Select the Actual Treatment and Its Exposure Dates

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

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

1# Reference solution for the yamaa benchmark adam-adsl-treatment (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(stringr)10library(lubridate)1112dm <- read_csv(13  "/app/input/dm.csv",14  col_types = cols(.default = col_character())15)16ex <- read_csv(17  "/app/input/ex.csv",18  col_types = cols(EXSEQ = col_integer(), .default = col_character())19)2021qual <- c("VITAMIN D3", "PLACEBO")2223parse_start <- function(x) {24  x[!is.na(x) & x == ""] <- NA_character_25  dt <- suppressWarnings(ymd_hms(x, quiet = TRUE, tz = "UTC"))26  d <- suppressWarnings(ymd(x, quiet = TRUE, tz = "UTC"))27  coalesce(dt, as.POSIXct(d, tz = "UTC"))28}2930parse_end <- function(x) {31  base <- parse_start(x)32  is_date <- !is.na(x) & x != "" & !str_detect(x, "T")33  base[is_date] <- base[is_date] + 23 * 3600 + 59 * 60 + 5934  base35}3637ex_dt <- ex |>38  mutate(39    `__start_dt` = parse_start(EXSTDTC),40    `__end_dt` = parse_end(EXENDTC)41  )4243first_start <- ex_dt |>44  filter(EXTRT %in% qual, !is.na(`__start_dt`)) |>45  arrange(`__start_dt`, EXSEQ) |>46  distinct(STUDYID, USUBJID, .keep_all = TRUE) |>47  select(STUDYID, USUBJID, TRT01RAW = EXTRT, `__stdtm` = `__start_dt`, `__stdtc` = EXSTDTC)4849last_end <- ex_dt |>50  filter(EXTRT %in% qual, !is.na(`__end_dt`)) |>51  arrange(desc(`__end_dt`)) |>52  distinct(STUDYID, USUBJID, .keep_all = TRUE) |>53  select(STUDYID, USUBJID, `__endtm` = `__end_dt`, `__endtc` = EXENDTC)5455adsl <- dm |>56  left_join(first_start, by = c("STUDYID", "USUBJID")) |>57  left_join(last_end, by = c("STUDYID", "USUBJID")) |>58  mutate(59    `__src` = coalesce(TRT01RAW, ACTARM),60    TRT01A = case_when(61      is.na(`__src`) | `__src` == "" ~ "NOT TREATED",62      .default = str_to_upper(`__src`)63    ),64    TRTSDTM = format(`__stdtm`, "%Y-%m-%dT%H:%M:%S", tz = "UTC"),65    TRTSDT = as.Date(`__stdtm`, tz = "UTC"),66    TRTSTMF = if_else(!is.na(`__stdtc`) & !str_detect(`__stdtc`, "T"), "H", NA_character_),67    TRTEDTM = format(`__endtm`, "%Y-%m-%dT%H:%M:%S", tz = "UTC"),68    TRTEDT = as.Date(`__endtm`, tz = "UTC"),69    TRTETMF = if_else(!is.na(`__endtc`) & !str_detect(`__endtc`, "T"), "H", NA_character_),70    TRTDURD = as.integer(TRTEDT - TRTSDT) + 1L,71    SAFFL = if_else(is.na(TRTSDT), "N", "Y")72  ) |>73  select(74    STUDYID, USUBJID, TRT01A, TRTSDT, TRTSDTM, TRTSTMF,75    TRTEDT, TRTEDTM, TRTETMF, TRTDURD, SAFFL76  ) |>77  arrange(USUBJID)7879dir.create("/app/output", showWarnings = FALSE)80write_csv(adsl, "/app/output/adsl.csv", na = "")

