apiserver/feed.py

@@ -46,7 +46,7 @@ def get_article(url):
     logging.info('Trying our server instead...')
 
     try:
-        r = requests.post(READ_API, data=dict(url=url), timeout=20)
+        r = requests.post(READ_API, data=dict(url=url), timeout=10)
         if r.status_code != 200:
             raise Exception('Bad response code ' + str(r.status_code))
         return r.text
@@ -94,6 +94,7 @@ def update_story(story, is_manual=False):
         return False
 
     if story.get('url', '') and not story.get('text', ''):
+        logging.info('inside if')
         if not get_content_type(story['url']).startswith('text/'):
             logging.info('URL invalid file type / content type:')
             logging.info(story['url'])

apiserver/feeds/manual.py

@@ -7,15 +7,12 @@ import requests
 import time
 from bs4 import BeautifulSoup
 
-USER_AGENT = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:77.0) Gecko/20100101 Firefox/77.0'
+USER_AGENT = 'Twitterbot/1.0'
 
 def api(route):
     try:
-        headers = {
-            'User-Agent': USER_AGENT,
-            'X-Forwarded-For': '66.249.66.1',
-        }
-        r = requests.get(route, headers=headers, timeout=10)
+        headers = {'User-Agent': USER_AGENT}
+        r = requests.get(route, headers=headers, timeout=5)
         if r.status_code != 200:
             raise Exception('Bad response code ' + str(r.status_code))
         return r.text

readerserver/main.js

@@ -4,7 +4,7 @@ const port = 33843;
 
 const request = require('request');
 const JSDOM = require('jsdom').JSDOM;
-const { Readability } = require('readability');
+const Readability = require('readability');
 
 app.use(express.urlencoded({ extended: true }));
 
@@ -36,11 +36,7 @@ app.post('/', (req, res) => {
 	const requestOptions = {
 		url: url,
 		//headers: {'User-Agent': 'Googlebot/2.1 (+http://www.google.com/bot.html)'},
-		//headers: {'User-Agent': 'Twitterbot/1.0'},
-		headers: {
-			'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:77.0) Gecko/20100101 Firefox/77.0',
-			'X-Forwarded-For': '66.249.66.1',
-		},
+		headers: {'User-Agent': 'Twitterbot/1.0'},
 	};
 
 	console.log('Parse request for:', url);

readerserver/yarn.lock

@@ -712,8 +712,8 @@ raw-body@2.4.0:
     unpipe "1.0.0"
 
 "readability@https://github.com/mozilla/readability":
-  version "0.3.0"
-  resolved "https://github.com/mozilla/readability#d5eea06a0095b3138dbd1f6233f656d690200509"
+  version "0.2.0"
+  resolved "https://github.com/mozilla/readability#2982216913af2c66b0690e88606b03116553ad92"
 
 request-promise-core@1.1.2:
   version "1.1.2"