Review ask

As a clinical trial statistician or programmer, answer three short questions in Your review: the task, the inputs and outputs, and the AI-assisted review. Agreeing with all three approves this benchmark; disagreeing with any needs a comment. Propose a fix with the Edit button on any source file, or add it to your comments.

We suggest spending no more than 5 minutes on a review. The other tabs beside Your review are background for assessment, not under review.

Summary

Lifecycle: draftEditUnder review
What this benchmark means
README.md

Public entries: Python and R. Passing coverage requires an installed inventory.

Goal: write each administration's collected duration as EXDUR.

Input: collected exposure records, one per administration (EXSEQ), with the treatment (EXTRT), the start date (EXSTDTC) and, where the case report form collected it, the duration in whole minutes.

Variables:

  • EXDUR is the collected duration as ISO 8601 duration text: P, then the days, then T and the hours and minutes, each part that is zero left out, so 120 minutes is PT2H, 90 minutes PT1H30M and 45 minutes PT45M. A day is the largest part, so a 46-hour continuous infusion is P1DT22H, never PT46H. A bolus injection recorded as zero minutes is PT0S. Empty when the duration was not collected; it is never computed from the start date.

Inputs

Under review
1 source file

input/ex_raw.csv

Edit
6 rows
input/ex_raw.csv
STUDYIDUSUBJIDEXSEQEXTRTEXSTDTCEXDURMIN
YAMAA-011OXALIPLATIN2025-03-04120
YAMAA-012FLUOROURACIL2025-03-040
YAMAA-013FLUOROURACIL2025-03-042760
YAMAA-011OXALIPLATIN2025-03-0590
YAMAA-012FLUOROURACIL2025-03-05—
YAMAA-011OXALIPLATIN2025-03-0645

Expected output

Under review
6 expected rows

expected/ex.csv

Edit
6 rows
expected/ex.csv
DOMAINSTUDYIDUSUBJIDEXSEQEXTRTEXSTDTCEXDUR
EXYAMAA-011OXALIPLATIN2025-03-04PT2H
EXYAMAA-012FLUOROURACIL2025-03-04PT0S
EXYAMAA-013FLUOROURACIL2025-03-04P1DT22H
EXYAMAA-011OXALIPLATIN2025-03-05PT1H30M
EXYAMAA-012FLUOROURACIL2025-03-05—
EXYAMAA-011OXALIPLATIN2025-03-06PT45M
Shaded columns are derived— = missing value

Human review and assessment

Your review

3 questions and your comments · View benchmark discussions

Mapping spec

Generated from the benchmark YAML; do not edit by hand

Variable NameVariable LabelTypeLengthControlled Terms or FormatOriginCoreConversion DefinitionVariable TypeVariable OrderComments for Define
DOMAINDomain AbbreviationCharAssignedSet constant value "EX".SDTM1
STUDYIDStudy IdentifierCharCollectedCopy value from EX_RAW.STUDYID.SDTM2
USUBJIDUnique Subject IdentifierCharCollectedCopy value from EX_RAW.USUBJID.SDTM3
EXSEQSequence NumberNumCollectedCopy value from EX_RAW.EXSEQ.SDTM4
EXTRTName of TreatmentCharCollectedCopy value from EX_RAW.EXTRT.SDTM5
EXSTDTCStart Date/Time of TreatmentCharCollectedCopy value from EX_RAW.EXSTDTC.SDTM6
EXDURDuration of TreatmentCharDerivedWrite EX_RAW.EXDURMIN, a number of minutes, as ISO 8601 duration text with days as the largest unit.SDTM7
VersionDateAuthorDescriptionReviewerSign-off
1.72026-10-20yamaa docs buildRead a specification in the 1.0 surface as its 0.5 counterpart; state that missing values stay missing where a mapping declares no missing handler, and leave out a declared missing answer that restates a handler's default
1.62026-10-17yamaa docs buildDescribe collected date parts written as ISO 8601 date text and times of day counted in seconds since midnight
1.52026-10-17yamaa docs buildName collected precision in words instead of letter codes; state imputation fallbacks, minimum source precision, bounds ignored outside the collected period, and invalid text with no handler as an error; describe date and datetime shifts, imputation flag codes, and partial text giving no date part
1.42026-10-15yamaa docs buildState datetime precision codes through the hour and minute, and complete only the time parts a collected datetime lacks
1.32026-09-29yamaa docs buildShow study codelist names, external dictionaries, and ISO 8601 formats in variable rows; list the document's codelist values and metadata
1.22026-09-29yamaa docs buildClassify ADaM copies from a declared input as Predecessor and recodes as Derived, and SDTM recodes of locally derived columns as Derived
1.12026-09-29yamaa docs buildResolve lookups into the donor record they select, and give every derivation shape a plain-language rule instead of its raw YAML
1.0yamaa docs buildInitial generation from spec.yaml

