{"id":"W4379276041","doi":"10.1093/bioadv/vbad068","title":"DataCurator.jl: efficient, portable and reproducible validation, curation and transformation of large heterogeneous datasets using human-readable recipes compiled into machine-verifiable templates","year":2023,"lang":"en","type":"article","venue":"Bioinformatics Advances","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia; Simon Fraser University","funders":"","keywords":"Computer science; Executable; Scalability; Python (programming language); Preprocessor; Workflow; Data mining; Verifiable secret sharing; Data curation; Database; Programming language","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005023274,0.0001791047,0.0003209612,0.0005057193,0.0007010265,0.0005820518,0.0004148812,0.00004365424,0.00003767332],"category_scores_gemma":[0.0006414028,0.0001427572,0.00003329574,0.001117534,0.0001118509,0.002805563,0.0003451617,0.00006644009,0.00003320528],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002037536,"about_ca_system_score_gemma":0.00003277618,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009174537,"about_ca_topic_score_gemma":0.00006470878,"domain_scores_codex":[0.9970427,0.00007451566,0.001213596,0.0005156636,0.0008547709,0.0002988273],"domain_scores_gemma":[0.997825,0.0002692273,0.0006319826,0.0009906247,0.0001928415,0.0000903447],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003296826,0.0008177079,0.03497032,0.003854626,0.0003283104,0.0000243926,0.02961265,0.6428764,0.01801957,0.01854708,0.06097408,0.1896452],"study_design_scores_gemma":[0.0005320617,0.00006565423,0.0003912138,0.0000825637,0.00003070918,0.00001288021,0.003010001,0.9360471,0.008890735,0.002650152,0.04804126,0.0002456584],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9373771,0.0008273702,0.05807723,0.000112346,0.0005470209,0.0006436531,0.001797902,0.0001407036,0.0004766941],"genre_scores_gemma":[0.9800448,0.0002346238,0.01582197,0.00003989186,0.00003073093,0.000009660789,0.003683472,0.00001272653,0.0001221502],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2931707,"threshold_uncertainty_score":0.5821468,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09799458101637326,"score_gpt":0.4068712497562927,"score_spread":0.3088766687399195,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}