Disposition from Several Source Forms

1library(yamaa)23ds <- yamaa_domain("spec.yaml")$output4ds

1import yamaa23ds = yamaa.yamaa_domain("spec.yaml").output4ds

1# Reference solution for the yamaa benchmark sdtm-ds-multi-form-disposition (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)1516wide <- odm |>17  group_by(StudyOID, SubjectKey, StudyEventOID) |>18  summarise(19    DTC = Value[ItemOID == "IT.DS.DTC"][1],20    COMP = Value[ItemOID == "IT.DS.COMP"][1],21    REASON = Value[ItemOID == "IT.DS.REASON"][1],22    REASONCD = Value[ItemOID == "IT.DS.REASONCD"][1],23    .groups = "drop"24  )2526ds <- wide |>27  rowwise() |>28  mutate(29    DSTERM = case_when(30      StudyEventOID == "CONSENT" ~ "INFORMED CONSENT OBTAINED",31      StudyEventOID == "RAND" ~ "RANDOMIZED",32      StudyEventOID == "EOT" & COMP == "COMPLETED" ~ "COMPLETED",33      StudyEventOID == "EOT" ~ REASON,34      StudyEventOID == "EOS" & COMP == "COMPLETED" ~ "COMPLETED",35      StudyEventOID == "EOS" & COMP == "SCREEN FAILURE" ~ "SCREEN FAILURE",36      StudyEventOID == "EOS" ~ REASON,37      .default = NA_character_38    ),39    DSDECOD = case_when(40      StudyEventOID == "CONSENT" ~ "INFORMED CONSENT OBTAINED",41      StudyEventOID == "RAND" ~ "RANDOMIZED",42      StudyEventOID == "EOT" & COMP == "COMPLETED" ~ "COMPLETED",43      StudyEventOID == "EOT" ~ REASONCD,44      StudyEventOID == "EOS" & COMP == "COMPLETED" ~ "COMPLETED",45      StudyEventOID == "EOS" & COMP == "SCREEN FAILURE" ~ "SCREEN FAILURE",46      StudyEventOID == "EOS" ~ REASONCD,47      .default = NA_character_48    ),49    DSCAT = case_when(50      StudyEventOID %in% c("CONSENT", "RAND") ~ "PROTOCOL MILESTONE",51      StudyEventOID %in% c("EOT", "EOS") ~ "DISPOSITION EVENT",52      .default = NA_character_53    ),54    DSSCAT = case_when(55      StudyEventOID == "CONSENT" ~ "INFORMED CONSENT",56      StudyEventOID == "RAND" ~ "RANDOMIZATION",57      StudyEventOID == "EOT" ~ "END OF TREATMENT",58      StudyEventOID == "EOS" ~ "END OF STUDY",59      .default = NA_character_60    )61  ) |>62  ungroup() |>63  filter(StudyEventOID %in% c("CONSENT", "RAND", "EOT", "EOS")) |>64  mutate(65    DOMAIN = "DS",66    STUDYID = StudyOID,67    USUBJID = paste(StudyOID, SubjectKey, sep = "-"),68    DSSTDTC = suppressWarnings(as.Date(na_if(DTC, ""))),69    DSTERM = na_if(DSTERM, ""),70    DSDECOD = na_if(DSDECOD, "")71  ) |>72  arrange(USUBJID, DSSTDTC, DSTERM) |>73  group_by(USUBJID) |>74  mutate(DSSEQ = row_number()) |>75  ungroup() |>76  select(DOMAIN, STUDYID, USUBJID, DSSEQ, DSTERM, DSDECOD, DSCAT, DSSCAT, DSSTDTC)7778dir.create("/app/output", showWarnings = FALSE)79write_csv(ds, "/app/output/ds.csv", na = "")

