Demographics from an EDC Snapshot Export

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

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

1# Reference solution for the yamaa benchmark sdtm-dm-odm-snapshot (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(xml2)89odm <- read_xml("/app/input/odm.xml")10# The snapshot's elements are in the ODM 1.3 default namespace; strip it so11# the paths below name the elements plainly.12xml_ns_strip(odm)1314# One subject's current value of one item. An item recorded as null has no15# value; audit records, signatures and comments hold no items.16item_value <- function(subject, oid) {17  item <- xml_find_first(subject, sprintf(".//ItemData[@ItemOID='%s']", oid))18  if (inherits(item, "xml_missing")) {19    return(NA_character_)20  }21  is_null <- xml_attr(item, "IsNull")22  if (!is.na(is_null) && tolower(is_null) %in% c("yes", "true", "1")) {23    return(NA_character_)24  }25  xml_attr(item, "Value")26}2728rows <- lapply(29  xml_find_all(odm, "/ODM/ClinicalData/SubjectData"),30  function(subject) {31    study <- xml_attr(xml_parent(subject), "StudyOID")32    key <- xml_attr(subject, "SubjectKey")33    age <- item_value(subject, "IT.DM.AGE")34    data.frame(35      DOMAIN = "DM",36      STUDYID = study,37      USUBJID = paste0(study, "-", key),38      SUBJID = key,39      BRTHDTC = item_value(subject, "IT.DM.BRTHDAT"),40      AGE = as.integer(age),41      AGEU = if (is.na(age)) NA_character_ else "YEARS",42      SEX = item_value(subject, "IT.DM.SEX"),43      stringsAsFactors = FALSE44    )45  }46)47dm <- do.call(rbind, rows)4849write.csv(dm, "/app/output/dm.csv", row.names = FALSE, na = "")

1# Reference solution for the yamaa benchmark sdtm-dm-odm-snapshot (Python2# track).3#4# Written from the benchmark's full prompt and its inputs alone, as an agent5# would write it. Harbor's oracle agent runs it to check that the prompt can6# be solved and that the grader scores a correct answer 1.78"""Build the SDTM DM dataset from the EDC's ODM 1.3 XML snapshot.910One record per subject, with columns:11DOMAIN, STUDYID, USUBJID, SUBJID, BRTHDTC, AGE, AGEU, SEX12"""1314import csv15import xml.etree.ElementTree as ET1617INPUT_PATH = "/app/input/odm.xml"18OUTPUT_PATH = "/app/output/dm.csv"1920ODM = "{http://www.cdisc.org/ns/odm/v1.3}"21COLUMNS = ["DOMAIN", "STUDYID", "USUBJID", "SUBJID", "BRTHDTC", "AGE", "AGEU", "SEX"]222324def current_values(subject):25    """Each item's current value by its OID. An item recorded as null has no26    value; audit records, signatures and comments hold no items."""27    values = {}28    for item in subject.iter(f"{ODM}ItemData"):29        if item.get("IsNull", "").lower() in ("yes", "true", "1"):30            values[item.get("ItemOID")] = ""31        else:32            values[item.get("ItemOID")] = item.get("Value", "")33    return values343536def main():37    root = ET.parse(INPUT_PATH).getroot()38    rows = []39    for clinical in root.iter(f"{ODM}ClinicalData"):40        study = clinical.get("StudyOID")41        for subject in clinical.iter(f"{ODM}SubjectData"):42            key = subject.get("SubjectKey")43            values = current_values(subject)44            age = values.get("IT.DM.AGE", "")45            rows.append(46                {47                    "DOMAIN": "DM",48                    "STUDYID": study,49                    "USUBJID": f"{study}-{key}",50                    "SUBJID": key,51                    "BRTHDTC": values.get("IT.DM.BRTHDAT", ""),52                    "AGE": age,53                    "AGEU": "YEARS" if age else "",54                    "SEX": values.get("IT.DM.SEX", ""),55                }56            )57    with open(OUTPUT_PATH, "w", newline="", encoding="utf-8") as fh:58        writer = csv.DictWriter(fh, fieldnames=COLUMNS)59        writer.writeheader()60        writer.writerows(rows)616263if __name__ == "__main__":64    main()

1/* opensas reference solution for the yamaa benchmark sdtm-dm-odm-snapshot.2   Reads the benchmark inputs and derives the requested output.3   Execute with the existing opensas runtime: opensas /app/output/result.sas. */45/* opensas v0.7.1 reads no XML. The runner must stage odm.xml as odm.csv,6   one row per collected item in file order, without deriving any values:7   the item's study OID, subject key, item OID and value, and IsNull as8   true or false. Audit records, signatures and comments give no rows. */9data items;10  length StudyOID SubjectKey ItemOID Value IsNull $1024;11  infile '/app/input/odm.csv' dsd dlm=',' firstobs=2 truncover lrecl=32767;12  input StudyOID : $char1024. SubjectKey : $char1024. ItemOID : $char1024.13        Value : $char1024. IsNull : $char1024.;14run;1516data dm;17  length DOMAIN $2 STUDYID USUBJID SUBJID BRTHDTC AGEU SEX $1024;18  retain BRTHDTC AGE SEX;19  set items;20  by StudyOID SubjectKey notsorted;21  if first.SubjectKey then do;22    BRTHDTC = '';23    AGE = .;24    SEX = '';25  end;26  if IsNull ne 'true' then do;27    if ItemOID = 'IT.DM.BRTHDAT' then BRTHDTC = Value;28    if ItemOID = 'IT.DM.AGE' then AGE = input(Value, best32.);29    if ItemOID = 'IT.DM.SEX' then SEX = Value;30  end;31  if last.SubjectKey then do;32    DOMAIN = 'DM';33    STUDYID = StudyOID;34    USUBJID = catx('-', StudyOID, SubjectKey);35    SUBJID = SubjectKey;36    if missing(AGE) then AGEU = '';37    else AGEU = 'YEARS';38    output;39  end;40  keep DOMAIN STUDYID USUBJID SUBJID BRTHDTC AGE AGEU SEX;41run;4243proc sql;44  create table final as45    select DOMAIN, STUDYID, USUBJID, SUBJID, BRTHDTC, AGE, AGEU, SEX from dm;46quit;47proc export data=final outfile='/app/output/dm.csv' dbms=csv replace; run;