v1.39 Bugfixes and maintenance
We also recommend reading the instructions above if you are running 4CAT via Docker and have in any way deviated from the default set-up.
Otherwise, you can upgrade 4CAT via the 'Restart or upgrade' button in the Control Panel. This release of 4CAT incorporates the following fixes and improvements:
- Fix several issues that could occur when trying to import CSV files (#404, 9963cd8, 02acd86, 0049de6, 151d498, cdfe75c)
- Fix an issue where the canonical host name for a Docker-based 4CAT front-end would not always be set correctly (#395, #403, 0d1dc05)
- Fix an issue where tags would linger in the database after they were no longer associated with any user, or would be stored in the wrong order of preference (c38a215, 79f58bf)
- Fix an issue where uploads from Zeeschuimer would fail for data sources whose ID starts with a number (e.g.
9gag
) - Fix an issue where making a user would crash the front-end if a user with the same e-mail already exists (14c1f9c)
- Fix a crash in the 'Post/topic matrix' processor (be6ea8c)
- Fix an issue where the user-specific setting for the max amount of downloaded TikTok videos would be ignored (df2462f)
- Fix a crash when duplicating a dataset and copying the dataset's owners (a3fdbad)
- Fix an issue where the TikTok metadata fetcher would fail with a 'proxy unavailable' error when trying to run multiple TikTok processors at the same time (04faf23)
- Fix an issue with the Telegram image downloader (23edf44)
- Fix an issue where network processors could crash with a divide by zero error if run on an empty dataset (e10cb2e)
- Add a clearer error message when trying to merge datasets that are not CSV or NDJSON (c5fbe02)
- Add an explicit edge weight to generated networks that is properly recognised by Gephi (c415870)
- Add the option to only capture the first frame of a video when extracting video frames (b3981c3)
- Add various image processors (e.g. image wall) as supported for video frame datasets (588290a)
- Add improved compatibility between video hash processor and image classification processor so that video frames can be visually categorised (a4e6904)
- Add data source options to explicitly define the Tumblr API key to use if no 4CAT-wide keys have been configured (fdbdca9)
- Add a 4CAT setting that determines which proxy headers will be taken into account for URLs generated in the front-end (5e47dac)
- Update the TikTok import processor to cope with the new data formats provided by Zeeschuimer (f24828b)
- Update code that determines place of a dataset in the queue to be more efficient (239726b)
- Update dataset importer to stream files, which should prevent issues with very large data files (7a1c4b9)
- Update the type of the
jobid
column of the jobs table to BIGINT to avoid issues with long-running 4CAT servers (3414a96) - Update jobs table to no longer have a useless 'status' column (9f493b2)
- Update processor presets so they do not linger in an unfinished state if one of their components crashes, finish with an error instead (b815c54, ddf9aab)
- Update various image dataset processors to produce more compatible .metadata.json files (87ec4d0)
- Updated version for the videohash library dependency (5f5e10f)