How do I upload a file to the knowledge base in parent-child structure mode through the API interface of DIFY?

[VincentWallace8012]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template :) and fill in all the required fields.

Provide a description of requested docs changes

curl --location --request POST 'http://127.0.0.1/v1/datasets/{dataset_id}/document/create-by-file'
--header 'Authorization: Bearer {api_key}'
--form 'data="{"indexing_technique":"high_quality","process_rule":{"rules":{"pre_processing_rules":[{"id":"remove_extra_spaces","enabled":true},{"id":"remove_urls_emails","enabled":true}],"segmentation":{"separator":"###","max_tokens":500}},"mode":"custom"}}";type=text/plain'
--form 'file=@"/path/to/file"'

The upload file interface of dify is like this, please adjust which parameter allows me to initialize the document into the knowledge base in parent-child mode？

[dosubot[bot]]

Hello @VincentWallace8012! I'm here to help you with any bugs, questions, or contributions while you wait for a human maintainer. Let me know how I can assist you today!

To upload a file to the knowledge base in parent-child structure mode through the Dify API, you need to adjust the process_rule.mode parameter to "hierarchical". Additionally, ensure that the doc_form is set to "hierarchical_model" and configure the rules section appropriately. This includes setting the parent_mode, segmentation, and subchunk_segmentation settings. Here's an example of how you might adjust your API request:

curl --location --request POST 'http://127.0.0.1/v1/datasets/{dataset_id}/document/create-by-file' \
--header 'Authorization: Bearer {api_key}' \
--form 'data="{
  \"indexing_technique\":\"high_quality\",
  \"process_rule\":{
    \"rules\":{
      \"pre_processing_rules\":[
        {\"id\":\"remove_extra_spaces\",\"enabled\":true},
        {\"id\":\"remove_urls_emails\",\"enabled\":true}
      ],
      \"segmentation\":{
        \"separator\":\"###\",
        \"max_tokens\":500
      },
      \"subchunk_segmentation\":{
        \"separator\":\"\n\",
        \"max_tokens\":200
      },
      \"parent_mode\":\"paragraph\"
    },
    \"mode\":\"hierarchical\"
  },
  \"doc_form\":\"hierarchical_model\"
}";type=text/plain' \
--form 'file=@"/path/to/file"'

This configuration sets the document to be processed in hierarchical mode, using paragraphs as parent segments and sentences as child segments ^[1][2].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Bug Report | Other}

[VincentWallace8012]

Thank you very much, I'm going to try it with this method