Derive absolute differential records

1library(yamaa)23adlb <- yamaa_domain("spec.yaml")$output4adlb

1import yamaa23adlb = yamaa.yamaa_domain("spec.yaml").output4adlb

1# Reference solution for the yamaa benchmark adam-adlb-lymphocytes (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)910raw <- read_csv(11  "/app/input/lb.csv",12  col_types = cols(LBSTRESN = col_double(), .default = col_character())13) |>14  rename(PARAMCD = LBTESTCD, AVAL = LBSTRESN, PARAM = LBTEST)1516# Keep every collected record unchanged, with DTYPE empty.17collected <- raw |>18  select(USUBJID, PARAMCD, AVAL, PARAM, VISIT) |>19  mutate(DTYPE = NA_character_)2021# Differential fraction code -> absolute code and parameter name.22differentials <- list(23  LYMLE = c("LYMPH", "Lymphocytes Abs (10^9/L)"),24  NEUTLE = c("NEUT", "Neutrophils Abs (10^9/L)"),25  MONOLE = c("MONO", "Monocytes Abs (10^9/L)"),26  EOSLE = c("EOS", "Eosinophils Abs (10^9/L)"),27  BASOLE = c("BASO", "Basophils Abs (10^9/L)")28)2930wbc <- raw |>31  filter(PARAMCD == "WBC") |>32  select(USUBJID, VISIT, WBCVAL = AVAL)3334# One absolute record per subject and visit for each differential35# fraction present with a WBC count and no absolute record yet.36calc <- bind_rows(lapply(names(differentials), function(frac_code) {37  abs_code <- differentials[[frac_code]][1]38  abs_param <- differentials[[frac_code]][2]39  frac <- raw |>40    filter(PARAMCD == frac_code) |>41    select(USUBJID, VISIT, FRACVAL = AVAL)42  existing <- raw |>43    filter(PARAMCD == abs_code) |>44    distinct(USUBJID, VISIT) |>45    mutate(HAS_ABS = TRUE)46  wbc |>47    inner_join(frac, by = c("USUBJID", "VISIT")) |>48    left_join(existing, by = c("USUBJID", "VISIT")) |>49    filter(!is.na(WBCVAL), !is.na(FRACVAL), is.na(HAS_ABS)) |>50    mutate(51      PARAMCD = abs_code,52      AVAL = WBCVAL * FRACVAL,53      PARAM = abs_param,54      DTYPE = "CALCULATION"55    ) |>56    select(USUBJID, PARAMCD, AVAL, PARAM, VISIT, DTYPE)57}))5859adlb <- bind_rows(collected, calc)6061dir.create("/app/output", showWarnings = FALSE)62write_csv(adlb, "/app/output/adlb.csv", na = "")

1# Reference solution for the yamaa benchmark adam-adlb-lymphocytes (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 pl1011raw = (12    pl.read_csv("/app/input/lb.csv", infer_schema=False)13    .with_columns(pl.col("LBSTRESN").cast(pl.Float64, strict=False))14    .rename({"LBTESTCD": "PARAMCD", "LBSTRESN": "AVAL", "LBTEST": "PARAM"})15)1617# Keep every collected record unchanged, with DTYPE empty.18collected = raw.select("USUBJID", "PARAMCD", "AVAL", "PARAM", "VISIT").with_columns(19    DTYPE=pl.lit(None, dtype=pl.String)20)2122# Differential fraction code -> (absolute code, absolute parameter name).23differentials = {24    "LYMLE": ("LYMPH", "Lymphocytes Abs (10^9/L)"),25    "NEUTLE": ("NEUT", "Neutrophils Abs (10^9/L)"),26    "MONOLE": ("MONO", "Monocytes Abs (10^9/L)"),27    "EOSLE": ("EOS", "Eosinophils Abs (10^9/L)"),28    "BASOLE": ("BASO", "Basophils Abs (10^9/L)"),29}3031wbc = raw.filter(pl.col("PARAMCD") == "WBC").select("USUBJID", "VISIT", WBCVAL="AVAL")3233frames = []34for frac_code, (abs_code, abs_param) in differentials.items():35    frac = (36        raw.filter(pl.col("PARAMCD") == frac_code)37        .select("USUBJID", "VISIT", FRACVAL="AVAL")38    )39    existing = (40        raw.filter(pl.col("PARAMCD") == abs_code)41        .select("USUBJID", "VISIT")42        .with_columns(pl.lit(True).alias("HAS_ABS"))43    )44    calc = (45        wbc.join(frac, on=["USUBJID", "VISIT"], how="inner")46        .join(existing, on=["USUBJID", "VISIT"], how="left")47        .filter(48            pl.col("WBCVAL").is_not_null()49            & pl.col("FRACVAL").is_not_null()50            & pl.col("HAS_ABS").is_null()51        )52        .with_columns(53            PARAMCD=pl.lit(abs_code),54            AVAL=pl.col("WBCVAL") * pl.col("FRACVAL"),55            PARAM=pl.lit(abs_param),56            DTYPE=pl.lit("CALCULATION"),57        )58        .select("USUBJID", "PARAMCD", "AVAL", "PARAM", "VISIT", "DTYPE")59    )60    frames.append(calc)6162adlb = pl.concat([collected, *frames], how="diagonal")6364Path("/app/output").mkdir(exist_ok=True)65adlb.write_csv("/app/output/adlb.csv")

