(runs: EvaluationRunExport[])
| 96 | } |
| 97 | |
| 98 | async function generateAnalysis(runs: EvaluationRunExport[]): Promise<string> { |
| 99 | const context = buildDynamicContext(runs); |
| 100 | |
| 101 | const { text } = await generateText({ |
| 102 | model: getZenLanguageModel(analyzerModelId), |
| 103 | system: AGENT_ANALYSIS_PROMPT, |
| 104 | prompt: context, |
| 105 | temperature: 0.3, |
| 106 | }); |
| 107 | return text.trim(); |
| 108 | } |
| 109 | |
| 110 | function usage(): void { |
| 111 | console.error("Usage: bun run scripts/analysis.ts <benchmark-file.json>"); |
no test coverage detected