MCPcopy Create free account
hub / github.com/ScaleML/AgentSPEX / load_benchmark_queries

Function load_benchmark_queries

src/benchmarks/writing_bench/evaluate.py:93–102  ·  view source on GitHub ↗

Load the full benchmark JSONL (index -> {query, checklist, domain1, ...}).

(jsonl_path: Path)

Source from the content-addressed store, hash-verified

91
92
93def load_benchmark_queries(jsonl_path: Path) -> Dict[int, dict]:
94 """Load the full benchmark JSONL (index -> {query, checklist, domain1, ...})."""
95 data = {}
96 with open(jsonl_path, "r", encoding="utf-8") as f:
97 for line in f:
98 line = line.strip()
99 if line:
100 obj = json.loads(line)
101 data[obj["index"]] = obj
102 return data
103
104
105def _evaluate_single(

Callers 1

evaluateFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected