Generate offline recommendation Task loops indefinitely

**Gorse version**
zhenghaoz/gorse-in-one:latest

**Describe the bug**
I am trying Gorse out with the provided github interactions dataset, everything runs fine on docker but I have noticed that the 'Generate offline recommendation' task keeps restarting indefinitely. I do understand that check_recommend_period and refresh_recommend_period variables dictate if offline recommendations will be recalculated (https://gorse.io/docs/master/concepts/how-it-works.html#worker-offline-recommendation) .
Currently these variables are set as:
- check_recommend_period = "1m"
- refresh_recommend_period = "24h"

I have confirmed that the Generate offline recommendation Task loops even if I don't add any additional users, items or feedbacks.

**To Reproduce**
- Ran docker compose locally
- successfully imported the demo data into mysql container
- Restarted all containers
- Let all tasks complete successfully on the "Tasks" page

**Expected behavior**
Offline recommendations not being calculated over and over again if there weren't any new users, items or feedbacks added to the system .

**Additional context**
Maybe an issue related/similar to this one ( https://github.com/gorse-io/gorse/issues/789 )
I am attaching my docker logs to the issue in case those can help.
[dockerLogGorse.txt](https://github.com/user-attachments/files/18351462/dockerLogGorse.txt)
[dockerLogRedis.txt](https://github.com/user-attachments/files/18351469/dockerLogRedis.txt)
[dockerLogMySql.txt](https://github.com/user-attachments/files/18351478/dockerLogMySql.txt)

Here is the config I am currently using, along with the config.toml file:
[config.toml.txt](https://github.com/user-attachments/files/18351523/config.toml.txt)
[docker-compose.yaml.txt](https://github.com/user-attachments/files/18351524/docker-compose.yaml.txt)

Please let me know if this is the expected behavior or if something needs to be adjusted on my end, my team is looking to use Gorse more on the daily to generate predictions on labeled data, and our main concern is that calculating offline recommendations every minute will use too much of our hardware resources as the underlaying data grows in size.

Start by reviewing the attached config.toml and docker-compose.yaml files alongside dockerLogGorse.txt, dockerLogRedis.txt, and dockerLogMySql.txt. Compare the configured check_recommend_period and refresh_recommend_period with the offline-recommendation behavior described in the documentation and issue #789. Done means identifying why the task repeats without new data and confirming that idle data no longer triggers unnecessary recalculation.

Written by the indexing model from the issue text.