Most people seem to have forgotten about email, unfortunately. It still is a great PIM (personal information management) tool.
email is a nice way to keep track of your reading queues, whether daily newspapers or monthly magazines or yearly journals or your favorite news channels or youtube feeds etc. With email you get a familiar interface, ability to (auto) classify contents into different folders as per your own taste and priorities, share them with someone, thread them, keep status as read or unread, trim your reading queue by deleting unwanted or read articles etc.
If you are like me and use a text based client like mutt, even nicer if you use procmail to classify your email (you can, of course, use 'filters' facility of your mail provider also) above conveniences are all the more empowering. But it will be a digression to get into the details of that.
feeds2mail (f2m for short) is inspired by rss2email, but with a more generalized notion of feeds, that includes RSS but adds more to it.
If you use email classification wisely, f2m can become a highly customizable feed aggregator, with highly organized contents. For example, different feeds for the same topic (say, feed for your city in different newspapers) can be brought together into one email folder. For this, there is a builtin support for generating + suffix in the email address to which the articles are mailed, which makes it easy to classify emails for procmail or any other filters.
Following are the feed types supprted by f2m. More detailed usage description of each type follows in later subsections.
Basic good old RSS feed is supported by f2m
Most publishers have a page called 'RSS Feeds' on which they list several of their feeds dedicated to different topics. metarss feed type in f2m is like a feed of feeds, where you specify the URL of feeds page and regular expressions to pick individual feeds by their topics.
A very handy feature you get with this is, the state of the feed is tracked at the metarss level. It helps eliminate duplicate items from various feeds, or items that appear today in one feed and tomorrow in another.
youtube itself provides an rss interface to its channels and playlists. In f2m by merely specifying the channel or playlist id, the feed can be activated.
Better organized publishers create specific channels or playlists for their popular shows. Sometimes they aren't easy to find. Do use the FILTERS widget on youtube search and restrict your search to channels or playlists (in two separate search queries). If you are lucky you might get a precise channel or playlist for the contents you want to track.
But if you do not find one. You can subscribe to the default channel of the publisher, which will typically have all the content it ever publishes - only a small fraction of which you are interested in. You can achieve the filtering by specifying txtregex property to filter a busy feed to retain only the articles of your interest. (All properties are described in the later sections below.)
The youtube feed items carry the video duration in the subject line (using yt-dlp package) of the email. This may help in deciding whether to view the item now or at leisure later.
Since youtube access mandated a login, for the duration functionality to work you must have logged in your default profile. We use 'cookiesfrombrowser' option to extract the duration information.
This feed type is to select items from an arbitrary URL use regular expression on its URL and/or hypertext using properties urlregex and txtregex respectively.
Use this only as last resort and form the regular expressions carefully.
urlgroup is to url what is metarss to rss, just grouping of URLs in one feed so that duplicates from various subfeeds can be eliminated. The 'url' property of the urlgroup has no semantics as that of the contained url list will get actually used to fetch articles. As a thumb rule just use the home page of the publication.
Many newspapers have stopped supporting RSS feeds. Either url or urlgroup type may help in these cases.
- python3 is required
Following python packages are required
- py-filelock
- py-feedparser
- yt-dlp
Feed settings reside in $HOME/.f2mrc.json. A sample settings file sample.f2mrc.json is included in the package.
The root level is a dictionary with following properties:
defaults
feeds
defaults is a dictionary where default properties of any feed can be specified. Most typically you might want to specify the following:
active : true | false, default is true, if false the feed is not processed.
Typically when you add a new feed, set this to false and set the same
property to true in the individual feed to run just the new feed.
dryrun : true | false, if true, only the first feed in the feedlist will be
processed, but no email will be generated, feeds' state will not be
updated. Set it to true if you want to test out a new feed. Remember to
place the feed under test at the first position in the lsist.
mailto : default email id to which generated emails should be sent
feeds is a list of dictionaries where each dictionary specifies properties of 1 feed. Following properties can be specified:
active : see description under defaults
channelid|playlistid : applicable to youtube type, to specify the
respective id
feedregex: In metarss, this regex helps identify the links to individual
rss fields, which are further chosen by applying txtregex of their title
texts.
mailto : Typically specified under defaults, but can be overridden for
individual feeds. Most typically you might not want to specify this for
feeds, you might use mailtosuf instead.
mailtosuf: A +suffix attached to the mailto address. For example, if your
mailto address is myname@localhost and mailtosuf is set to mytag, then
email will go to myname+mytag@localhost. The mail filter rules can make use
of this to classify the email into different folders.
remarks: Play no role in the software. This is for your own documentation
purpose.
subfeeds: In metarss, this a list of dictionaries that specify individual
rss feed's properties. Note that, as this is a meta feed, url property is
not required to be specified in these subfeeds.
subpref: A subject prefix, that will be prefixed to the mail subject, if
specified. It is put in square brackets. Typically you might want to
indicate the source of the article by doing so. You might use a common tag
for articles from multiple sources. Then in your mailbox this helps spot
the publiher of it. It is recommended to keep its length small, at the
most 4 may be so that it doesn't eat up into the title's space. For
example, subpref FP will be prefixed as [FP] to the subject line and it
might help you spot that the article is from firstpost when looking at the
mailbox. (You of course know it when you open the URL, but many times you
might not even open the email if the subject doesn't interest you.)
txtregex : For all feed types and subfeeds in metarss, this regular
expression, if specified, is applied to select the item from the feed. For
metarsss, it is applied to select rss feeds from a feeds page.
urlregex : For feed ot type "url" this regex selects the URLs on the page.
typ : youtube|rss|metarss|url|urlgroup, as explained above
Make sure that you have $HOME/.f2mrc.json in place, as described above.
You might want to create a script such as with name f2m with contents like below in a directory in your PATH, typically in $HOME/bin or /usr/local/bin
exec <f2m_install_dir>/feeds2mail/f2m.py
Typing just f2m will process the feeds and generate email as per your specification. As of now there are no command line arguments.
Most typically way to run f2m is to place it in cron jobs. The cron daemon would be typically set to email you the stdout contents.
It's typical to run it once or twice a day. Be aware that some feeds change annoyingly fast, for example - if it is of a youtube feed of a news channel. Yo might want to adjust your cron frequency accordingly.
On stdout, f2m prints the feed url and number of new items for which it has generated email. Under this line with an indent, it may report one or more of the following diagnostics if applicable:
duplicates: Count of duplicate items (appearing more than once to be precise) in metarss feeds. This typically happens with newspaper feeds. For example some items may appear under a 'city' feed as well as 'nation' feed. f2m sends them only once and attributes them to the first feed specification they match with in the subfeeds list.
feed too old: Feeds with latest item being 365 days or more older than today. May be this feed has stopped adding articles and you might want to remove it from your feeds list.
feed blank: This happens to some feeds over time. Instead of retaining the last updated state, publisher's system might keep them blank, or the URL might just get invalidated. But note that this message would occur even if for some transient technical reasons the feed could not be accessed. So do verify things before removing this feed from your feed list.
f2m uses py-filelock to lock the feeds' state to avoid the file getting corrupted due to concurrent runs. Check if another process is running if you get an error suggesting so.
Usually you should not require doing this. f2m stores lists of 'seen' articles in $HOME/.f2mdb/feedstat.json, to filter the already emailed articles from the mails. Preferably do not tinker with this file manually unless you know what you are doing.
-
Multithreading to speed up the run
-
Make a proper python package, include pre-requisites, include the launch script
-
Purge history of removed feeds