Repeated Concomitant Medications from ODM

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

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

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 1.67import csv8from collections import OrderedDict910INPUT = "/app/input/odm.csv"11OUTPUT = "/app/output/cm.csv"1213# Visit ordering: screening before baseline; any other visit follows in14# order of first appearance in the data.15VISIT_ORDER = {"SCREENING": 0, "BASELINE": 1}1617with open(INPUT, newline="", encoding="utf-8") as f:18    rows = list(csv.DictReader(f))1920visit_seen = {}21for row in rows:22    v = row["StudyEventOID"]23    if v not in visit_seen:24        visit_seen[v] = len(visit_seen)252627def visit_key(visit):28    if visit in VISIT_ORDER:29        return (0, VISIT_ORDER[visit])30    return (1, visit_seen[visit])313233# Group IG.CM items into courses keyed by subject, visit, visit repeat,34# and form repeat.35courses = OrderedDict()36for row in rows:37    if row["ItemGroupOID"] != "IG.CM":38        continue39    key = (40        row["StudyOID"],41        row["SubjectKey"],42        row["StudyEventOID"],43        int(row["StudyEventRepeatKey"]),44        int(row["ItemGroupRepeatKey"]),45    )46    item = row["ItemOID"].rsplit(".", 1)[-1]47    courses.setdefault(key, {})[item] = row["Value"]4849records = []50for (studyid, usubjid, visit, visit_rep, form_rep), items in courses.items():51    cmtrt = items.get("CMTRT", "")52    if not cmtrt:53        # A course with no reported treatment name contributes no record.54        continue55    records.append(56        {57            "STUDYID": studyid,58            "USUBJID": usubjid,59            "CMTRT": cmtrt,60            "CMSTDTC": items.get("CMSTDTC", ""),61            "CMENDTC": items.get("CMENDTC", ""),62            "CMROUTE": items.get("CMROUTE", ""),63            "CMINDC": items.get("CMINDC", ""),64            "_sort": (usubjid, visit_key(visit), visit_rep, form_rep),65        }66    )6768records.sort(key=lambda r: r["_sort"])6970out = []71seq_by_subject = {}72for rec in records:73    seq_by_subject[rec["USUBJID"]] = seq_by_subject.get(rec["USUBJID"], 0) + 174    out.append(75        {76            "DOMAIN": "CM",77            "STUDYID": rec["STUDYID"],78            "USUBJID": rec["USUBJID"],79            "CMSEQ": seq_by_subject[rec["USUBJID"]],80            "CMTRT": rec["CMTRT"],81            "CMSTDTC": rec["CMSTDTC"],82            "CMENDTC": rec["CMENDTC"],83            "CMROUTE": rec["CMROUTE"],84            "CMINDC": rec["CMINDC"],85        }86    )8788with open(OUTPUT, "w", newline="", encoding="utf-8") as f:89    writer = csv.DictWriter(90        f,91        fieldnames=[92            "DOMAIN",93            "STUDYID",94            "USUBJID",95            "CMSEQ",96            "CMTRT",97            "CMSTDTC",98            "CMENDTC",99            "CMROUTE",100            "CMINDC",101        ],102    )103    writer.writeheader()104    writer.writerows(out)

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;