Discontinuation Reasons

1library(yamaa)23yamaa_spec("spec.yaml")$run()$output

1import yamaa23yamaa.spec("spec.yaml").run().output

1# Reference solution for the yamaa benchmark sdtm-ds-reasons (R track).2#3# Written from the benchmark's full prompt and its inputs alone, as an agent4# would write it. Harbor's oracle agent runs it to check that the prompt can5# be solved and that the grader scores a correct answer 1.67library(dplyr, warn.conflicts = FALSE)8library(readr)9library(tidyr)1011odm <- read_csv(12  "/app/input/odm.csv",13  col_types = cols(.default = col_character())14)1516wide <- odm |>17  filter(StudyEventOID %in% c("EOT", "EOS")) |>18  group_by(StudyOID, SubjectKey, StudyEventOID) |>19  summarise(20    DTC = Value[ItemOID == "IT.DS.DTC"][1],21    COMP = Value[ItemOID == "IT.DS.COMP"][1],22    REASON = Value[ItemOID == "IT.DS.REASON"][1],23    REASONCD = Value[ItemOID == "IT.DS.REASONCD"][1],24    .groups = "drop"25  )2627ds <- wide |>28  mutate(29    DOMAIN = "DS",30    STUDYID = StudyOID,31    USUBJID = SubjectKey,32    DSSEQ = ifelse(StudyEventOID == "EOT", 1L, 2L),33    DSCAT = "DISPOSITION EVENT",34    DSSCAT = ifelse(StudyEventOID == "EOT", "STUDY TREATMENT", "STUDY"),35    DSTERM = ifelse(COMP == "COMPLETED", "COMPLETED", na_if(REASON, "")),36    DSDECOD = ifelse(COMP == "COMPLETED", "COMPLETED", na_if(REASONCD, "")),37    DSSTDTC = suppressWarnings(as.Date(na_if(DTC, "")))38  ) |>39  select(DOMAIN, STUDYID, USUBJID, DSSEQ, DSCAT, DSSCAT, DSTERM, DSDECOD, DSSTDTC) |>40  arrange(USUBJID, DSSEQ)4142dir.create("/app/output", showWarnings = FALSE)43write_csv(ds, "/app/output/ds.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-ds-reasons (Python track).2#3# Written from the benchmark's full prompt and its inputs alone, as an agent4# would write it. Harbor's oracle agent runs it to check that the prompt can5# be solved and that the grader scores a correct answer 1.67"""Build the SDTM DS dataset from the ODM extract.89One record per subject for each of the two milestones: end of10treatment (EOT) and end of study (EOS). Each milestone's completion11flag, reason text, controlled code, and start date are collected as12separate items in the ODM rows; the record for a milestone carries13the values of that milestone's items, in whatever order the input14rows hold them.15"""1617import csv18from pathlib import Path1920INPUT = Path("/app/input")21OUTPUT = Path("/app/output")2223# (StudyEventOID, DSSEQ, DSSCAT) for each milestone, in DSSEQ order.24MILESTONES = [25    ("EOT", 1, "STUDY TREATMENT"),26    ("EOS", 2, "STUDY"),27]2829ITEM_COMPLETION = "IT.DS.COMP"30ITEM_REASON = "IT.DS.REASON"31ITEM_REASON_CODE = "IT.DS.REASONCD"32ITEM_DATE = "IT.DS.DTC"3334COLUMNS = [35    "DOMAIN",36    "STUDYID",37    "USUBJID",38    "DSSEQ",39    "DSCAT",40    "DSSCAT",41    "DSTERM",42    "DSDECOD",43    "DSSTDTC",44]454647def main() -> None:48    with open(INPUT / "odm.csv", newline="") as handle:49        odm_rows = list(csv.DictReader(handle))5051    # Collect each milestone's items per subject:52    # items[(subject, event)][item oid] = value.53    items: dict[tuple[str, str], dict[str, str]] = {}54    study_by_subject: dict[str, str] = {}55    for row in odm_rows:56        subject = row["SubjectKey"]57        study_by_subject.setdefault(subject, row["StudyOID"])58        key = (subject, row["StudyEventOID"])59        items.setdefault(key, {})[row["ItemOID"]] = row["Value"]6061    records = []62    for subject in sorted(study_by_subject):63        for event_oid, dsseq, dsscat in MILESTONES:64            collected = items.get((subject, event_oid), {})65            if collected.get(ITEM_COMPLETION) == "COMPLETED":66                dsterm = "COMPLETED"67                dsdecod = "COMPLETED"68            else:69                dsterm = collected.get(ITEM_REASON, "")70                dsdecod = collected.get(ITEM_REASON_CODE, "")71            records.append(72                {73                    "DOMAIN": "DS",74                    "STUDYID": study_by_subject[subject],75                    "USUBJID": subject,76                    "DSSEQ": str(dsseq),77                    "DSCAT": "DISPOSITION EVENT",78                    "DSSCAT": dsscat,79                    "DSTERM": dsterm,80                    "DSDECOD": dsdecod,81                    "DSSTDTC": collected.get(ITEM_DATE, ""),82                }83            )8485    OUTPUT.mkdir(parents=True, exist_ok=True)86    with open(OUTPUT / "ds.csv", "w", newline="") as handle:87        writer = csv.DictWriter(handle, fieldnames=COLUMNS)88        writer.writeheader()89        writer.writerows(records)909192if __name__ == "__main__":93    main()