Agent prompt

The request an agent receives with the input files; each tier states no more than the one above

prompts/sdtm-ex-infusion-duration/full.md

Edit
19 lines

Following CDISC SDTM standards, use the provided EX_RAW dataset to create an EX dataset with one record per administration, that is, per subject per EXSEQ.

The output dataset should contain the following columns in this order: DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXSTDTC, EXDUR

DOMAIN is "EX". Records are keyed by STUDYID, USUBJID and EXSEQ, where EXSEQ is the sequence number EX_RAW gives each administration.

EXDUR is the duration collected in EXDURMIN, a whole number of minutes, written as ISO 8601 duration text: "P", then the days, then "T" and the hours and minutes, each part that is zero left out. So 120 minutes is "PT2H", 90 minutes is "PT1H30M" and 45 minutes is "PT45M". A day of 24 hours is the largest part, so a 46-hour continuous infusion is "P1DT22H", never "PT46H". A bolus injection recorded as zero minutes is "PT0S". EXDUR has no value when EXDURMIN is empty; it is never computed from the start date.

Read the source datasets from /app/input and save the completed dataset as /app/output/ex.csv.

prompts/sdtm-ex-infusion-duration/conventions.md

Edit
14 lines

Following CDISC SDTM standards, use the provided EX_RAW dataset to create an EX dataset with one record per administration, that is, per subject per EXSEQ.

The output dataset should contain the following columns in this order: DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXSTDTC, EXDUR

DOMAIN is "EX". EXDUR is ISO 8601 duration text: "P", then the days, then "T" and the hours and minutes, each part that is zero left out. So 90 minutes is "PT1H30M", and a day of 24 hours is the largest part, so a 46-hour continuous infusion is "P1DT22H", never "PT46H". Zero minutes is "PT0S", or EXDUR has no value.

Read the source datasets from /app/input and save the completed dataset as /app/output/ex.csv.

prompts/sdtm-ex-infusion-duration/brief.md

Edit
8 lines

Following CDISC SDTM standards, use the provided EX_RAW dataset to create an EX dataset with one record per administration, that is, per subject per EXSEQ.

The output dataset should contain the following columns in this order: DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXSTDTC, EXDUR

Read the source datasets from /app/input and save the completed dataset as /app/output/ex.csv.

yamaa code

Public calls generated from benchmark.yaml; results must match the expected output

R

Edit
3 lines
library(yamaa)yamaa_spec("spec.yaml")$run()$output

Python

Edit
3 lines
import yamaayamaa.spec("spec.yaml").run().output

Solution without yamaa

Written from the full prompt and the inputs alone; CI checks it reproduces the expected output

solutions/sdtm-ex-infusion-duration/result.R

Edit
43 lines
# Reference solution for the yamaa benchmark sdtm-ex-infusion-duration (R track).## Written from the benchmark's full prompt and its inputs alone, as an agent# would write it. Harbor's oracle agent runs it to check that the prompt can# be solved and that the grader scores a correct answer 1.library(dplyr, warn.conflicts = FALSE)library(readr)ex_raw <- read_csv(  "/app/input/ex_raw.csv",  col_types = cols(.default = col_character()))# Whole minutes as ISO 8601 duration text, a day the largest part and every# part that is zero left out; zero minutes is PT0S.duration_text <- function(minutes) {  vapply(minutes, function(m) {    if (is.na(m)) return(NA_character_)    if (m == 0L) return("PT0S")    days <- m %/% 1440L    hours <- (m %% 1440L) %/% 60L    mins <- m %% 60L    text <- "P"    if (days > 0L) text <- paste0(text, days, "D")    if (hours > 0L || mins > 0L) {      text <- paste0(text, "T")      if (hours > 0L) text <- paste0(text, hours, "H")      if (mins > 0L) text <- paste0(text, mins, "M")    }    text  }, character(1), USE.NAMES = FALSE)}ex <- ex_raw |>  mutate(    DOMAIN = "EX",    EXDUR = duration_text(as.integer(EXDURMIN))  ) |>  select(DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXSTDTC, EXDUR)dir.create("/app/output", showWarnings = FALSE)write_csv(ex, "/app/output/ex.csv", na = "")

