Replicate Readings

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

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

1# Reference solution for the yamaa benchmark sdtm-vs-replicates (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)1011fmt_text <- function(x) {12  if (is.na(x)) {13    return(NA_character_)14  }15  if (x == floor(x)) {16    formatC(x, format = "f", digits = 0)17  } else {18    format(x, scientific = FALSE, trim = TRUE, digits = 15)19  }20}2122odm <- read_csv(23  "/app/input/odm.csv",24  col_types = cols(.default = col_character())25)2627readings <- odm |>28  filter(29    ItemOID %in% c("IT.VS.SYSBP", "IT.VS.DIABP"),30    StudyEventOID %in% c("SE.VISIT1", "SE.VISIT2"),31    !is.na(Value) & Value != ""32  ) |>33  mutate(34    VISIT = case_when(35      StudyEventOID == "SE.VISIT1" ~ "VISIT 1",36      .default = "VISIT 2"37    ),38    VSTESTCD = case_when(39      ItemOID == "IT.VS.DIABP" ~ "DIABP",40      .default = "SYSBP"41    ),42    VSTEST = case_when(43      ItemOID == "IT.VS.DIABP" ~ "Diastolic Blood Pressure",44      .default = "Systolic Blood Pressure"45    ),46    VSREPNUM = suppressWarnings(as.integer(ItemGroupRepeatKey)),47    val = suppressWarnings(as.numeric(Value)),48    STUDYID = StudyOID,49    USUBJID = SubjectKey50  ) |>51  filter(!is.na(val), !is.na(VSREPNUM)) |>52  select(STUDYID, USUBJID, VISIT, VSTESTCD, VSTEST, VSREPNUM, val)5354vs <- readings |>55  transmute(56    DOMAIN = "VS",57    STUDYID, USUBJID, VISIT, VSTESTCD, VSTEST,58    VSREPNUM,59    VSSPID = as.character(VSREPNUM),60    VSORRES = vapply(val, fmt_text, character(1)),61    VSORRESU = "mmHg",62    VSSTRESN = val,63    VSSTRESC = VSORRES64  ) |>65  arrange(USUBJID, VISIT, VSTESTCD, VSREPNUM) |>66  group_by(USUBJID) |>67  mutate(VSSEQ = row_number()) |>68  ungroup() |>69  arrange(USUBJID, VSSEQ) |>70  select(71    DOMAIN, STUDYID, USUBJID, VSSEQ, VSSPID, VISIT, VSTESTCD, VSTEST,72    VSORRES, VSORRESU, VSSTRESN, VSSTRESC73  )7475dir.create("/app/output", showWarnings = FALSE)76write_csv(vs, "/app/output/vs.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-vs-replicates (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 1.67import csv8import re9from collections import OrderedDict1011INPUT = "/app/input"12OUTPUT = "/app/output"1314TESTS = {15    "SYSBP": "Systolic Blood Pressure",16    "DIABP": "Diastolic Blood Pressure",17}1819COLUMNS = [20    "DOMAIN", "STUDYID", "USUBJID", "VSSEQ", "VSSPID", "VISIT", "VSTESTCD",21    "VSTEST", "VSORRES", "VSORRESU", "VSSTRESN", "VSSTRESC",22]232425def fmt(value):26    """Write a number as text, dropping trailing zeros (82.50 -> 82.5)."""27    if value == int(value):28        return str(int(value))29    return repr(value)303132def visit_name(event_oid):33    name = event_oid.split(".", 1)[1] if "." in event_oid else event_oid34    m = re.match(r"^(.*?)(\d+)$", name)35    if m:36        return m.group(1) + " " + m.group(2)37    return name383940def main():41    with open(INPUT + "/odm.csv", newline="") as f:42        rows = list(csv.DictReader(f))4344    studyid = rows[0]["StudyOID"]4546    # readings[(subject, visit, testcd)] = [(repnum, value), ...]47    readings = OrderedDict()48    subjects = []49    for row in rows:50        subject = row["SubjectKey"]51        if subject not in subjects:52            subjects.append(subject)53        if row["ItemGroupOID"] != "IG.VS":54            continue55        testcd = row["ItemOID"].rsplit(".", 1)[-1]56        if testcd not in TESTS:57            continue58        value = (row["Value"] or "").strip()59        if value == "":60            continue  # reading not taken: no record61        key = (subject, visit_name(row["StudyEventOID"]), testcd)62        readings.setdefault(key, []).append(63            (int(row["ItemGroupRepeatKey"]), float(value))64        )6566    out = []67    for subject in subjects:68        records = []69        keys = [k for k in readings if k[0] == subject]70        for key in sorted(keys, key=lambda k: (k[1], k[2])):71            _, visit, testcd = key72            for repnum, value in sorted(readings[key]):73                records.append({74                    "DOMAIN": "VS",75                    "STUDYID": studyid,76                    "USUBJID": subject,77                    "VSSPID": str(repnum),78                    "VISIT": visit,79                    "VSTESTCD": testcd,80                    "VSTEST": TESTS[testcd],81                    "VSORRES": fmt(value),82                    "VSORRESU": "mmHg",83                    "VSSTRESN": fmt(value),84                    "VSSTRESC": fmt(value),85                })86        for i, rec in enumerate(records, start=1):87            rec["VSSEQ"] = str(i)88            out.append(rec)8990    with open(OUTPUT + "/vs.csv", "w", newline="") as f:91        writer = csv.DictWriter(f, fieldnames=COLUMNS)92        writer.writeheader()93        writer.writerows(out)949596if __name__ == "__main__":97    main()

1/* opensas reference solution for the yamaa benchmark sdtm-vs-replicates.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;1011data readings; set odm; where ItemOID in ('IT.VS.SYSBP','IT.VS.DIABP') and not missing(Value); length STUDYID USUBJID VISIT VSTESTCD $1024;12STUDYID=StudyOID; USUBJID=SubjectKey; VISIT=catx(' ','VISIT',scan(StudyEventOID,-1,'T')); if StudyEventOID='SE.VISIT1' then VISIT='VISIT 1'; if StudyEventOID='SE.VISIT2' then VISIT='VISIT 2'; VSTESTCD=scan(ItemOID,-1,'.'); VSREPNUM=input(ItemGroupRepeatKey,best32.); VSSTRESN=input(Value,best32.); run;13data prepared; set readings; length DOMAIN $2 VSSPID VSTEST VSORRES VSORRESU VSSTRESC $1024; DOMAIN='VS'; VSSPID=strip(ItemGroupRepeatKey); VSTEST='Diastolic Blood Pressure'; if VSTESTCD='SYSBP' then VSTEST='Systolic Blood Pressure'; VSORRES=strip(put(VSSTRESN,best32.)); VSSTRESC=VSORRES; VSORRESU='mmHg'; run;14proc sort data=prepared; by STUDYID USUBJID VISIT VSTESTCD VSREPNUM; run;15data result; set prepared; by STUDYID USUBJID; retain VSSEQ;16  if first.USUBJID then VSSEQ=0;17  VSSEQ+1;18run;1920proc sql;21  create table final as select DOMAIN, STUDYID, USUBJID, VSSEQ, VSSPID, VISIT, VSTESTCD, VSTEST, VSORRES, VSORRESU, VSSTRESN, VSSTRESC from result;22quit;23proc export data=final outfile="/app/output/vs.csv" dbms=csv replace; run;