1# Reference solution for the yamaa benchmark adam-adsl-treatment (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 ADSL (one record per subject) from DM and EX.89Reads /app/input/dm.csv and /app/input/ex.csv and writes10/app/output/adsl.csv with columns:11STUDYID, USUBJID, TRT01A, TRTSDT, TRTSDTM, TRTSTMF, TRTEDT, TRTEDTM,12TRTETMF, TRTDURD, SAFFL13"""1415import csv16from datetime import datetime, time1718INPUT_DIR = "/app/input"19OUTPUT_PATH = "/app/output/adsl.csv"2021NAMED_TREATMENTS = {"VITAMIN D3", "PLACEBO"}22COLUMNS = [23    "STUDYID",24    "USUBJID",25    "TRT01A",26    "TRTSDT",27    "TRTSDTM",28    "TRTSTMF",29    "TRTEDT",30    "TRTEDTM",31    "TRTETMF",32    "TRTDURD",33    "SAFFL",34]353637def parse_dtc(value):38    """Split an ISO 8601 date/time collection value.3940    Returns (date, time_or_None, has_value). A value without a "T"41    separator was collected as a date only.42    """43    value = (value or "").strip()44    if not value:45        return None, None, False46    if "T" in value:47        date_part, time_part = value.split("T", 1)48        parsed_time = datetime.strptime(time_part.strip(), "%H:%M:%S").time()49        return (50            datetime.strptime(date_part.strip(), "%Y-%m-%d").date(),51            parsed_time,52            True,53        )54    return datetime.strptime(value, "%Y-%m-%d").date(), None, True555657def read_rows(path):58    with open(path, newline="") as handle:59        return list(csv.DictReader(handle))606162def main():63    dm_rows = read_rows(f"{INPUT_DIR}/dm.csv")64    ex_rows = read_rows(f"{INPUT_DIR}/ex.csv")6566    ex_by_subject = {}67    for row in ex_rows:68        ex_by_subject.setdefault(row["USUBJID"], []).append(row)6970    output_rows = []71    for dm in dm_rows:72        subject_ex = ex_by_subject.get(dm["USUBJID"], [])7374        # Qualifying exposure records: named treatment and a start value.75        qualifying = []76        for row in subject_ex:77            if row["EXTRT"].strip() not in NAMED_TREATMENTS:78                continue79            start_date, start_time, has_start = parse_dtc(row["EXSTDTC"])80            if has_start:81                effective = datetime.combine(82                    start_date, start_time or time(0, 0, 0)83                )84                qualifying.append(85                    (effective, int(row["EXSEQ"]), row, start_date, start_time)86                )87        qualifying.sort(key=lambda item: (item[0], item[1]))8889        trtsdt = trtsdtm = trtstmf = ""90        trtedt = trtedtm = trtetmf = ""91        trtdurd = ""9293        if qualifying:94            _, _, first_row, start_date, start_time = qualifying[0]95            trt01a = first_row["EXTRT"].strip().upper()96            trtsdt = start_date.isoformat()97            trtsdtm = (98                f"{start_date.isoformat()}T"99                f"{(start_time or time(0, 0, 0)).isoformat()}"100            )101            trtstmf = "H" if start_time is None else ""102        else:103            actarm = (dm.get("ACTARM") or "").strip()104            trt01a = actarm.upper() if actarm else "NOT TREATED"105106        # Latest end among records naming a treatment, with or without a107        # start; an end collected without a time counts as 23:59:59.108        end_candidates = []109        for row in subject_ex:110            if row["EXTRT"].strip() not in NAMED_TREATMENTS:111                continue112            end_date, end_time, has_end = parse_dtc(row["EXENDTC"])113            if has_end:114                effective = datetime.combine(115                    end_date, end_time or time(23, 59, 59)116                )117                end_candidates.append((effective, end_date, end_time))118119        if end_candidates:120            _, end_date, end_time = max(121                end_candidates, key=lambda item: item[0]122            )123            trtedt = end_date.isoformat()124            trtedtm = (125                f"{end_date.isoformat()}T"126                f"{(end_time or time(23, 59, 59)).isoformat()}"127            )128            trtetmf = "H" if end_time is None else ""129130        if trtsdt and trtedt:131            start = datetime.strptime(trtsdt, "%Y-%m-%d").date()132            end = datetime.strptime(trtedt, "%Y-%m-%d").date()133            trtdurd = str((end - start).days + 1)134135        output_rows.append(136            {137                "STUDYID": dm["STUDYID"],138                "USUBJID": dm["USUBJID"],139                "TRT01A": trt01a,140                "TRTSDT": trtsdt,141                "TRTSDTM": trtsdtm,142                "TRTSTMF": trtstmf,143                "TRTEDT": trtedt,144                "TRTEDTM": trtedtm,145                "TRTETMF": trtetmf,146                "TRTDURD": trtdurd,147                "SAFFL": "Y" if trtsdt else "N",148            }149        )150151    with open(OUTPUT_PATH, "w", newline="") as handle:152        writer = csv.DictWriter(handle, fieldnames=COLUMNS)153        writer.writeheader()154        writer.writerows(output_rows)155156157if __name__ == "__main__":158    main()

