Repeated Concomitant Medications from ODM

1library(yamaa)23cm <- yamaa_domain("spec.yaml")$output4cm

1import yamaa23cm = yamaa.yamaa_domain("spec.yaml").output4cm

1# Reference solution for the yamaa benchmark sdtm-cm-odm-repeated (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)1011odm <- read_csv(12  "/app/input/odm.csv",13  col_types = cols(.default = col_character())14)1516# One course per form repeat within a visit repeat: subject, visit, visit17# repeat, and form repeat. A course without a reported treatment name gives18# no record; a repeat number reused at a later visit starts a new course.19courses <- odm |>20  filter(ItemGroupOID == "IG.CM") |>21  pivot_wider(22    id_cols = c(23      StudyOID, SubjectKey, StudyEventOID, StudyEventRepeatKey,24      ItemGroupRepeatKey25    ),26    names_from = ItemOID,27    values_from = Value,28    values_fn = first29  ) |>30  mutate(31    CMTRT = `IT.CM.CMTRT`,32    CMSTDTC = `IT.CM.CMSTDTC`,33    CMENDTC = `IT.CM.CMENDTC`,34    CMROUTE = `IT.CM.CMROUTE`,35    CMINDC = `IT.CM.CMINDC`36  ) |>37  filter(!is.na(CMTRT) & CMTRT != "") |>38  # Courses run by visit (screening before baseline), visit repeat, and39  # form repeat; the same medication in two courses stays two records.40  mutate(41    VISITORD = case_when(42      StudyEventOID == "SCREENING" ~ 0L,43      StudyEventOID == "BASELINE" ~ 1L,44      .default = 2L45    ),46    VISITREP = as.integer(StudyEventRepeatKey),47    IGREP = as.integer(ItemGroupRepeatKey)48  ) |>49  arrange(SubjectKey, VISITORD, VISITREP, IGREP) |>50  group_by(SubjectKey) |>51  mutate(CMSEQ = row_number()) |>52  ungroup() |>53  mutate(DOMAIN = "CM", STUDYID = StudyOID, USUBJID = SubjectKey) |>54  select(55    DOMAIN, STUDYID, USUBJID, CMSEQ, CMTRT,56    CMSTDTC, CMENDTC, CMROUTE, CMINDC57  )5859dir.create("/app/output", showWarnings = FALSE)60write_csv(courses, "/app/output/cm.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-cm-odm-repeated (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 pl1011odm = pl.read_csv("/app/input/odm.csv", infer_schema=False)1213# One course per form repeat within a visit repeat: subject, visit, visit14# repeat, and form repeat. A course without a reported treatment name gives15# no record; a repeat number reused at a later visit starts a new course.16courses = (17    odm.filter(pl.col("ItemGroupOID") == "IG.CM")18    .pivot(19        index=[20            "StudyOID",21            "SubjectKey",22            "StudyEventOID",23            "StudyEventRepeatKey",24            "ItemGroupRepeatKey",25        ],26        on="ItemOID",27        values="Value",28        aggregate_function="first",29    )30    .with_columns(31        CMTRT=pl.col("IT.CM.CMTRT"),32        CMSTDTC=pl.col("IT.CM.CMSTDTC"),33        CMENDTC=pl.col("IT.CM.CMENDTC"),34        CMROUTE=pl.col("IT.CM.CMROUTE"),35        CMINDC=pl.col("IT.CM.CMINDC"),36    )37    .filter(pl.col("CMTRT").fill_null("") != "")38    # Courses run by visit (screening before baseline), visit repeat, and39    # form repeat; the same medication in two courses stays two records.40    .with_columns(41        VISITORD=pl.when(pl.col("StudyEventOID") == "SCREENING")42        .then(0)43        .when(pl.col("StudyEventOID") == "BASELINE")44        .then(1)45        .otherwise(2),46        VISITREP=pl.col("StudyEventRepeatKey").cast(pl.Int64),47        IGREP=pl.col("ItemGroupRepeatKey").cast(pl.Int64),48    )49    .sort(["SubjectKey", "VISITORD", "VISITREP", "IGREP"])50    .with_columns(CMSEQ=pl.col("SubjectKey").cum_count().over("SubjectKey"))51    .with_columns(DOMAIN=pl.lit("CM"), STUDYID=pl.col("StudyOID"), USUBJID=pl.col("SubjectKey"))52    .select(53        "DOMAIN",54        "STUDYID",55        "USUBJID",56        "CMSEQ",57        "CMTRT",58        "CMSTDTC",59        "CMENDTC",60        "CMROUTE",61        "CMINDC",62    )63)6465Path("/app/output").mkdir(exist_ok=True)66courses.write_csv("/app/output/cm.csv")

1/* opensas reference solution for the yamaa benchmark sdtm-cm-odm-repeated.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;1011proc sort data=odm out=ordered; by StudyOID SubjectKey StudyEventOID StudyEventRepeatKey ItemGroupRepeatKey; run;12data wide;13  length CMTRT CMSTDTC CMENDTC CMROUTE CMINDC CMONGO $1024;14  retain CMTRT CMSTDTC CMENDTC CMROUTE CMINDC CMONGO;15  set ordered;16  where ItemGroupOID='IG.CM';17  by StudyOID SubjectKey StudyEventOID StudyEventRepeatKey ItemGroupRepeatKey;18  if first.ItemGroupRepeatKey then do;19    CMTRT=''; CMSTDTC=''; CMENDTC=''; CMROUTE=''; CMINDC=''; CMONGO='';20  end;21    if ItemOID='IT.CM.CMTRT' then CMTRT=Value;22  if ItemOID='IT.CM.CMSTDTC' then CMSTDTC=Value;23  if ItemOID='IT.CM.CMENDTC' then CMENDTC=Value;24  if ItemOID='IT.CM.CMROUTE' then CMROUTE=Value;25  if ItemOID='IT.CM.CMINDC' then CMINDC=Value;26  if ItemOID='IT.CM.CMONGO' then CMONGO=Value;27  if last.ItemGroupRepeatKey then output;28run;29data prepared; set wide; if not missing(CMTRT); length DOMAIN $2 STUDYID USUBJID $1024; DOMAIN='CM'; STUDYID=StudyOID; USUBJID=SubjectKey;30if CMONGO='Y' then CMENDTC=''; VISITORD=2; if StudyEventOID='SCREENING' then VISITORD=0; if StudyEventOID='BASELINE' then VISITORD=1;31VISITREP=input(StudyEventRepeatKey,best32.); IGREP=input(ItemGroupRepeatKey,best32.); run;32proc sort data=prepared; by STUDYID USUBJID VISITORD VISITREP IGREP; run;33data result; set prepared; by STUDYID USUBJID; retain CMSEQ;34  if first.USUBJID then CMSEQ=0;35  CMSEQ+1;36run;3738proc sql;39  create table final as select DOMAIN, STUDYID, USUBJID, CMSEQ, CMTRT, CMSTDTC, CMENDTC, CMROUTE, CMINDC from result;40quit;41proc export data=final outfile="/app/output/cm.csv" dbms=csv replace; run;