fix: embedding demension change bug

[huangzhuo1949]

Summary

Close #12913

Screenshots

Before	After
...	...

Checklist

Important

Please review the checklist below before submitting your pull request.

• This change requires a documentation update, included: Dify Document
• I understand that this PR may be closed in case there was no previous discussion or issues. (This doesn't apply to typos!)
• I've added a test for each change that was introduced, and I tried as much as possible to make a single atomic change.
• I've updated the documentation accordingly.
• I ran dev/reformat(backend) and cd web && npx lint-staged(frontend) to appease the lint gods

[JohnJyong]

the logic is exist in the deal_dataset_vector_index_task.py, pls descripe the detail of how to replicate the issue @huangzhuo1949

[huangzhuo1949]

the logic is exist in the deal_dataset_vector_index_task.py, pls descripe the detail of how to replicate the issue @huangzhuo1949

replicate step by step
1、upload some file to dataset
2、 change the embedding model of dataset while some file is indexing
3、worker1 has inited vector class by old embedding model because it has started indexing file
4、woker2 inited vector class by new embedding model
5、then the following phenomena occur: expected 1536 dimensions, not 1024

[bowenliang123]

It's risky in huge memory usage to loading all the docs risks for.
Count the docs on db side instead.