1/* opensas reference solution for the yamaa benchmark adam-adsl-treatment.2   Reads the benchmark inputs and derives the requested output.3   Execute with the existing opensas runtime: opensas /app/output/result.sas. */45data dm;6  length STUDYID USUBJID ACTARM $1024;7  infile "/app/input/dm.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;8  input STUDYID : $char1024. USUBJID : $char1024. ACTARM : $char1024.;9run;1011data ex;12  length STUDYID USUBJID EXTRT EXSTDTC EXENDTC $1024;13  infile "/app/input/ex.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;14  input STUDYID : $char1024. USUBJID : $char1024. EXSEQ EXTRT : $char1024. EXDOSE EXSTDTC : $char1024. EXENDTC : $char1024.;15run;1617data exposure;18  length startmoment endmoment $19;19  set ex;20  if upcase(EXTRT) in ('VITAMIN D3','PLACEBO');21  startmoment=EXSTDTC; endmoment=EXENDTC;22  if lengthn(EXSTDTC)=10 then startmoment=cats(EXSTDTC,'T00:00:00');23  else if lengthn(EXSTDTC)=16 then startmoment=cats(EXSTDTC,':00');24  if lengthn(EXENDTC)=10 then endmoment=cats(EXENDTC,'T23:59:59');25  else if lengthn(EXENDTC)=16 then endmoment=cats(EXENDTC,':00');26run;27proc sort data=exposure(where=(EXSTDTC ne '')) out=starts; by STUDYID USUBJID startmoment EXSEQ; run;28data firstex; set starts; by STUDYID USUBJID; if first.USUBJID; run;29proc sort data=exposure(where=(EXENDTC ne '')) out=ends; by STUDYID USUBJID descending endmoment; run;30data lastex; set ends; by STUDYID USUBJID; if first.USUBJID; run;31proc sql; create table joined as select a.*,b.EXTRT,b.EXSTDTC,b.startmoment,c.EXENDTC,c.endmoment from dm as a32  left join firstex as b on a.STUDYID=b.STUDYID and a.USUBJID=b.USUBJID left join lastex as c on a.STUDYID=c.STUDYID and a.USUBJID=c.USUBJID; quit;33data result;34  length TRT01A $1024 TRTSDT TRTEDT $10 TRTSDTM TRTEDTM $19 TRTSTMF TRTETMF SAFFL $1;35  set joined;36  TRT01A=upcase(coalescec(EXTRT,ACTARM,'NOT TREATED'));37  TRTSDT=''; TRTEDT=''; TRTSTMF=''; TRTETMF=''; TRTDURD=.;38  if lengthn(EXSTDTC)>=10 then TRTSDT=substr(EXSTDTC,1,10);39  if lengthn(EXENDTC)>=10 then TRTEDT=substr(EXENDTC,1,10);40  TRTSDTM=tranwrd(startmoment,'T',' '); TRTEDTM=tranwrd(endmoment,'T',' ');41  if lengthn(EXSTDTC)=10 then TRTSTMF='H';42  if lengthn(EXENDTC)=10 then TRTETMF='H';43  if not missing(TRTSDT) and not missing(TRTEDT) then TRTDURD=input(TRTEDT,yymmdd10.)-input(TRTSDT,yymmdd10.)+1;44  SAFFL='N'; if not missing(TRTSDT) then SAFFL='Y';45run;4647proc sql;48  create table final as select STUDYID, USUBJID, TRT01A, TRTSDT, TRTSDTM, TRTSTMF, TRTEDT, TRTEDTM, TRTETMF, TRTDURD, SAFFL from result;49quit;50proc export data=final outfile="/app/output/adsl.csv" dbms=csv replace; run;