I’m running version 0.23.0 with Docker Compose and decided to try out the Typesense integration. After performing a full reindex using the vikunja index CLI command, the logs showed no errors, so I began testing.
In the Kanban view, the fast search can no longer find tasks by partial title names. While I can still search for task titles, I now need to use filtering options, such as filtering for “Test task” to find tasks with partial names. Additionally, I can search for comments on the Kanban board using fuzzy matching, like “hello123”, and the results are correct.
Another issue I’m encountering is that I have over 600 tasks. When I open the List, Gantt, and Table views, I receive a UI error, and the Docker Compose logs display the following:
vikunja | 2024-07-01T09:47:34.569848488Z: WEB ▶ 192.168.96.1 GET 304 /assets/ProjectList-qm15u6OH.css 4.132968ms - Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36
vikunja | 2024-07-01T09:47:34.588363854Z: ERROR ▶ handler/ReadAllWeb 266 status: 422 response: {"message": "Only upto 250 hits can be fetched per page."}
vikunja | 2024-07-01T09:47:34.5884174Z: WEB ▶ 192.168.96.1 GET 500 /api/v1/projects/4/tasks?sort_by[]=position&order_by[]=asc&filter_by[]=done&filter_value[]=false&filter_comparator[]=equals&filter_concat=and&page=1 2.364987ms - Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36
Is there a configuration option to prevent this error in these views? I believe this is due to a Typesense limit per page that cannot be changed. I’m using Typesense typesense/typesense:26.0.
Maybe I have to reindex again? Because now nothing shows on the Kaban view either
Tried to reindex using the unstable version but seems that I can’t access the container shell
$ docker exec -it vikunja sh
OCI runtime exec failed: exec failed: unable to start container process: exec: "sh": executable file not found in $PATH: unknown
If I go back to the stable version I can access the container shell but if I try to reindex I now get
/app/vikunja # /app/vikunja/vikunja index
2024-07-01T13:24:44.131056627Z: INFO ▶ config/InitConfig 001 No config file found, using default or config from environment variables.
2024-07-01T13:24:44.166884686Z: INFO ▶ migration/Migrate 05c Ran all migrations successfully.
2024-07-01T13:24:44.167161143Z: INFO ▶ cmd/func4 05d Indexing… This may take a while.
2024-07-01T13:24:44.183861534Z: CRITICAL ▶ cmd/func4 061 Could not reindex all tasks into Typesense: could not get all tasks: Error 1054 (42S22): Unknown column 'bucket_id' in 'field list'
Is there a change in the database structure beween the latest stable and the latest unstable?
Just moved to the unstable version and did a docker exec vikunja /app/vikunja/vikunja index
2024-07-01T15:02:22.161067276Z: INFO ▶ config/InitConfig 001 No config file found, using default or config from environment variables.
2024-07-01T15:02:22.218714406Z: INFO ▶ migration/Migrate 069 Ran all migrations successfully.
2024-07-01T15:02:22.238496018Z: INFO ▶ cmd/func4 06a Indexing all tasks… This may take a while.
2024-07-01T15:02:25.350786948Z: INFO ▶ cmd/func4 ebc Done!
All good with the Kanban view using the unstable version
Yes I did to VIKUNJA_SERVICE_MAXITEMSPERPAGE: 1000.
To avoid using pagination on the API calls.
I am using the API to import data from Wekan to Vikunja and to speed coding a bit I changed the maxitemsperpage. It’s useful for a scenario where I am importing the comments from one side to the other.
I’ve pushed a change in 2982acc0ec to restrict the max length to 250 when using Typesense. That’s a limitation of Typesense, nothing we can do about it. With the change, the error should go away.