1# Reference solution for the yamaa benchmark sdtm-ds-multi-form-disposition (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)1213wide = (14    odm.group_by("StudyOID", "SubjectKey", "StudyEventOID")15    .agg(16        pl.col("Value").filter(pl.col("ItemOID") == "IT.DS.DTC").first().alias("DTC"),17        pl.col("Value").filter(pl.col("ItemOID") == "IT.DS.COMP").first().alias("COMP"),18        pl.col("Value").filter(pl.col("ItemOID") == "IT.DS.REASON").first().alias("REASON"),19        pl.col("Value")20        .filter(pl.col("ItemOID") == "IT.DS.REASONCD")21        .first()22        .alias("REASONCD"),23    )24)252627def build_row(event: str, dtc, comp, reason, reasoncd) -> dict | None:28    if event == "CONSENT":29        return {30            "DSTERM": "INFORMED CONSENT OBTAINED",31            "DSDECOD": "INFORMED CONSENT OBTAINED",32            "DSCAT": "PROTOCOL MILESTONE",33            "DSSCAT": "INFORMED CONSENT",34        }35    if event == "RAND":36        return {37            "DSTERM": "RANDOMIZED",38            "DSDECOD": "RANDOMIZED",39            "DSCAT": "PROTOCOL MILESTONE",40            "DSSCAT": "RANDOMIZATION",41        }42    if event == "EOT":43        if comp == "COMPLETED":44            term, decod = "COMPLETED", "COMPLETED"45        else:46            term, decod = reason, reasoncd47        return {48            "DSTERM": term,49            "DSDECOD": decod,50            "DSCAT": "DISPOSITION EVENT",51            "DSSCAT": "END OF TREATMENT",52        }53    if event == "EOS":54        if comp == "COMPLETED":55            term, decod = "COMPLETED", "COMPLETED"56        elif comp == "SCREEN FAILURE":57            term, decod = "SCREEN FAILURE", "SCREEN FAILURE"58        else:59            term, decod = reason, reasoncd60        return {61            "DSTERM": term,62            "DSDECOD": decod,63            "DSCAT": "DISPOSITION EVENT",64            "DSSCAT": "END OF STUDY",65        }66    return None676869records: list[dict] = []70for row in wide.to_dicts():71    study = row["StudyOID"]72    subj = row["SubjectKey"]73    event = row["StudyEventOID"]74    out = build_row(event, row["DTC"], row["COMP"], row["REASON"], row["REASONCD"])75    if out is None:76        continue77    usubjid = f"{study}-{subj}"78    records.append(79        {80            "STUDYID": study,81            "USUBJID": usubjid,82            "DSTERM": out["DSTERM"] if out["DSTERM"] != "" else None,83            "DSDECOD": out["DSDECOD"] if out["DSDECOD"] != "" else None,84            "DSCAT": out["DSCAT"],85            "DSSCAT": out["DSSCAT"],86            "DSSTDTC": row["DTC"] if row["DTC"] != "" else None,87        }88    )8990ds = pl.DataFrame(91    records,92    schema={93        "STUDYID": pl.String,94        "USUBJID": pl.String,95        "DSTERM": pl.String,96        "DSDECOD": pl.String,97        "DSCAT": pl.String,98        "DSSCAT": pl.String,99        "DSSTDTC": pl.String,100    },101).with_columns(102    DOMAIN=pl.lit("DS"),103    DSSTDTC_DT=pl.col("DSSTDTC").str.to_date(strict=False),104)105106ds = (107    ds.sort(["USUBJID", "DSSTDTC_DT", "DSTERM"])108    .with_columns(109        DSSEQ=pl.col("USUBJID").cum_count().over("USUBJID"),110    )111    .with_columns(DSSEQ=pl.col("DSSEQ").cast(pl.Int64))112    .select(113        "DOMAIN",114        "STUDYID",115        "USUBJID",116        "DSSEQ",117        "DSTERM",118        "DSDECOD",119        "DSCAT",120        "DSSCAT",121        "DSSTDTC",122    )123)124125Path("/app/output").mkdir(exist_ok=True)126ds.write_csv("/app/output/ds.csv")

1/* opensas reference solution for the yamaa benchmark sdtm-ds-multi-form-disposition.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; run;12data wide;13  length COMP DTC REASON REASONCD $1024;14  retain COMP DTC REASON REASONCD;15  set ordered;1617  by StudyOID SubjectKey StudyEventOID;18  if first.StudyEventOID then do;19    COMP=''; DTC=''; REASON=''; REASONCD='';20  end;21    if ItemOID='IT.DS.COMP' then COMP=Value;22  if ItemOID='IT.DS.DTC' then DTC=Value;23  if ItemOID='IT.DS.REASON' then REASON=Value;24  if ItemOID='IT.DS.REASONCD' then REASONCD=Value;25  if last.StudyEventOID then output;26run;27data prepared; set wide; length DOMAIN $2 STUDYID USUBJID DSTERM DSDECOD DSCAT DSSCAT DSSTDTC $1024;28DOMAIN='DS'; STUDYID=StudyOID; USUBJID=SubjectKey; DSCAT='DISPOSITION EVENT'; DSSTDTC=DTC; DSTERM=REASON; DSDECOD=REASONCD;29if COMP='COMPLETED' then do; DSTERM='COMPLETED'; DSDECOD='COMPLETED'; end;30USUBJID=catx('-',StudyOID,SubjectKey);31if StudyEventOID='CONSENT' then do; DSTERM='INFORMED CONSENT OBTAINED'; DSDECOD=DSTERM; DSCAT='PROTOCOL MILESTONE'; DSSCAT='INFORMED CONSENT'; end;32if StudyEventOID='RAND' then do; DSTERM='RANDOMIZED'; DSDECOD=DSTERM; DSCAT='PROTOCOL MILESTONE'; DSSCAT='RANDOMIZATION'; end;33if StudyEventOID='EOT' then DSSCAT='END OF TREATMENT';34if StudyEventOID='EOS' then do; DSSCAT='END OF STUDY'; if COMP='SCREEN FAILURE' then do; DSTERM='SCREEN FAILURE'; DSDECOD=DSTERM; end; end; run;35proc sort data=prepared; by STUDYID USUBJID DSSTDTC DSTERM; run;36data result; set prepared; by STUDYID USUBJID; retain DSSEQ;37  if first.USUBJID then DSSEQ=0;38  DSSEQ+1;39run;4041proc sql;42  create table final as select DOMAIN, STUDYID, USUBJID, DSSEQ, DSTERM, DSDECOD, DSCAT, DSSCAT, DSSTDTC from result;43quit;44proc export data=final outfile="/app/output/ds.csv" dbms=csv replace; run;