bugfix when document not have metadata ,then cant retrieved

[likenamehaojie]

What problem does this PR solve?

Fix bug: when the document doesn't have meta data. It can't be retrieved.

Type of change

• Bug Fix (non-breaking change which fixes an issue)

[KevinHuSh]

Appreciations!
Could you paste a screenshot or something to illustrate why people need this feature?

[likenamehaojie]

老用户如果使用dify外挂知识库后，直接升级到20.4版本后，会什么都检索不到@KevinHuSh

[likenamehaojie]

这是一个dify的对接SDK的修复
之前我提交了一个片断，是dify sdk不能接收元数据请求
合并之后
如果dify外挂ragflow会出现拿不到那些没有设置元数据的文档，当然前提是dify在外挂时可以默认不传递元数据过滤，只不过ragflow中提供的（dify sdk）已经主动过滤了 没有 设置元数据 的文档（这是之前我提交代码的一个bug）这个pr已经修复，主要就是解决这个问题@KevinHuSh

[KevinHuSh]

We've considered the meta data as a filter rather than a ranker(matched docs rank ahead of unmatched docs).
Could you elaberate on the scenario you're encountered? Share with us please.

[likenamehaojie]

Background:
This is to address different permission scopes for multiple documents in a knowledge base within Dify. Based on different logged-in users, restrict the documents that different users can retrieve from the same knowledge base.

In v0.20.3, Ragflow added document-level metadata filtering. For this, we modified the SDKs of both Dify and Ragflow.
In the previous version, we resolved this issue but simultaneously introduced a new bug. This bug causes Dify users to retrieve no content when using an external Ragflow without enabling metadata (specifically, when metadata is not enabled for documents in Ragflow).@KevinHuSh

[KevinHuSh]

Appreciations!
By our evaluations, this feature is not going to be merged into main branch.