Missed Temperature Diary Days

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

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

1# Reference solution for the yamaa benchmark sdtm-vs-diary-missed-days (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 1.67suppressPackageStartupMessages({8  library(dplyr)9  library(readr)10})1112read_chr <- function(path) {13  read_csv(path, col_types = cols(.default = col_character()),14           na = character(), trim_ws = FALSE)15}1617diary <- read_chr("/app/input/diary_days.csv")18temp  <- read_chr("/app/input/temperature.csv")19vac   <- read_chr("/app/input/vaccination.csv")2021blank_to_na <- function(x) ifelse(is.na(x) | trimws(x) == "", NA_character_, x)2223diary <- diary %>%24  mutate(VSTPTNUM = as.integer(DIARYDAY), VSTPT = DAYLABEL) %>%25  select(STUDYID, VSTPTNUM, VSTPT)2627given <- vac %>%28  mutate(VACDAT = blank_to_na(VACDAT)) %>%29  filter(!is.na(VACDAT)) %>%30  mutate(VACDATE = as.Date(VACDAT))3132temp <- temp %>%33  mutate(TEMPDAT = blank_to_na(TEMPDAT),34         TEMP = blank_to_na(TEMP),35         TEMPU = blank_to_na(TEMPU)) %>%36  filter(!is.na(TEMPDAT)) %>%37  select(STUDYID, USUBJID, VSDTC = TEMPDAT, VSORRES = TEMP, VSORRESU = TEMPU)3839vs <- given %>%40  inner_join(diary, by = "STUDYID", relationship = "many-to-many") %>%41  mutate(VSDTC = format(VACDATE + (VSTPTNUM - 1L), "%Y-%m-%d"),42         VSTPTREF = VISIT,43         VSRFTDTC = VACDAT) %>%44  left_join(temp, by = c("STUDYID", "USUBJID", "VSDTC")) %>%45  mutate(46    DOMAIN = "VS",47    VSTESTCD = "TEMP",48    VSTEST = "Temperature",49    entered = !is.na(VSORRES),50    VSSTAT = ifelse(entered, NA_character_, "NOT DONE"),51    VSREASND = ifelse(entered, NA_character_, "NOT ENTERED")52  ) %>%53  arrange(USUBJID, VACDATE, VSTPTNUM) %>%54  group_by(USUBJID) %>%55  mutate(VSSEQ = row_number()) %>%56  ungroup() %>%57  select(DOMAIN, STUDYID, USUBJID, VSSEQ, VSTESTCD, VSTEST, VSORRES, VSORRESU,58         VSSTAT, VSREASND, VSDTC, VSTPT, VSTPTNUM, VSTPTREF, VSRFTDTC)5960dir.create("/app/output", showWarnings = FALSE)61write_csv(vs, "/app/output/vs.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-vs-diary-missed-days (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 datetime as dt9from pathlib import Path1011INPUT = Path("/app/input")12OUTPUT = Path("/app/output")1314COLUMNS = [15    "DOMAIN", "STUDYID", "USUBJID", "VSSEQ", "VSTESTCD", "VSTEST", "VSORRES",16    "VSORRESU", "VSSTAT", "VSREASND", "VSDTC", "VSTPT", "VSTPTNUM", "VSTPTREF",17    "VSRFTDTC",18]192021def read(name):22    with open(INPUT / name, newline="", encoding="utf-8") as handle:23        return [24            {k.strip(): (v or "").strip() for k, v in row.items()}25            for row in csv.DictReader(handle)26        ]272829def main():30    days = sorted(read("diary_days.csv"), key=lambda r: int(r["DIARYDAY"]))31    temps = {}32    for row in read("temperature.csv"):33        temps.setdefault((row["USUBJID"], row["TEMPDAT"]), row)3435    given = [v for v in read("vaccination.csv") if v["VACDAT"]]36    given.sort(key=lambda v: (v["USUBJID"], v["VACDAT"], int(v["DOSENUM"] or 0)))3738    records = []39    seq = {}40    for vac in given:41        start = dt.date.fromisoformat(vac["VACDAT"])42        for day in days:43            num = int(day["DIARYDAY"])44            date = (start + dt.timedelta(days=num - 1)).isoformat()45            temp = temps.get((vac["USUBJID"], date))46            seq[vac["USUBJID"]] = seq.get(vac["USUBJID"], 0) + 147            records.append({48                "DOMAIN": "VS",49                "STUDYID": vac["STUDYID"],50                "USUBJID": vac["USUBJID"],51                "VSSEQ": seq[vac["USUBJID"]],52                "VSTESTCD": "TEMP",53                "VSTEST": "Temperature",54                "VSORRES": temp["TEMP"] if temp else "",55                "VSORRESU": temp["TEMPU"] if temp else "",56                "VSSTAT": "" if temp else "NOT DONE",57                "VSREASND": "" if temp else "NOT ENTERED",58                "VSDTC": date,59                "VSTPT": day["DAYLABEL"],60                "VSTPTNUM": num,61                "VSTPTREF": vac["VISIT"],62                "VSRFTDTC": vac["VACDAT"],63            })6465    OUTPUT.mkdir(parents=True, exist_ok=True)66    with open(OUTPUT / "vs.csv", "w", newline="", encoding="utf-8") as handle:67        writer = csv.DictWriter(handle, fieldnames=COLUMNS)68        writer.writeheader()69        writer.writerows(records)707172if __name__ == "__main__":73    main()

