MCPcopy Create free account
hub / github.com/VectifyAI/OpenKB / run_knowledge_lint

Function run_knowledge_lint

openkb/agent/linter.py:92–119  ·  view source on GitHub ↗

Run the semantic knowledge lint agent against the wiki. Args: kb_dir: Root of the knowledge base. model: LLM model name. Returns: The agent's lint report as a Markdown string.

(kb_dir: Path, model: str)

Source from the content-addressed store, hash-verified

90
91
92async def run_knowledge_lint(kb_dir: Path, model: str) -> str:
93 """Run the semantic knowledge lint agent against the wiki.
94
95 Args:
96 kb_dir: Root of the knowledge base.
97 model: LLM model name.
98
99 Returns:
100 The agent's lint report as a Markdown string.
101 """
102 from openkb.config import load_config
103
104 openkb_dir = kb_dir / ".openkb"
105 config = load_config(openkb_dir / "config.yaml")
106 language: str = config.get("language", "en")
107
108 wiki_root = str(kb_dir / "wiki")
109 agent = build_lint_agent(wiki_root, model, language=language)
110
111 prompt = (
112 "Please audit this knowledge base wiki for semantic quality issues: "
113 "contradictions, gaps, staleness, redundancy, and missing concept and "
114 "entity pages. Start with index.md, then read summaries, concepts, and "
115 "entities as needed. Produce a structured Markdown report."
116 )
117
118 result = await Runner.run(agent, prompt, max_turns=MAX_TURNS)
119 return result.final_output or "Knowledge lint completed. No output produced."

Callers 4

run_lintFunction · 0.90

Calls 4

load_configFunction · 0.90
build_lint_agentFunction · 0.85
runMethod · 0.80
getMethod · 0.45