Skip to content

m5_004

Latest
Compare
Choose a tag to compare
@kerchner kerchner released this 29 Oct 15:30
· 5 commits to master since this release
  • Docker support (see https://github.com/gwu-libraries/social-feed-manager/blob/m5_004/docker/README.md)
  • The export_xls and export_csv management commands have been removed and replaced by a single export command with a --format option, which can be xls, csv, or json.
  • Extract files now in XLSX format (was XLS). The XLSX format removes the limitation of 65,536 rows. SFM m5_004 substituted the openpyxl library for xlwt.
  • SFM now uses tweepy v3.4.0. This seems to eliminate the problem observed in m5_003, which used tweepy 3.2.0, where filterstream jobs stopped writing and/or slowed down the server.
  • Several documentation improvements, mostly around installation and around supervisor/filterstreams setup.
  • UI cleanup: Improved consistency of date-time format rendering.
  • Enhanced validation logic around checking People values when adding or updating filterstreams: If one or more People account names wasn't found when checking against Twitter, the TwitterFilter does save, but presents a warning listing the invalid accounts.
  • Now allows deletion of filterstreams even if Supervisor isn't running.
  • Now allows creation/update of filterstreams even if Supervisor isn't running (however, these won't automatically reflect updates when Supervisor is restarted (See #376).
  • Updated Apache2 configuration file to note recommended changes for deployment on Ubuntu 14 / Apache2 v2.4+
  • Enhanced fetch_urls to more gracefully handle (skip) truncated URLs that may appear in retweets with an ellipsis, and to pull URLs from 'media' if possible.
  • And several other minor enhancements.

See the complete list of changes for milestone m5_004 in github (https://github.com/gwu-libraries/social-feed-manager/issues?q=is%3Aissue+is%3Aclosed+milestone%3Am5_004) as well as the code changes from m5_003 to m5_004 (m5_003...m5_004)

Upgrading from m5_003 to m5_004

  • The requirements.txt file has changed, so Python library dependencies must be updated within your virtualenv. To update dependencies:
% source ENV/bin/activate
% pip install -r requirements.txt

Contributors

  • Dan Chudnov, Dan Kerchner, Laura Wrubel, Justin Littman, Ankushi Sharma, and Rajat Vij at The George Washington University Libraries
  • Ed Summers at Maryland Institute for the Humanities (MITH)
  • Martin Klein at UCLA Research Library