Basic Demographics

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

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

1# Reference solution for the yamaa benchmark sdtm-dm-basic (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)10library(stringr)1112odm <- read_csv(13  "/app/input/odm.csv",14  col_types = cols(.default = col_character())15)1617wide <- odm |>18  pivot_wider(19    id_cols = c(StudyOID, SubjectKey),20    names_from = ItemOID,21    values_from = Value,22    values_fn = first23  )24for (column in c(25  "IT.DM.SITEID", "IT.DM.SEX", "IT.DM.AGE", "IT.DM.ARM", "IT.DM.COUNTRY"26)) {27  if (!column %in% names(wide)) wide[[column]] <- NA_character_28}29dm <- wide |>30  mutate(31    DOMAIN = "DM",32    STUDYID = StudyOID,33    USUBJID = SubjectKey,34    SUBJID = SubjectKey,35    SITEID = `IT.DM.SITEID`,36    # M for Male, F for Female; anything missing, blank, uncollected, or37    # otherwise recorded becomes U.38    SEX = case_when(39      `IT.DM.SEX` == "Male" ~ "M",40      `IT.DM.SEX` == "Female" ~ "F",41      .default = "U"42    ),43    AGE = suppressWarnings(as.integer(`IT.DM.AGE`)),44    AGEU = if_else(is.na(AGE), NA_character_, "YEARS"),45    ARM = na_if(str_trim(`IT.DM.ARM`), ""),46    # The actual arm always equals the planned arm.47    ACTARM = ARM,48    ARMNRS = if_else(is.na(ARM), "NOT ASSIGNED", NA_character_),49    COUNTRY = `IT.DM.COUNTRY`50  ) |>51  arrange(USUBJID) |>52  select(53    DOMAIN, STUDYID, USUBJID, SUBJID, SITEID, AGE, AGEU, SEX, ARM, ACTARM,54    ARMNRS, COUNTRY55  )5657dir.create("/app/output", showWarnings = FALSE)58write_csv(dm, "/app/output/dm.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-dm-basic (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 DM dataset from the long-format ODM extract.89One record per subject, with columns:10DOMAIN, STUDYID, USUBJID, SUBJID, SITEID, AGE, AGEU, SEX, ARM, ACTARM,11ARMNRS, COUNTRY12"""1314import csv15from collections import OrderedDict1617INPUT_PATH = "/app/input/odm.csv"18OUTPUT_PATH = "/app/output/dm.csv"1920ITEM_SITEID = "IT.DM.SITEID"21ITEM_SEX = "IT.DM.SEX"22ITEM_AGE = "IT.DM.AGE"23ITEM_ARM = "IT.DM.ARM"24ITEM_COUNTRY = "IT.DM.COUNTRY"252627def main():28    subjects = OrderedDict()  # SubjectKey -> {ItemOID: Value}29    study_ids = {}  # SubjectKey -> StudyOID30    with open(INPUT_PATH, newline="", encoding="utf-8") as fh:31        for row in csv.DictReader(fh):32            key = row["SubjectKey"]33            if key not in subjects:34                subjects[key] = {}35                study_ids[key] = row["StudyOID"]36            subjects[key][row["ItemOID"]] = row["Value"]3738    rows = []39    for key, items in subjects.items():40        collected_sex = items.get(ITEM_SEX, "")41        if collected_sex == "Male":42            sex = "M"43        elif collected_sex == "Female":44            sex = "F"45        else:46            sex = "U"47        age = (items.get(ITEM_AGE, "") or "").strip()48        arm = items.get(ITEM_ARM, "") or ""49        actarm = arm50        armnrs = "" if arm else "NOT ASSIGNED"51        rows.append(52            {53                "DOMAIN": "DM",54                "STUDYID": study_ids[key],55                "USUBJID": key,56                "SUBJID": key,57                "SITEID": items.get(ITEM_SITEID, ""),58                "AGE": age,59                "AGEU": "YEARS" if age else "",60                "SEX": sex,61                "ARM": arm,62                "ACTARM": actarm,63                "ARMNRS": armnrs,64                "COUNTRY": items.get(ITEM_COUNTRY, ""),65            }66        )6768    fieldnames = [69        "DOMAIN",70        "STUDYID",71        "USUBJID",72        "SUBJID",73        "SITEID",74        "AGE",75        "AGEU",76        "SEX",77        "ARM",78        "ACTARM",79        "ARMNRS",80        "COUNTRY",81    ]82    with open(OUTPUT_PATH, "w", newline="", encoding="utf-8") as fh:83        writer = csv.DictWriter(fh, fieldnames=fieldnames)84        writer.writeheader()85        writer.writerows(rows)868788if __name__ == "__main__":89    main()

1/* opensas reference solution for the yamaa benchmark sdtm-dm-basic.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; run;12data wide;13  length SITEID SEX AGE ARM COUNTRY $1024;14  retain SITEID SEX AGE ARM COUNTRY;15  set ordered;16  by StudyOID SubjectKey;17  if first.SubjectKey then do;18    SITEID=''; SEX=''; AGE=''; ARM=''; COUNTRY='';19  end;20    if ItemOID='IT.DM.SEX' then SEX=Value;21  if ItemOID='IT.DM.AGE' then AGE=Value;22  if ItemOID='IT.DM.ARM' then ARM=Value;23  if ItemOID='IT.DM.SITEID' then SITEID=Value;24  if ItemOID='IT.DM.COUNTRY' then COUNTRY=Value;25  if last.SubjectKey then output;26run;27data result; set wide; length DOMAIN $2 STUDYID USUBJID SUBJID AGEU ACTARM ARMNRS $1024;28DOMAIN='DM'; STUDYID=StudyOID; USUBJID=SubjectKey; SUBJID=SubjectKey;29if SEX='Male' then SEX='M'; else if SEX='Female' then SEX='F'; else SEX='U'; if not missing(AGE) then AGEU='YEARS'; ACTARM=ARM; if missing(ARM) then ARMNRS='NOT ASSIGNED';30run;31proc sql;32  create table final as select DOMAIN, STUDYID, USUBJID, SUBJID, SITEID, AGE, AGEU, SEX, ARM, ACTARM, ARMNRS, COUNTRY from result;33quit;34proc export data=final outfile="/app/output/dm.csv" dbms=csv replace; run;