1/* opensas reference solution for the yamaa benchmark sdtm-ds-reasons.2   Reads the benchmark inputs and derives the requested output.3   Execute with the existing opensas runtime: opensas /app/output/result.sas. */45data odm;6  length StudyOID MetaDataVersionOID SubjectKey StudyEventOID StudyEventRepeatKey ItemGroupOID ItemGroupRepeatKey ItemOID Value $1024;7  infile "/app/input/odm.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;8  input StudyOID : $char1024. MetaDataVersionOID : $char1024. SubjectKey : $char1024. StudyEventOID : $char1024. StudyEventRepeatKey : $char1024. ItemGroupOID : $char1024. ItemGroupRepeatKey : $char1024. ItemOID : $char1024. Value : $char1024.;9run;1011proc sort data=odm out=ordered; by StudyOID SubjectKey StudyEventOID; run;12data wide;13  length COMP DTC REASON REASONCD $1024;14  retain COMP DTC REASON REASONCD;15  set ordered;1617  by StudyOID SubjectKey StudyEventOID;18  if first.StudyEventOID then do;19    COMP=''; DTC=''; REASON=''; REASONCD='';20  end;21    if ItemOID='IT.DS.COMP' then COMP=Value;22  if ItemOID='IT.DS.DTC' then DTC=Value;23  if ItemOID='IT.DS.REASON' then REASON=Value;24  if ItemOID='IT.DS.REASONCD' then REASONCD=Value;25  if last.StudyEventOID then output;26run;27data prepared; set wide; length DOMAIN $2 STUDYID USUBJID DSTERM DSDECOD DSCAT DSSCAT DSSTDTC $1024;28DOMAIN='DS'; STUDYID=StudyOID; USUBJID=SubjectKey; DSCAT='DISPOSITION EVENT'; DSSTDTC=DTC; DSTERM=REASON; DSDECOD=REASONCD;29if COMP='COMPLETED' then do; DSTERM='COMPLETED'; DSDECOD='COMPLETED'; end;30if StudyEventOID='EOT' then do; DSSEQ=1; DSSCAT='STUDY TREATMENT'; end; else do; DSSEQ=2; DSSCAT='STUDY'; end; run; data result; set prepared; run;31proc sql;32  create table final as select DOMAIN, STUDYID, USUBJID, DSSEQ, DSCAT, DSSCAT, DSTERM, DSDECOD, DSSTDTC from result;33quit;34proc export data=final outfile="/app/output/ds.csv" dbms=csv replace; run;