Skip to content

Latest commit

 

History

History
140 lines (115 loc) · 6.83 KB

README.md

File metadata and controls

140 lines (115 loc) · 6.83 KB
Raw

Brevity

Build Status

Brevity is the soul of tweet

A small utility to shorten posts to a tweet-length summary. Appends an optional permalink or citation. Also supports autolinking.

Brevity checks URLs against the full list of ICANN TLDs, to avoid linking things that look like web addresses but aren't.

Ports

Usage

Shorten

The primary method brevity.shorten takes a plain text string of arbitrary length returns a shortened string that meets twitter's length requirements (accounting for t.co URL shortening).

>>> import brevity
>>> brevity.shorten("This is the text of a fairly long tweet that will need to be shortened before we can post it to twitter. Since it is longer than 280 characters, it will also include an ellipsis and link to the original note. 123567890 123567890 123567890 123567890 123567890 123567890 123567890 123567890", permalink="http://example.com/2015/03/fairly-long-note")
'This is the text of a fairly long tweet that will need to be shortened before we can post it to twitter. Since it is longer than 280 characters, it will also include an ellipsis and link to the original note. 123567890 123567890 123567890 123567890… http://example.com/2015/03/fairly-long-note'

The permalink, permashortlink, and permashortcitation parameters are all optional and all have slightly different behavior. Permalinks will only be added if the main text needs to be shortened, with the intention that followers can click the link for the full note contents.

To identify all tweets as POSSE copies, you may additionally provide a permashortlink or permashortcitation. If a note is short enough to post to twitter without truncation, the PSL/PSC will be appended to the note text in parentheses.

>>> brevity.shorten("This note is pithy and to the point", permalink="http://example.com/2015/03/to-the-point", permashortlink="http://exm.pl/y1x3")
'This note is pithy and to the point (http://exm.pl/y1x3)'
>>> brevity.shorten("This note is pithy and to the point", permalink="http://example.com/2015/03/to-the-point", permashortcitation="exm.pl y1x3")
'This note is pithy and to the point (exm.pl y1x3)'

If you do not have a URL shortener, but still want to tag all tweets with their permalinks, it is perfectly fine to use the same url for your permalink and permashortlink. It will be appended after an ellipsis for long notes, or in parentheses for short ones.

Note that to be used in a permashortcitation, the bare domain must not be autolinked by Twitter (Otherwise, what should be 5-6 characters will count for 22). This typically means it cannot be a .com, .net, or .org.

Setting the optional parameter format='article' implies that text is the title of a longer article (that can be found at permalink). The composed text will be Article Title: permalink and the permalink will be included regardless of the length of the title. The values of permashortlink and permashortcitation are ignored.

Autolink

The method brevity.autolink, based heavily on CASSIS auto_link, takes a text string, that may contain some HTML, and surrounds web addresses with well-formed <a> tags.

>>> import brevity
>>> brevity.autolink('this links to nebenan.hamburg')
'this links to <a class="auto-link" href="http://nebenan.hamburg">nebenan.hamburg</a>'

Like the CASSIS method it is based on, autolink is idempotent -- applying it to its own output will not change the result. In practice, this means <a> tags in existing HTML will not be affected.

Acknowledgments

Brevity's URL-recognition is based very heavily on Tantek Çelik's excellent CASSIS.

Changes

  • 0.2.18 - unreleased: shorten: add new ellipsis and punctuation kwargs, support link_length=None to mean links count as normal characters.
  • 0.2.17 - 2018-03-15: Twitter's API no longer has a special carve-out for ccTLDs; they count as 23 characters like every other URL. Simplified the logic and updated the tests to match
  • 0.2.16 - Fix character-counting bug that caused index out of range error.
  • 0.2.15 - Implement Twitter's new weighted character counting introduced on 2017-11-07.
  • 0.2.14 - 2017-04-23: Fix crash when given very long tokens
  • 0.2.13 - 2017-04-23: Account for leading special characters like /, @, and $
  • 0.2.12 - 2017-01-03: Update list of TLDs
  • 0.2.11 - 2016-09-25: Remove special formatting for media attachments since Twitter no longer counts them against us!
  • 0.2.10 - 2016-09-05: Rearrange link regex so that TLDs are validated by the regex instead of after the fact. Makes the shortener behave more like Twitter's.
  • 0.2.9 - 2016-05-26: Bugfix: when trimming trailing characters from a URL, fixes an issue where characters after the first would be included several times.
  • 0.2.8 - 2016-04-19: Add support for article+media where text represents the title of an article and includes an image.
  • 0.2.7 - 2016-02-14: Bugfix: make sure check for ccTLDs is case-insensitive, so that we know Twitter won't autolink e.g., Gogs.io
  • 0.2.6 - 2016-02-01: Change shorten() parameter format_as_title to format, and allow different format styles, including note with media (to account for extra characters used by Twitter for an attachment)
  • 0.2.5 - 2016-01-28: Added ports to to php and js
  • 0.2.4 - 2016-01-28: Changed license from BSD to CC0. Extract test cases into reusable tests.json file. Fix bug where permalink citations were counted as text rather than links.
  • 0.2.3 - 2015-11-19: when truncating to the nearest word, strip trailing delimiters, so there's no punctuation before the ellipsis
  • 0.2.2 - 2015-11-19: add format_as_title parameter to shorten to support formatting article titles instead of note posts
  • 0.2.1 - 2015-10-25: all links default to 23 characters now that Twitter serves all t.co links over https
  • 0.2.0 - 2015-09-20: added autolink function
  • 0.1.6 - 2015-06-05: move data text file into brevity.py for easier distribution/reuse.
  • 0.1.5 - 2015-06-05: match all TLDs recognized by IANA; ignore things that look like domains but aren't.
  • 0.1.4 - 2015-04-22: match URLs that include exclamation points.
  • 0.1.3 - 2015-03-29: improved description in pypi.
  • 0.1.0 - 2015-02-15: initial check-in.