solutions/sdtm-ex-infusion-duration/result.py

Edit
67 lines
# Reference solution for the yamaa benchmark sdtm-ex-infusion-duration (Python track).## Written from the benchmark's full prompt and its inputs alone, as an agent# would write it. Harbor's oracle agent runs it to check that the prompt can# be solved and that the grader scores a correct answer 1."""Build the SDTM EX dataset from EX_RAW.One record per administration. EXDUR writes the collected duration in wholeminutes as ISO 8601 duration text, with a day as the largest part and everypart that is zero left out."""import csvINPUT_PATH = "/app/input/ex_raw.csv"OUTPUT_PATH = "/app/output/ex.csv"COLUMNS = ["DOMAIN", "STUDYID", "USUBJID", "EXSEQ", "EXTRT", "EXSTDTC", "EXDUR"]def duration_text(minutes):    """Whole minutes as ISO 8601 duration text: zero minutes is PT0S."""    if minutes == 0:        return "PT0S"    days, rest = divmod(minutes, 24 * 60)    hours, mins = divmod(rest, 60)    text = "P"    if days:        text += f"{days}D"    if hours or mins:        text += "T"        if hours:            text += f"{hours}H"        if mins:            text += f"{mins}M"    return textdef main():    with open(INPUT_PATH, newline="", encoding="utf-8") as f:        raw_rows = list(csv.DictReader(f))    out_rows = []    for row in raw_rows:        collected = (row["EXDURMIN"] or "").strip()        out_rows.append(            {                "DOMAIN": "EX",                "STUDYID": row["STUDYID"],                "USUBJID": row["USUBJID"],                "EXSEQ": row["EXSEQ"],                "EXTRT": row["EXTRT"],                "EXSTDTC": row["EXSTDTC"],                # No collected duration leaves EXDUR without a value.                "EXDUR": duration_text(int(collected)) if collected else "",            }        )    with open(OUTPUT_PATH, "w", newline="", encoding="utf-8") as f:        writer = csv.DictWriter(f, fieldnames=COLUMNS)        writer.writeheader()        writer.writerows(out_rows)if __name__ == "__main__":    main()

solutions/sdtm-ex-infusion-duration/result.sas

Edit
38 lines
/* opensas reference solution for the yamaa benchmark sdtm-ex-infusion-duration.   Reads the benchmark inputs and derives the requested output.   Execute with the existing opensas runtime: opensas /app/output/result.sas. */data ex_raw;  length STUDYID USUBJID EXTRT EXSTDTC $1024;  infile "/app/input/ex_raw.csv" dsd dlm="," firstobs=2 truncover lrecl=32767;  input STUDYID : $char1024. USUBJID : $char1024. EXSEQ EXTRT : $char1024. EXSTDTC : $char1024. EXDURMIN;run;/* EXDUR: the collected whole minutes as ISO 8601 duration text, a day the   largest part and every part that is zero left out; zero minutes is PT0S. */data result;  set ex_raw;  length DOMAIN $2 EXDUR $64;  DOMAIN = 'EX';  EXDUR = '';  if not missing(EXDURMIN) then do;    if EXDURMIN = 0 then EXDUR = 'PT0S';    else do;      days = floor(EXDURMIN / 1440);      hours = floor((EXDURMIN - days * 1440) / 60);      mins = EXDURMIN - days * 1440 - hours * 60;      EXDUR = 'P';      if days > 0 then EXDUR = cats(EXDUR, strip(put(days, 12.)), 'D');      if hours > 0 or mins > 0 then do;        EXDUR = cats(EXDUR, 'T');        if hours > 0 then EXDUR = cats(EXDUR, strip(put(hours, 12.)), 'H');        if mins > 0 then EXDUR = cats(EXDUR, strip(put(mins, 12.)), 'M');      end;    end;  end;run;proc sql;  create table final as select DOMAIN, STUDYID, USUBJID, EXSEQ, EXTRT, EXSTDTC, EXDUR from result;quit;proc export data=final outfile="/app/output/ex.csv" dbms=csv replace; run;