Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
from controllers.openapi._models import KNOWLEDGE_TOP_K
from controllers.openapi.knowledge_bases import knowledge_base_row
def _record(**extra: object) -> dict[str, object]:
return {
"id": "ds-1",
"name": "Docs",
"description": None,
"provider": "vendor",
"indexing_technique": "high_quality",
"document_count": 3,
"embedding_available": True,
**extra,
}
def test_row_carries_a_retrieval_fragment() -> None:
row = knowledge_base_row(_record())
assert row.node_data.dataset_ids == ["ds-1"]
assert row.node_data.retrieval_mode == "multiple"
assert row.node_data.multiple_retrieval_config.top_k == KNOWLEDGE_TOP_K
assert row.node_data.multiple_retrieval_config.reranking_enable is False
def test_unavailable_embedding_is_reported() -> None:
assert knowledge_base_row(_record(embedding_available=False)).usable is False
assert knowledge_base_row(_record(embedding_available=None, provider="external")).usable is True