1/* opensas reference solution for the yamaa benchmark sdtm-vs-diary-missed-days.2   Reads the benchmark inputs and derives the requested output.3   Execute with the existing opensas runtime: opensas /app/output/result.sas. */45data vaccination;6  length STUDYID USUBJID VISIT $1024;7  infile "/app/input/vaccination.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;8  input STUDYID : $char1024. USUBJID : $char1024. DOSENUM VISIT : $char1024. VACDAT : yymmdd10.;9  /* A vaccination with no date was not given and expects no diary. */10  if not missing(VACDAT);11run;1213data diary_days;14  length STUDYID DAYLABEL $1024;15  infile "/app/input/diary_days.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;16  input STUDYID : $char1024. DIARYDAY DAYLABEL : $char1024.;17run;1819data temperature;20  length STUDYID USUBJID TEMP TEMPU $1024;21  infile "/app/input/temperature.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;22  input STUDYID : $char1024. USUBJID : $char1024. TEMPDAT : yymmdd10. TEMP : $char1024. TEMPU : $char1024.;23run;2425/* Each vaccination given with each diary day; Day 1 is the day of26   vaccination. The entered temperature, when there is one, is matched by27   subject and diary date. */28proc sql;29  create table expected as select v.STUDYID, v.USUBJID, v.VISIT, v.VACDAT, d.DIARYDAY, d.DAYLABEL, v.VACDAT + d.DIARYDAY - 1 as DIARYDAT from vaccination as v inner join diary_days as d on v.STUDYID=d.STUDYID;30  create table joined as select e.*, t.TEMP, t.TEMPU from expected as e left join temperature as t on e.USUBJID=t.USUBJID and e.DIARYDAT=t.TEMPDAT;31quit;3233proc sort data=joined; by STUDYID USUBJID VACDAT DIARYDAY; run;34data result;35  length DOMAIN $2 VSTESTCD VSTEST VSORRES VSORRESU VSSTAT VSREASND VSDTC VSTPT VSTPTREF VSRFTDTC $1024;36  set joined;37  by STUDYID USUBJID;38  retain VSSEQ;39  if first.USUBJID then VSSEQ=0;40  VSSEQ+1;41  DOMAIN='VS'; VSTESTCD='TEMP'; VSTEST='Temperature';42  VSORRES=TEMP; VSORRESU=TEMPU; VSSTAT=''; VSREASND='';43  if missing(TEMP) then do; VSSTAT='NOT DONE'; VSREASND='NOT ENTERED'; end;44  VSDTC=put(DIARYDAT, yymmdd10.); VSTPT=DAYLABEL; VSTPTNUM=DIARYDAY;45  VSTPTREF=VISIT; VSRFTDTC=put(VACDAT, yymmdd10.);46run;4748proc sql;49  create table final as select DOMAIN, STUDYID, USUBJID, VSSEQ, VSTESTCD, VSTEST, VSORRES, VSORRESU, VSSTAT, VSREASND, VSDTC, VSTPT, VSTPTNUM, VSTPTREF, VSRFTDTC from result;50quit;51proc export data=final outfile="/app/output/vs.csv" dbms=csv replace; run;