Dose Intervals

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

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

1# Reference solution for the yamaa benchmark sdtm-ex-intervals (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)910ec_raw <- read_csv(11  "/app/input/ec_raw.csv",12  col_types = cols(.default = col_character())13)1415ex <- ec_raw |>16  mutate(17    EXDOSE_NUM = suppressWarnings(as.numeric(ECDOSE)),18    ECSTDTC_DT = suppressWarnings(as.Date(ECSTDTC)),19    ECADJ_CLEAN = na_if(ECADJ, "")20  ) |>21  group_by(STUDYID, USUBJID, ECTRT, EXDOSE_NUM, ECDOSU, ECDOSFRQ) |>22  summarise(23    START_DT = min(ECSTDTC_DT, na.rm = TRUE),24    END_DT = max(ECSTDTC_DT, na.rm = TRUE),25    EXADJ = {26      vals <- sort(unique(ECADJ_CLEAN[!is.na(ECADJ_CLEAN)]))27      if (length(vals) == 0) NA_character_ else vals[1]28    },29    .groups = "drop"30  ) |>31  mutate(32    DOMAIN = "EX",33    EXTRT = ECTRT,34    EXDOSE = EXDOSE_NUM,35    EXDOSU = ECDOSU,36    EXDOSFRQ = ECDOSFRQ,37    EXSTDTC = START_DT,38    EXENDTC = END_DT39  ) |>40  arrange(STUDYID, USUBJID, START_DT, EXTRT) |>41  group_by(STUDYID, USUBJID) |>42  mutate(EXSEQ = row_number()) |>43  ungroup() |>44  select(45    DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXDOSE, EXDOSU,46    EXDOSFRQ, EXSTDTC, EXENDTC, EXADJ47  )4849dir.create("/app/output", showWarnings = FALSE)50write_csv(ex, "/app/output/ex.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-ex-intervals (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 EX dataset (one record per constant-dose interval) from EC_RAW.89Reads /app/input/ec_raw.csv and writes /app/output/ex.csv with columns:10DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXDOSE, EXDOSU, EXDOSFRQ,11EXSTDTC, EXENDTC, EXADJ.12"""1314import csv15import os1617INPUT_DIR = "/app/input"18OUTPUT_DIR = "/app/output"192021def main():22    with open(os.path.join(INPUT_DIR, "ec_raw.csv"), newline="") as f:23        rows = list(csv.DictReader(f))2425    # Administrations of one subject and treatment, in date order.26    rows.sort(key=lambda r: (r["USUBJID"], r["ECTRT"], r["ECSTDTC"]))2728    # Consecutive administrations at one dose, dose unit, and dosing29    # frequency form one interval. A dose interruption (dose 0 with a30    # recorded reason) is its own interval like any other dose level.31    intervals = []32    for r in rows:33        key = (34            r["STUDYID"],35            r["USUBJID"],36            r["ECTRT"],37            r["ECDOSE"],38            r["ECDOSU"],39            r["ECDOSFRQ"],40        )41        if intervals and intervals[-1]["key"] == key:42            iv = intervals[-1]43            iv["end"] = r["ECSTDTC"]44            if not iv["adj"] and r["ECADJ"]:45                iv["adj"] = r["ECADJ"]46        else:47            intervals.append(48                {49                    "key": key,50                    "start": r["ECSTDTC"],51                    "end": r["ECSTDTC"],52                    "adj": r["ECADJ"],53                }54            )5556    # EXSEQ numbers each subject's intervals from 1 in the order they57    # started, by start date then treatment.58    by_subject = {}59    for iv in intervals:60        by_subject.setdefault(iv["key"][1], []).append(iv)6162    out_rows = []63    for usubjid in sorted(by_subject):64        subject_intervals = sorted(65            by_subject[usubjid], key=lambda iv: (iv["start"], iv["key"][2])66        )67        for seq, iv in enumerate(subject_intervals, start=1):68            studyid, _, trt, dose, dosu, dosfrq = iv["key"]69            out_rows.append(70                {71                    "DOMAIN": "EX",72                    "STUDYID": studyid,73                    "USUBJID": usubjid,74                    "EXSEQ": seq,75                    "EXTRT": trt,76                    "EXDOSE": dose,77                    "EXDOSU": dosu,78                    "EXDOSFRQ": dosfrq,79                    "EXSTDTC": iv["start"],80                    "EXENDTC": iv["end"],81                    "EXADJ": iv["adj"],82                }83            )8485    fieldnames = [86        "DOMAIN",87        "STUDYID",88        "USUBJID",89        "EXSEQ",90        "EXTRT",91        "EXDOSE",92        "EXDOSU",93        "EXDOSFRQ",94        "EXSTDTC",95        "EXENDTC",96        "EXADJ",97    ]98    os.makedirs(OUTPUT_DIR, exist_ok=True)99    with open(os.path.join(OUTPUT_DIR, "ex.csv"), "w", newline="") as f:100        writer = csv.DictWriter(f, fieldnames=fieldnames)101        writer.writeheader()102        writer.writerows(out_rows)103104105if __name__ == "__main__":106    main()

1/* opensas reference solution for the yamaa benchmark sdtm-ex-intervals.2   Reads the benchmark inputs and derives the requested output.3   Execute with the existing opensas runtime: opensas /app/output/result.sas. */45data ec_raw;6  length STUDYID USUBJID ECTRT ECDOSU ECDOSFRQ ECSTDTC ECADJ $1024;7  infile "/app/input/ec_raw.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;8  input STUDYID : $char1024. USUBJID : $char1024. ECTRT : $char1024. ECDOSE ECDOSU : $char1024. ECDOSFRQ : $char1024. ECSTDTC : $char1024. ECADJ : $char1024.;9run;1011proc sql;12create table intervals as select STUDYID,USUBJID,ECTRT as EXTRT,ECDOSE as EXDOSE,ECDOSU as EXDOSU,ECDOSFRQ as EXDOSFRQ,min(ECSTDTC) as EXSTDTC,max(ECSTDTC) as EXENDTC from ec_raw group by STUDYID,USUBJID,ECTRT,ECDOSE,ECDOSU,ECDOSFRQ;13create table reasons as select STUDYID,USUBJID,ECTRT,ECDOSE,ECDOSU,ECDOSFRQ,min(ECADJ) as EXADJ from ec_raw where not missing(ECADJ) group by STUDYID,USUBJID,ECTRT,ECDOSE,ECDOSU,ECDOSFRQ;14create table prepared as select 'EX' as DOMAIN,a.*,b.EXADJ from intervals a left join reasons b on a.STUDYID=b.STUDYID and a.USUBJID=b.USUBJID and a.EXTRT=b.ECTRT and a.EXDOSE=b.ECDOSE and a.EXDOSU=b.ECDOSU and a.EXDOSFRQ=b.ECDOSFRQ; quit;15proc sort data=prepared; by STUDYID USUBJID EXSTDTC EXTRT; run;16data result; set prepared; by STUDYID USUBJID; retain EXSEQ;17  if first.USUBJID then EXSEQ=0;18  EXSEQ+1;19run;2021proc sql;22  create table final as select DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXDOSE, EXDOSU, EXDOSFRQ, EXSTDTC, EXENDTC, EXADJ from result;23quit;24proc export data=final outfile="/app/output/ex.csv" dbms=csv replace; run;