[Bug]: HTTP API Lacks Type and Valid Value Checks for all parameters under parser_config During Dataset Creation

[asiroliu]

Is there an existing issue for the same bug?

• I have checked the existing issues.

RAGFlow workspace code commit ID

x

RAGFlow image version

cb1febb7d8ae （infiniflow/ragflow:v0.17.0-slim）

Other environment information

Actual behavior

The API accepts and persists invalid values for parser_config.chunk_token_count during dataset creation, including negative numbers, non-integer types, and out-of-range values.

Expected behavior

The API should reject requests with:

Negative integers (e.g., -1)
Non-integer types (e.g., 3.14, "1024")
Values exceeding reasonable limits (e.g., >2048)

Steps to reproduce

1. Send a POST request with invalid chunk_token_count values (e.g., -1, 3.14, "1024")

response = requests.post(
    f'http://127.0.0.1:9380/api/v1/datasets',
    json={
        "name": "test",
        "chunk_method": "naive",
        "parser_config": {"chunk_token_count": -1}  # Test with -1/3.14/"1024"
    }
)

2. Observe the successful response:

{
    "code": 0,
    "data": {
        "parser_config": {"chunk_token_count": -1},
        "name": "test",
        // ...other fields
    }
}

Additional information

No response

[asiroliu]

Currently, all parameters under parser_config lack type checking and validation of valid values.

[asiroliu]

1. When chunk_token_num, task_page_size, auto_keywords, or auto_questions are set to ​floating-point numbers (e.g., 3.14), no error is thrown.
2. When these fields are set to ​string types (e.g., "1234"), the error message is unclear and appears as:

{'code': 100, 'data': None, 'message': 'TypeError("\'<=\' not supported between instances of \'int\' and \'str\'")'}