initial commit

.gitignore

+rssemailconfig.py
+*.pyc
+*.pyo

rss-email.py

+#! /usr/bin/env python3
+
+import os
+import sys
+import subprocess
+import feedparser
+import pickle
+from smtplib import SMTP
+from email.mime.multipart import MIMEMultipart
+from email.mime.text import MIMEText
+from bs4 import BeautifulSoup
+
+feed            = {'file': 'feed.pickle'}
+mail            = {'from': 'Murukesh Mohanan <murukesh@cse.iitb.ac.in>'}
+smtp            = {}
+old_entries     = set()
+
+from rssemailconfig import *
+
+def get_new_posts ():
+    current_entries = feedparser.parse(subprocess.Popen(['curl', '-u', ':'.join(feed['creds']), feed['url']], stdout=subprocess.PIPE).communicate()[0]).entries
+
+    post_ids = [post.id for post in current_entries]
+
+    try:
+        old_entries = pickle.load (open(feed['file'], 'br'))
+    except (EOFError, FileNotFoundError):
+        old_entries = set()
+
+    new_post_ids = set (id for id in post_ids if id not in old_entries)
+    new_posts = [post for post in current_entries if post.id in new_post_ids]
+    return new_posts
+
+def convert_post_to_email (post):
+    msg = MIMEMultipart('alternative')
+    msg['Subject'] = mail['prefix'] + post.title
+    msg['From'] = mail['from']
+    msg['To'] = mail['to']
+    part1 = MIMEText(BeautifulSoup(post['summary']).get_text(), 'plain')
+    html_email = post['summary'] + '<br><hr><a target="_blank" href="' + post['id'] + '">' + post['id'] + '</a>'
+    part2 = MIMEText(html_email, 'html')
+    msg.attach(part1)
+    msg.attach(part2)
+    return msg
+
+def mail_feed (feed):
+    smtp = SMTP(smtp_url)
+    smtp.starttls()
+    smtp.login(*smtp_creds)
+    bad_posts = set()
+    for post in feed:
+        msg = convert_post_to_email (post)
+        try:
+            smtp.send_message(msg)
+        except:
+            bad_posts.add(post.id)
+    return bad_posts
+
+def save_sent_posts (feed, bad_posts):
+    old_entries.update(post.id for post in feed if post.id not in bad_posts)
+    with open (feed_file, 'bw') as f:
+        pickle.dump (old_entries, f)
+    return
+
+if __name__ == "__main__":
+    new_posts = get_new_posts ()
+    if len(new_posts) == 0:
+        print ("No new posts.")
+        sys.exit(0)
+    bad_posts =  mail_feed (new_posts)
+    save_sent_posts (new_posts, bad_posts)

rssemailconfig.example.py

+
+feed['url']	= 'http://placements.iitb.ac.in/blog/?feed=rss2'
+feed['name'] = 'IITB Placement Blog'
+feed['creds'] = ['murukesh', 'password']
+feed['file'] = 'feed.pickle'
+
+smtp['url']	= 'smtp-auth.iitb.ac.in'
+smtp['creds'] = ['murukesh', 'password']
+
+mail['to'] = 'murukesh.mohanan@gmail.com'
+