1/* opensas reference solution for the yamaa benchmark adam-adlb-lymphocytes.2   Reads the benchmark inputs and derives the requested output.3   Execute with the existing opensas runtime: opensas /app/output/result.sas. */45data lb;6  length USUBJID LBTESTCD LBTEST VISIT $1024;7  infile "/app/input/lb.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;8  input USUBJID : $char1024. LBTESTCD : $char1024. LBSTRESN LBTEST : $char1024. VISIT : $char1024.;9run;1011proc sql; create table derived as12select a.USUBJID,'LYMPH' as PARAMCD,a.LBSTRESN*b.LBSTRESN as AVAL,'Lymphocytes Abs (10^9/L)' as PARAM,a.VISIT,'CALCULATION' as DTYPE from lb as a inner join lb as b on a.USUBJID=b.USUBJID and a.VISIT=b.VISIT left join lb as c on a.USUBJID=c.USUBJID and a.VISIT=c.VISIT and c.LBTESTCD='LYMPH' where a.LBTESTCD='LYMLE' and b.LBTESTCD='WBC' and not missing(a.LBSTRESN) and not missing(b.LBSTRESN) and missing(c.LBTESTCD)13union all14select a.USUBJID,'NEUT' as PARAMCD,a.LBSTRESN*b.LBSTRESN as AVAL,'Neutrophils Abs (10^9/L)' as PARAM,a.VISIT,'CALCULATION' as DTYPE from lb as a inner join lb as b on a.USUBJID=b.USUBJID and a.VISIT=b.VISIT left join lb as c on a.USUBJID=c.USUBJID and a.VISIT=c.VISIT and c.LBTESTCD='NEUT' where a.LBTESTCD='NEUTLE' and b.LBTESTCD='WBC' and not missing(a.LBSTRESN) and not missing(b.LBSTRESN) and missing(c.LBTESTCD)15union all16select a.USUBJID,'MONO' as PARAMCD,a.LBSTRESN*b.LBSTRESN as AVAL,'Monocytes Abs (10^9/L)' as PARAM,a.VISIT,'CALCULATION' as DTYPE from lb as a inner join lb as b on a.USUBJID=b.USUBJID and a.VISIT=b.VISIT left join lb as c on a.USUBJID=c.USUBJID and a.VISIT=c.VISIT and c.LBTESTCD='MONO' where a.LBTESTCD='MONOLE' and b.LBTESTCD='WBC' and not missing(a.LBSTRESN) and not missing(b.LBSTRESN) and missing(c.LBTESTCD)17union all18select a.USUBJID,'EOS' as PARAMCD,a.LBSTRESN*b.LBSTRESN as AVAL,'Eosinophils Abs (10^9/L)' as PARAM,a.VISIT,'CALCULATION' as DTYPE from lb as a inner join lb as b on a.USUBJID=b.USUBJID and a.VISIT=b.VISIT left join lb as c on a.USUBJID=c.USUBJID and a.VISIT=c.VISIT and c.LBTESTCD='EOS' where a.LBTESTCD='EOSLE' and b.LBTESTCD='WBC' and not missing(a.LBSTRESN) and not missing(b.LBSTRESN) and missing(c.LBTESTCD)19union all20select a.USUBJID,'BASO' as PARAMCD,a.LBSTRESN*b.LBSTRESN as AVAL,'Basophils Abs (10^9/L)' as PARAM,a.VISIT,'CALCULATION' as DTYPE from lb as a inner join lb as b on a.USUBJID=b.USUBJID and a.VISIT=b.VISIT left join lb as c on a.USUBJID=c.USUBJID and a.VISIT=c.VISIT and c.LBTESTCD='BASO' where a.LBTESTCD='BASOLE' and b.LBTESTCD='WBC' and not missing(a.LBSTRESN) and not missing(b.LBSTRESN) and missing(c.LBTESTCD); quit;21data collected; length PARAMCD $16 PARAM $1024 DTYPE $11; set lb; PARAMCD=LBTESTCD; PARAM=LBTEST; AVAL=LBSTRESN; run;22data result; set collected derived; run;2324proc sql;25  create table final as select USUBJID, PARAMCD, AVAL, PARAM, VISIT, DTYPE from result;26quit;27proc export data=final outfile="/app/output/adlb.csv" dbms=csv replace; run;