Planned Time Points Around Dosing

1library(yamaa)23vs <- yamaa_domain("spec.yaml")$output4vs

1import yamaa23vs = yamaa.yamaa_domain("spec.yaml").output4vs

1# Reference solution for the yamaa benchmark sdtm-vs-planned-time-points (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(lubridate)9library(readr)10library(tidyr)1112study_day <- function(ref, coll) {13  if (is.na(ref) || is.na(coll)) {14    return(NA_integer_)15  }16  delta <- as.integer(coll - ref)17  if (delta >= 0L) delta + 1L else delta18}1920time_point <- function(form) {21  name <- toupper(form)22  if (grepl("PREDOSE", name, fixed = TRUE)) {23    list(vstpt = "PRE-DOSE", vstptnum = 1L, vseltm = "-PT15M")24  } else if (grepl("30MIN", name, fixed = TRUE)) {25    list(vstpt = "30 MIN POST-DOSE", vstptnum = 2L, vseltm = "PT30M")26  } else if (grepl("1H", name, fixed = TRUE)) {27    list(vstpt = "1 H POST-DOSE", vstptnum = 3L, vseltm = "PT1H")28  } else if (grepl("4H", name, fixed = TRUE)) {29    list(vstpt = "4 H POST-DOSE", vstptnum = 4L, vseltm = "PT4H")30  } else {31    NULL32  }33}3435dm <- read_csv(36  "/app/input/dm.csv",37  col_types = cols(.default = col_character())38)39odm <- read_csv(40  "/app/input/odm.csv",41  col_types = cols(.default = col_character())42)43mapping <- read_csv(44  "/app/input/vs_mapping.csv",45  col_types = cols(.default = col_character())46)4748ref_map <- setNames(suppressWarnings(ymd(dm$RFSTDTC)), dm$USUBJID)49test_map <- setNames(mapping$VSTESTCD, mapping$ItemOID)50unit_map <- setNames(mapping$UNIT, mapping$ItemOID)5152forms <- odm |>53  filter(StudyEventOID == "SE.D1", grepl("^FO\\.VS_", FormOID)) |>54  group_by(SubjectKey, FormOID, StudyOID) |>55  summarise(56    vsdtc = Value[ItemOID == "IT.VS.VSDTC"][1],57    .groups = "drop"58  )5960results <- odm |>61  filter(62    StudyEventOID == "SE.D1",63    grepl("^FO\\.VS_", FormOID),64    ItemOID %in% names(test_map),65    !is.na(Value) & Value != ""66  ) |>67  left_join(forms, by = c("SubjectKey", "FormOID", "StudyOID")) |>68  rowwise() |>69  mutate(70    point = list(time_point(FormOID)),71    VSTPT = point$vstpt,72    VSTPTNUM = point$vstptnum,73    VSELTM = point$vseltm,74    VSTESTCD = test_map[[ItemOID]],75    VSORRES = Value,76    VSORRESU = unit_map[[ItemOID]],77    VSSTRESN = suppressWarnings(as.numeric(Value)),78    VSSTRESU = VSORRESU,79    VSDTC = vsdtc,80    ref = ref_map[[SubjectKey]],81    VSDY = study_day(ref, suppressWarnings(as_date(ymd_hms(VSDTC)))),82    test_ord = case_when(83      VSTESTCD == "PULSE" ~ 1L,84      VSTESTCD == "SYSBP" ~ 2L,85      .default = 3L86    )87  ) |>88  ungroup() |>89  filter(!is.na(VSTPT), !is.na(VSDTC) & VSDTC != "") |>90  arrange(SubjectKey, VSTPTNUM, test_ord) |>91  group_by(SubjectKey) |>92  mutate(VSSEQ = row_number()) |>93  ungroup() |>94  arrange(SubjectKey, VSSEQ) |>95  transmute(96    DOMAIN = "VS",97    STUDYID = StudyOID,98    USUBJID = SubjectKey,99    VSSEQ,100    VISIT = "DAY 1",101    VSTPT, VSTPTNUM, VSELTM, VSTESTCD, VSORRES, VSORRESU, VSSTRESN, VSSTRESU,102    VSDTC, VSDY103  )104105dir.create("/app/output", showWarnings = FALSE)106write_csv(results, "/app/output/vs.