google Cloud Function that accepts URL of the article and parses its content",
"description": "Accepts URL of the article and parses its content",
"main": "index.js",
"dependencies": {
"@google-cloud/functions-framework": "^3.1.2",
@ -20,8 +20,6 @@
"winston": "^3.3.3"
},
"scripts": {
"start": "npx functions-framework --port=9090 --target=puppeteer",
"start_preview": "npx functions-framework --target=preview",
"test": "yarn mocha"
}
}