I was a happy Google Reader user. Then Google decided to kill it off because it wasn't making them enough cash. From there, I bounced to Feedly, Blogtrottr, rss2email and a bunch of others in-between. Much like Goldilocks, nothing was quite right for me. I boiled my requirements down to a minimal list:
- RSS to email gateway. I need to heavily filter some feeds so why not make use of email filters that are more powerful than most feed-level filtering options?
- SMTP output. I am sick of messing with MTAs to get my local mail forwarded to an inbox I actually read. Actually this subsequently evolved into IMAP output when I realised I could evade some unhelpful non-optional spam filtering by directly copying entries into a mailbox over IMAP.
- Text file configuration that I can check-in to my utils repo. I don't like losing these things when my disk crashes.
- Easily extensible. I had some ideas for RSS-like monitoring of non-RSS sources.
What you see is my attempt to solve this with some scripting. It's a pretty small codebase and (IMHO) pretty easily extensible via the feeders. Be aware it is currently in no way polished and lacks even basic error handling.
- Python ≥ 3.6
- BeautifulSoup
- feedparser
Needtoknow expects two configuration files provided to it as JSON:
conf.json: How to access your mailbox.feeds.json: What feeds to monitor.
By default, it looks for these files in the directory ~/.needtoknow. You can
specify a different directory with the --config command line option.
The mailbox configuration file is expected to contain a dictionary detailing how to connect to your mailbox:
{
"host":"imap.gmail.com",
"login":"[email protected]",
"password":"correct horse battery staple"
}Note that your password is in clear text in this file. You should make efforts to locate this file in a secure area or otherwise prevent access to it by unauthorised users.
Only SSL connections are supported and the port to use defaults to 993. If you need to override the default port, you can add an extra entry to the dictionary:
{
"host":"imap.gmail.com",
"login":"[email protected]",
"password":"correct horse battery staple",
"port":"42"
}By default, the above will result in new items being posted to your IMAP inbox. If you want feed items to be posted to a different mailbox folder, you can specify this as another dictionary entry:
{
"host":"imap.gmail.com",
"login":"[email protected]",
"password":"correct horse battery staple",
"folder":"my-feeds"
}Note that this folder must already exist.
The feeds you want to monitor are provided as a dictionary, wherein each entry represents a single feed. The mandatory elements to define for a feed are its name, the "feeder" to use and its URL:
{
"Hacker News":{
"feeder":"rss",
"url":"https://news.ycombinator.com/rss"
},
"Slashdot":{
"feeder":"rss",
"url":"http://rss.slashdot.org/Slashdot/slashdot"
}
}The feeder defines the type of the feed and how to present its contents. For a list of feeders, look under the directory src/feeders/. Several feeders have their own options that can be tweaked via further dictionary entries. A full explanation of each feeder and its features is beyond the scope of this README and you will have to read their source to understand their capabilities.
Feed items will show up in your mailbox as emails from the name you have given
them in your feeds.json. Perhaps you wish items from a specific feed to
appear as if they were sent by a person. You can achieve this by naming the feed
with a format representing a name and email address:
{
"Paul Graham <[email protected]>":{
"feeder":"rss",
"url":"http://www.aaronsw.com/2002/feeds/pgessays.rss"
}
}The advantage of this is that your mail client recognises this and will let you naturally send a reply based on items from this feed to that person. [1]
This is barely scratching the surface of what is possible with different feeders and their configuration options, so as mentioned above please read their source to learn more.
Once you have needtoknow configured, you probably want to run it on a schedule. You can do this with cron or your favourite scheduling utility. Note that needtoknow outputs diagnostics and errors to stderr, so if you are using cron any errors will be delivered to your local system mailbox. More detailed information is emitted if stderr is a TTY because needtoknow thinks a human is paying attention, so it can be helpful to configure your cron environment with a pseudo TTY if you are debugging something.
Want to modify this code? Fork away. If you have any questions, let me know. If you want a feature, but are too lazy to implement it yourself, ask me when I'm having a good day and I may do it for you :)
All files in this repository are in the public domain. Use them in any way you wish. However, be aware that they come with no warranty. If you reuse this code, I assume you will read and understand it first.
| [1] | If you actually do this with Paul Graham, I highly doubt he will answer your email. |