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-vs-planned-time-points (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 datetime import date, datetime8from pathlib import Path910import polars as pl111213def parse_date(text: str | None) -> date | None:14    if text is None or text == "":15        return None16    try:17        return datetime.fromisoformat(text).date()18    except ValueError:19        try:20            return date.fromisoformat(text[:10])21        except ValueError:22            return None232425def study_day(ref: date | None, coll: date | None) -> int | None:26    if ref is None or coll is None:27        return None28    delta = (coll - ref).days29    return delta + 1 if delta >= 0 else delta303132def time_point(form: str) -> tuple[str, int, str] | None:33    # The planned time point for each collection form, in schedule order.34    name = form.upper()35    if "PREDOSE" in name:36        return ("PRE-DOSE", 1, "-PT15M")37    if "30MIN" in name:38        return ("30 MIN POST-DOSE", 2, "PT30M")39    if "1H" in name:40        return ("1 H POST-DOSE", 3, "PT1H")41    if "4H" in name:42        return ("4 H POST-DOSE", 4, "PT4H")43    return None444546dm = pl.read_csv("/app/input/dm.csv", infer_schema=False)47odm = pl.read_csv("/app/input/odm.csv", infer_schema=False)48mapping = pl.read_csv("/app/input/vs_mapping.csv", infer_schema=False)4950ref_map = {row["USUBJID"]: parse_date(row["RFSTDTC"]) for row in dm.to_dicts()}51test_map = {52    row["ItemOID"]: (row["VSTESTCD"], row["UNIT"]) for row in mapping.to_dicts()53}54# Within each time point: pulse, systolic, then diastolic.55test_order = {"PULSE": 1, "SYSBP": 2, "DIABP": 3}5657# One form per subject per planned time point: its datetime and results.58forms: dict[tuple[str, str], dict] = {}59for row in odm.to_dicts():60    if row["StudyEventOID"] != "SE.D1":61        continue62    if not row["FormOID"].startswith("FO.VS_"):63        continue64    key = (row["SubjectKey"], row["FormOID"])65    entry = forms.setdefault(66        key,67        {68            "subject": row["SubjectKey"],69            "study": row["StudyOID"],70            "form": row["FormOID"],71            "dt": None,72            "results": {},73        },74    )75    if row["ItemOID"] == "IT.VS.VSDTC":76        entry["dt"] = row["Value"]77    elif row["ItemOID"] in test_map:78        entry["results"][row["ItemOID"]] = row["Value"]7980records = []81for entry in forms.values():82    point = time_point(entry["form"])83    if point is None or entry["dt"] in (None, ""):84        continue85    vstpt, vstptnum, vseltm = point86    coll = parse_date(entry["dt"])87    vsdy = study_day(ref_map.get(entry["subject"]), coll)88    for item, value in entry["results"].items():89        if value is None or value == "":90            continue91        testcd, unit = test_map[item]92        records.append(93            {94                "DOMAIN": "VS",95                "STUDYID": entry["study"],96                "USUBJID": entry["subject"],97                "VISIT": "DAY 1",98                "VSTPT": vstpt,99                "VSTPTNUM": vstptnum,100                "VSELTM": vseltm,101                "VSTESTCD": testcd,102                "_tord": test_order.get(testcd, 9),103                "VSORRES": value,104                "VSORRESU": unit,105                "VSSTRESN": float(value),106                "VSSTRESU": unit,107                "VSDTC": entry["dt"],108                "VSDY": vsdy,109            }110        )111112vs = (113    pl.DataFrame(records)114    .sort(["USUBJID", "VSTPTNUM", "_tord"])115    .with_columns(116        VSSEQ=pl.int_range(1, pl.len() + 1).over("USUBJID"),117    )118    .sort(["USUBJID", "VSSEQ"])119    .select(120        "DOMAIN",121        "STUDYID",122        "USUBJID",123        "VSSEQ",124        "VISIT",125        "VSTPT",126        "VSTPTNUM",127        "VSELTM",128        "VSTESTCD",129        "VSORRES",130        "VSORRESU",131        "VSSTRESN",132        "VSSTRESU",133        "VSDTC",134        "VSDY",135    )136)137138Path("/app/output").mkdir(exist_ok=True)139vs.write_csv("/app/output/vs.csv")

