| from researchbench.data_io import ( | |
| best_title_match, | |
| contains_forbidden_text, | |
| normalize_doi, | |
| normalize_title, | |
| ) | |
| def test_normalize_doi_removes_doi_punctuation() -> None: | |
| assert normalize_doi("10.1016/j.ascom.2023.100771") == "101016jascom2023100771" | |
| assert normalize_doi("10.1038-S41586-024-07860-9") == "101038s41586024078609" | |
| def test_normalize_title_and_best_title_match() -> None: | |
| candidates = [ | |
| {"title": "Dictionary learning allows model-free pseudotime estimation of transcriptomic data"}, | |
| {"title": "Perfect metamaterial absorber"}, | |
| ] | |
| idx, score = best_title_match( | |
| "Dictionary Learning Allows Model Free Pseudotime Estimation of Transcriptomic Data", | |
| candidates, | |
| ) | |
| assert idx == 0 | |
| assert score == 1.0 | |
| assert normalize_title("PtTe2-based type-II dirac semimetal") == "ptte2 based type ii dirac semimetal" | |
| def test_forbidden_text_detection_avoids_common_scientific_false_positives() -> None: | |
| row = { | |
| "url": "https://github.com/example/project", | |
| "title": "Dusk-to-nighttime enhancement of mid-latitude Nm F2", | |
| } | |
| assert not contains_forbidden_text(row) | |
| assert contains_forbidden_text({"path": "/Users/alice/private/file.json"}) | |
| assert contains_forbidden_text({"token": "sk-proj-" + "x" * 32}) | |