Implementation:Infiniflow Ragflow KnowledgebaseService Create With Name

Knowledge Sources	RAGFlow
Domains	RAG, Knowledge_Management
Last Updated	2026-02-12 06:00 GMT

Overview

Concrete tool for creating knowledge base records provided by the RAGFlow KnowledgebaseService.

Description

The KnowledgebaseService.create_with_name class method encapsulates the full creation logic for a knowledge base (dataset). It validates the name, deduplicates within the tenant scope, verifies tenant existence, builds the payload with defaults, and persists via peewee ORM. This method is the canonical entry point used by both the REST API endpoint (POST /v1/kb/create) and internal callers.

Usage

Import and call this method when programmatically creating a new knowledge base. The REST endpoint at api/apps/kb_app.py delegates to this method after authentication.

Code Reference

Source Location

• Repository: ragflow
• File: api/db/services/knowledgebase_service.py
• Lines: L376-423

Signature

class KnowledgebaseService(CommonService):
    model = Knowledgebase

    @classmethod
    @DB.connection_context()
    def create_with_name(
        cls,
        *,
        name: str,
        tenant_id: str,
        parser_id: str | None = None,
        **kwargs
    ) -> tuple[bool, dict | str]:
        """Create a dataset (knowledgebase) by name with kb_app defaults.

        Args:
            name: Dataset name (required, validated for length and emptiness).
            tenant_id: Owner tenant ID.
            parser_id: Parser type (defaults to "naive"). Options: "naive", "paper",
                       "book", "laws", "presentation", "table", "qa", "picture",
                       "one", "audio", "email", "tag".
            **kwargs: Optional fields - description, language, permission, avatar,
                      parser_config.

        Returns:
            (ok: bool, model_or_msg): On success (True, Knowledgebase instance);
                                       on failure (False, error_message).
        """

Import

from api.db.services.knowledgebase_service import KnowledgebaseService

I/O Contract

Inputs

Name	Type	Required	Description
name	str	Yes	Dataset name (max bytes: DATASET_NAME_LIMIT)
tenant_id	str	Yes	Owner tenant UUID
parser_id	str or None	No	Parser type (default: "naive")
description	str	No	Human-readable description
language	str	No	Document language hint
permission	str	No	Access permission level
avatar	str	No	Avatar image path
parser_config	dict	No	Parser-specific configuration

Outputs

Name	Type	Description
ok	bool	True on success, False on validation failure
result	Knowledgebase or str	Knowledgebase model instance on success, error message on failure

Usage Examples

Create a Basic Knowledge Base

from api.db.services.knowledgebase_service import KnowledgebaseService

# Create with defaults (naive parser)
ok, result = KnowledgebaseService.create_with_name(
    name="Technical Documentation",
    tenant_id="user-uuid-123"
)

if ok:
    kb = result
    print(f"Created KB: {kb.id}")
else:
    print(f"Error: {result}")

Create with Specific Parser

ok, result = KnowledgebaseService.create_with_name(
    name="Legal Contracts",
    tenant_id="user-uuid-123",
    parser_id="laws",
    description="Collection of legal documents",
    language="English"
)

Related Pages

Implements Principle

• Principle:Infiniflow_Ragflow_Knowledge_Base_Creation