1/* opensas reference solution for the yamaa benchmark sdtm-vs-planned-time-points.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 RFSTDTC $1024;7  infile "/app/input/dm.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;8  input STUDYID : $char1024. USUBJID : $char1024. RFSTDTC : $char1024.;9run;1011data odm;12  length StudyOID MetaDataVersionOID SubjectKey StudyEventOID StudyEventRepeatKey FormOID FormRepeatKey ItemGroupOID ItemGroupRepeatKey ItemOID Value $1024;13  infile "/app/input/odm.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;14  input StudyOID : $char1024. MetaDataVersionOID : $char1024. SubjectKey : $char1024. StudyEventOID : $char1024. StudyEventRepeatKey : $char1024. FormOID : $char1024. FormRepeatKey : $char1024. ItemGroupOID : $char1024. ItemGroupRepeatKey : $char1024. ItemOID : $char1024. Value : $char1024.;15run;1617data vs_mapping;18  length ItemOID VSTESTCD VSTEST UNIT $1024;19  infile "/app/input/vs_mapping.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;20  input ItemOID : $char1024. VSTESTCD : $char1024. VSTEST : $char1024. UNIT : $char1024.;21run;2223proc sort data=odm out=ordered; by StudyOID SubjectKey StudyEventOID StudyEventRepeatKey FormOID FormRepeatKey; run;24data dates;25  length VSDTC $1024;26  retain VSDTC;27  set ordered;28  where ItemGroupOID='IG.VS';29  by StudyOID SubjectKey StudyEventOID StudyEventRepeatKey FormOID FormRepeatKey;30  if first.FormRepeatKey then do;31    VSDTC='';32  end;33    if ItemOID='IT.VS.VSDTC' then VSDTC=Value;34  if last.FormRepeatKey then output;35run;36proc sql; create table joined as select a.*,b.VSDTC,c.VSTESTCD,c.UNIT,d.RFSTDTC from odm a inner join dates b on a.StudyOID=b.StudyOID and a.SubjectKey=b.SubjectKey and a.StudyEventOID=b.StudyEventOID and a.StudyEventRepeatKey=b.StudyEventRepeatKey and a.FormOID=b.FormOID and a.FormRepeatKey=b.FormRepeatKey inner join vs_mapping c on a.ItemOID=c.ItemOID left join dm d on a.StudyOID=d.STUDYID and a.SubjectKey=d.USUBJID; quit;37data prepared; set joined; length DOMAIN $2 STUDYID USUBJID VISIT VSTPT VSELTM VSORRES VSORRESU VSSTRESU $1024; DOMAIN='VS'; STUDYID=StudyOID; USUBJID=SubjectKey; VISIT='DAY 1'; VSTPT=''; VSELTM=''; VSTPTNUM=.;38if FormOID='FO.VS_PREDOSE' then do; VSTPT='PRE-DOSE'; VSTPTNUM=1; VSELTM='-PT15M'; end;39if FormOID='FO.VS_30MIN' then do; VSTPT='30 MIN POST-DOSE'; VSTPTNUM=2; VSELTM='PT30M'; end;40if FormOID='FO.VS_1H' then do; VSTPT='1 H POST-DOSE'; VSTPTNUM=3; VSELTM='PT1H'; end;41if FormOID='FO.VS_4H' then do; VSTPT='4 H POST-DOSE'; VSTPTNUM=4; VSELTM='PT4H'; end;42VSORRES=Value; VSORRESU=UNIT; VSSTRESU=UNIT; VSSTRESN=input(Value,best32.); VSDY=.; if lengthn(VSDTC)>=10 and lengthn(RFSTDTC)>=10 then do; VSDY=input(substr(VSDTC,1,10),yymmdd10.)-input(substr(RFSTDTC,1,10),yymmdd10.); if VSDY>=0 then VSDY+1; end;43TESTORD=3; if VSTESTCD='PULSE' then TESTORD=1; if VSTESTCD='SYSBP' then TESTORD=2; run;44proc sort data=prepared; by STUDYID USUBJID VSTPTNUM TESTORD; run;45data result; set prepared; by STUDYID USUBJID; retain VSSEQ;46  if first.USUBJID then VSSEQ=0;47  VSSEQ+1;48run;4950proc sql;51  create table final as select DOMAIN, STUDYID, USUBJID, VSSEQ, VISIT, VSTPT, VSTPTNUM, VSELTM, VSTESTCD, VSORRES, VSORRESU, VSSTRESN, VSSTRESU, VSDTC, VSDY from result;52quit;53proc export data=final outfile="/app/output/vs.csv" dbms=csv replace; run;