File size: 2,779 Bytes
3c332d2 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 | from researchbench.generation import estimate_generation_model_calls, parse_matched_score, run_generation_record, score_generation
from researchbench.clients import MockClient
def test_parse_matched_score_accepts_zero_to_five() -> None:
assert parse_matched_score("**Matched score starts:** 4 **Matched score ends**") == 4.0
assert parse_matched_score("Reason: ok\nMatched score: 4") == 4.0
assert parse_matched_score("Matched score: 5.0") == 5.0
assert parse_matched_score("Matched score: 6") is None
assert parse_matched_score("no score") is None
def test_score_generation_normalizes_average_matched_score() -> None:
predictions = [
{"sample_id": "s1", "final_hypothesis": "hypothesis", "matched_score": 2},
{"sample_id": "s2", "final_hypothesis": "hypothesis", "matched_score": 4},
]
data = [
{"sample_id": "s1", "gold_hypothesis": "gold 1"},
{"sample_id": "s2", "gold_hypothesis": "gold 2"},
]
result = score_generation(predictions, data, MockClient())
assert result["summary"]["average_score"] == 3.0
assert result["summary"]["accuracy"] == 0.6
def test_run_generation_accepts_legacy_ture_retrieve_key() -> None:
record = {
"sample_id": "s1",
"research_question": "question",
"background_survey": "survey",
"ture_retrieve": [{"title": "Inspiration", "abstract": "abstract"}],
}
result = run_generation_record(record, MockClient(), num_mutations=1, num_itr_self_refine=1)
assert "moosechem_trace" in result
assert result["moosechem_trace"]["inspiration_results"][0]["inspiration_title"] == "Inspiration"
assert result["final_hypothesis"]
def test_run_generation_defaults_to_no_background_survey_for_moosechem_reproducibility() -> None:
record = {
"sample_id": "s1",
"research_question": "question",
"background_survey": "private survey text",
"gold_inspirations": [{"title": "Inspiration", "abstract": "abstract"}],
}
result = run_generation_record(record, MockClient(), num_mutations=1, num_itr_self_refine=1)
prompt = result["moosechem_trace"]["inspiration_results"][0]["mutations"]["0"]["trace"][0]["prompt"]
assert "Survey not provided. Please overlook the survey." in prompt
assert "private survey text" not in prompt
assert result["moosechem_trace"]["params"]["if_use_background_survey"] is False
def test_estimate_generation_model_calls_default_three_inspirations() -> None:
record = {
"sample_id": "s1",
"gold_inspirations": [
{"title": "A", "abstract": "a"},
{"title": "B", "abstract": "b"},
{"title": "C", "abstract": "c"},
],
}
assert estimate_generation_model_calls(record) == 105
|