Derive the Last Known Alive Date from Multiple Sources

1library(yamaa)23adsl <- yamaa_domain("spec.yaml")$output4adsl

1import yamaa23adsl = yamaa.yamaa_domain("spec.yaml").output4adsl

1# Reference solution for the yamaa benchmark adam-adsl-alive-date (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)910adsl_raw <- read_csv(11  "/app/input/adsl.csv",12  col_types = cols(TRTEDT = col_date(), .default = col_character())13)14adae <- read_csv(15  "/app/input/adae.csv",16  col_types = cols(AENDT = col_date(), .default = col_character())17)18advs <- read_csv(19  "/app/input/advs.csv",20  col_types = cols(ADATE = col_date(), .default = col_character())21)2223# The contact text completed to a day: a full date stands, a year and month24# take the month's first day, a year alone takes January first, and anything25# else leaves the date missing.26adsl_raw <- adsl_raw |>27  mutate(28    .full = suppressWarnings(as.Date(LSTCNTDC, format = "%Y-%m-%d")),29    .ym = suppressWarnings(as.Date(paste0(LSTCNTDC, "-01"), format = "%Y-%m-%d")),30    .y = suppressWarnings(as.Date(paste0(LSTCNTDC, "-01-01"), format = "%Y-%m-%d")),31    LSTCNTDT = coalesce(.full, .ym, .y)32  )3334# Each subject's latest adverse event end and vital signs dates.35adae_last <- adae |>36  group_by(STUDYID, USUBJID) |>37  summarise(38    ADAE_LST = if (all(is.na(AENDT))) as.Date(NA) else max(AENDT, na.rm = TRUE),39    .groups = "drop"40  )41advs_last <- advs |>42  group_by(STUDYID, USUBJID) |>43  summarise(44    ADVS_LST = if (all(is.na(ADATE))) as.Date(NA) else max(ADATE, na.rm = TRUE),45    .groups = "drop"46  )4748adsl <- adsl_raw |>49  left_join(adae_last, by = c("STUDYID", "USUBJID")) |>50  left_join(advs_last, by = c("STUDYID", "USUBJID")) |>51  mutate(52    LSTALVDT = suppressWarnings(do.call(53      pmax,54      list(TRTEDT, LSTCNTDT, ADAE_LST, ADVS_LST, na.rm = TRUE)55    ))56  ) |>57  mutate(58    LSTALVDT = if_else(59      is.na(TRTEDT) & is.na(LSTCNTDT) & is.na(ADAE_LST) & is.na(ADVS_LST),60      as.Date(NA),61      LSTALVDT62    )63  ) |>64  select(STUDYID, USUBJID, TRTEDT, LSTCNTDC, LSTCNTDT, LSTALVDT)6566dir.create("/app/output", showWarnings = FALSE)67write_csv(adsl, "/app/output/adsl.csv", na = "")

1# Reference solution for the yamaa benchmark adam-adsl-alive-date (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 1.67from pathlib import Path89import polars as pl1011adsl_raw = pl.read_csv("/app/input/adsl.csv", infer_schema=False).with_columns(12    pl.col("TRTEDT").str.to_date("%Y-%m-%d", strict=False)13)14adae = pl.read_csv("/app/input/adae.csv", infer_schema=False).with_columns(15    pl.col("AENDT").str.to_date("%Y-%m-%d", strict=False)16)17advs = pl.read_csv("/app/input/advs.csv", infer_schema=False).with_columns(18    pl.col("ADATE").str.to_date("%Y-%m-%d", strict=False)19)2021# The contact text completed to a day: a full date stands, a year and month22# take the month's first day, a year alone takes January first, and anything23# else leaves the date missing.24adsl_raw = adsl_raw.with_columns(25    LSTCNTDT=pl.coalesce(26        pl.col("LSTCNTDC").str.to_date("%Y-%m-%d", strict=False),27        (pl.col("LSTCNTDC") + "-01").str.to_date("%Y-%m-%d", strict=False),28        (pl.col("LSTCNTDC") + "-01-01").str.to_date("%Y-%m-%d", strict=False),29    )30)3132# Each subject's latest adverse event end and vital signs dates.33adae_last = adae.group_by(["STUDYID", "USUBJID"]).agg(34    ADAE_LST=pl.col("AENDT").max()35)36advs_last = advs.group_by(["STUDYID", "USUBJID"]).agg(37    ADVS_LST=pl.col("ADATE").max()38)3940adsl = (41    adsl_raw.join(adae_last, on=["STUDYID", "USUBJID"], how="left", maintain_order="left")42    .join(advs_last, on=["STUDYID", "USUBJID"], how="left", maintain_order="left")43    .with_columns(44        LSTALVDT=pl.max_horizontal("TRTEDT", "LSTCNTDT", "ADAE_LST", "ADVS_LST")45    )46    .select("STUDYID", "USUBJID", "TRTEDT", "LSTCNTDC", "LSTCNTDT", "LSTALVDT")47)4849Path("/app/output").mkdir(exist_ok=True)50adsl.write_csv("/app/output/adsl.csv")

1/* opensas reference solution for the yamaa benchmark adam-adsl-alive-date.2   Reads the benchmark inputs and derives the requested output.3   Execute with the existing opensas runtime: opensas /app/output/result.sas. */45data adae;6  length STUDYID USUBJID AENDT $1024;7  infile "/app/input/adae.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;8  input STUDYID : $char1024. USUBJID : $char1024. AENDT : $char1024.;9run;1011data adsl;12  length STUDYID USUBJID TRTEDT LSTCNTDC $1024;13  infile "/app/input/adsl.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;14  input STUDYID : $char1024. USUBJID : $char1024. TRTEDT : $char1024. LSTCNTDC : $char1024.;15run;1617data advs;18  length STUDYID USUBJID ADATE $1024;19  infile "/app/input/advs.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;20  input STUDYID : $char1024. USUBJID : $char1024. ADATE : $char1024.;21run;2223proc sql;24  create table ends as select STUDYID,USUBJID,max(AENDT) as AENDT from adae group by STUDYID,USUBJID;25  create table vitals as select STUDYID,USUBJID,max(ADATE) as ADATE from advs group by STUDYID,USUBJID;26  create table joined as select a.*,b.AENDT,c.ADATE from adsl as a left join ends as b on a.STUDYID=b.STUDYID and a.USUBJID=b.USUBJID left join vitals as c on a.STUDYID=c.STUDYID and a.USUBJID=c.USUBJID;27quit;28data result;29  length LSTCNTDT LSTALVDT $10 completed $10;30  set joined;31  if prxmatch('/^[0-9]{4}(-[0-9]{2}(-[0-9]{2})?)?$/',strip(LSTCNTDC)) then do;32    completed=LSTCNTDC;33    if lengthn(LSTCNTDC)=4 then completed=cats(LSTCNTDC,'-01-01');34    if lengthn(LSTCNTDC)=7 then completed=cats(LSTCNTDC,'-01');35    if not missing(input(completed,yymmdd10.)) then LSTCNTDT=completed;36  end;37  lastdate=max(input(TRTEDT,yymmdd10.),input(LSTCNTDT,yymmdd10.),input(AENDT,yymmdd10.),input(ADATE,yymmdd10.));38  if not missing(lastdate) then LSTALVDT=put(lastdate,yymmdd10.);39run;4041proc sql;42  create table final as select STUDYID, USUBJID, TRTEDT, LSTCNTDC, LSTCNTDT, LSTALVDT from result;43quit;44proc export data=final outfile="/app/output/adsl.csv" dbms=csv replace; run;