search

korpling / ANNIS

ANNIS is an open source, versatile web browser-based search and visualization architecture for complex multilevel linguistic corpora with diverse types of annotation.
http://corpus-tools.org/annis/
Apache License 2.0
69 stars 25 forks source link

Corpora with version metadatum create duplicate resolver entries on import with overwrite #549

Open amir-zeldes opened 7 years ago

amir-zeldes commented 7 years ago

What is the used ANNIS version? 3.5.0-SNAPSHOT (rev. 39d5ba46d0, built 2016-12-09 10:30:36)

What browser and operating system did you use? Chrome, Win 7 64

What steps will reproduce the problem?

  1. Import GUM
  2. Increment the version datum and reimport + overwrite
  3. Open any example query

What is the expected result? Normal visualizers as defined by resolver_vis_map

What happens instead? All visualizers are duplicated (again and again, for each import+overwrite)

Please provide any additional information below. Also please add a screenshot if possible. It seems on import corpus metadata containing the string 'version' is recognized and the version field in corpus.annis is filled. Once a new version of the corpus is imported, the old version is overwritten, but old entries in resolver_vis_map are not deleted. Instead, the new entries are appended, creating duplicate visualizations.

amir-zeldes commented 7 years ago

This seems to persist even if I delete and then import the corpus fresh.

rainsfordtm commented 3 years ago

I've just encountered the same issue (v. 3.6.0). I confirm it is caused by a "version" metadatum in the corpus. and fixed by removing it. Also the most serious issue I've encountered working with ANNIS, since it corrupts the resolver_vis_map table in the database and the only fixes I found were (i) reset the whole database or (ii) never use the corpus name again, neither of which are convenient workarounds!