yamaa As a clinical trial statistician or programmer, challenge anything in this benchmark: describe a problem in Comments, or propose a fix with the Edit button on any source file.
- Task: Would a real study need this derivation, specified this way? Does it follow CDISC ADaM: variable names, controlled terms, record structure? Flag any rule that is ambiguous or that a sponsor would handle differently.
- Inputs: Do they look like data collected in a real trial, and do they include the situations you would expect to meet in one?
- Expected output: Would independent QC programming reproduce every value from these inputs and rules?
The other tabs beside Comments are background for assessment, not under review.
Summary
EditUnder reviewWhat this benchmark meansREADME.mdGoal: build the change parameters _ALB and _BILI from albumin
and bilirubin lab results, carrying each visit's previous result
(PREV_AVAL) and the change from it (CHG).
Input: laboratory (LB) records carrying a test code (LBTESTCD), a
numeric result (LBSTRESN), and a visit number (VISITNUM). Five
subjects: one with both tests across several visits, one with both
tests but a single albumin visit, one whose first albumin visit has
no result plus a single bilirubin visit, one albumin-only, and one
with hemoglobin only, which produces no rows.
Variables:
AVISITN: the analysis visit number, taken from the lab visit number.AVAL: the numeric lab result for the visit.PREV_AVAL: theAVALof the visit just before inAVISITNorder, for the same subject and parameter, so an_ALBrow never reads a_BILIvalue; blank on the subject's first visit for the parameter.CHG:AVAL - PREV_AVAL, the change since that visit; blank wheneverPREV_AVALis.
Note: a record with no numeric result gets no row, so the previous
value and the change skip that visit and compare with the latest
earlier visit that has a result. CHG here is the change from the
previous visit, not the ADaMIG "change from baseline" definition.
Single-visit subjects have blank PREV_AVAL and CHG; equal
consecutive values give CHG 0.
Inputs
Under review1 source fileinput/lb.csv
Edit19 rows| STUDYID | USUBJID | LBTESTCD | LBSTRESN | VISITNUM |
|---|---|---|---|---|
| YAMAA-01 | ALB | 4.0 | 1 | |
| YAMAA-01 | ALB | 4.5 | 2 | |
| YAMAA-01 | ALB | — | 3 | |
| YAMAA-01 | ALB | 4.75 | 4 | |
| YAMAA-01 | BILI | 1.0 | 1 | |
| YAMAA-01 | BILI | 1.25 | 2 | |
| YAMAA-01 | BILI | 1.5 | 3 | |
| YAMAA-01 | ALB | 3.75 | 1 | |
| YAMAA-01 | BILI | 2.0 | 1 | |
| YAMAA-01 | BILI | 2.25 | 2 | |
| YAMAA-01 | ALB | — | 1 | |
| YAMAA-01 | ALB | 3.5 | 2 | |
| YAMAA-01 | ALB | 3.75 | 3 | |
| YAMAA-01 | BILI | 0.75 | 1 | |
| YAMAA-01 | ALB | 4.0 | 1 | |
| YAMAA-01 | ALB | 4.0 | 2 | |
| YAMAA-01 | ALB | 3.5 | 3 | |
| YAMAA-01 | HGB | 14.2 | 1 | |
| YAMAA-01 | HGB | 14.5 | 2 |
Expected output
Under review15 expected rowsexpected/adlbc.csv
Edit15 rows| STUDYID | USUBJID | PARAMCD | AVISITN | AVAL | PREV_AVAL | CHG |
|---|---|---|---|---|---|---|
| YAMAA-01 | _ALB | 1 | 4 | — | — | |
| YAMAA-01 | _ALB | 2 | 4.5 | 4 | 0.5 | |
| YAMAA-01 | _ALB | 4 | 4.75 | 4.5 | 0.25 | |
| YAMAA-01 | _BILI | 1 | 1 | — | — | |
| YAMAA-01 | _BILI | 2 | 1.25 | 1 | 0.25 | |
| YAMAA-01 | _BILI | 3 | 1.5 | 1.25 | 0.25 | |
| YAMAA-01 | _ALB | 1 | 3.75 | — | — | |
| YAMAA-01 | _BILI | 1 | 2 | — | — | |
| YAMAA-01 | _BILI | 2 | 2.25 | 2 | 0.25 | |
| YAMAA-01 | _ALB | 2 | 3.5 | — | — | |
| YAMAA-01 | _ALB | 3 | 3.75 | 3.5 | 0.25 | |
| YAMAA-01 | _BILI | 1 | 0.75 | — | — | |
| YAMAA-01 | _ALB | 1 | 4 | — | — | |
| YAMAA-01 | _ALB | 2 | 4 | 4 | 0 | |
| YAMAA-01 | _ALB | 3 | 3.5 | 4 | -0.5 |
Comments and assessment materials
| Dataset Name | Variable Name | Variable Label | Type | Codelist/Controlled Terms | Core | Computational Method | Origin |
|---|---|---|---|---|---|---|---|
| ADLBC | STUDYID | Study Identifier | character | Copy value from LB.STUDYID. | Predecessor | ||
| ADLBC | USUBJID | Unique Subject Identifier | character | Copy value from LB.USUBJID. | Predecessor | ||
| ADLBC | PARAMCD | Parameter Code | character | Set per constructed row; see the Row construction sheet. | Derived | ||
| ADLBC | AVISITN | Analysis Visit Number | numeric | Copy value from LB.VISITNUM. | Predecessor | ||
| ADLBC | AVAL | Analysis Value | numeric | Copy value from LB.LBSTRESN. | Predecessor | ||
| ADLBC | PREV_AVAL | Previous Analysis Value | numeric | Take AVAL from 1 row earlier within each group of USUBJID, PARAMCD, ordered by AVISITN. | Derived | ||
| ADLBC | CHG | Change from Previous Value | numeric | Compute AVAL - PREV_AVAL. | Derived |
| Template | Include when | PARAMCD |
|---|---|---|
| change_params | LB.LBTESTCD IN ('ALB', 'BILI') AND LB.LBSTRESN IS NOT NULL | Recode LB.LBTESTCD: "ALB" → "_ALB", "BILI" → "_BILI"; missing values are errors; unlisted values stay missing |
| Version | Date | Author | Description | Reviewer | Sign-off |
|---|---|---|---|---|---|
| 1.3 | 2026-09-29 | yamaa docs build | Show study codelist names, external dictionaries, and ISO 8601 formats in variable rows; list the document's codelist values and metadata | ||
| 1.2 | 2026-09-29 | yamaa docs build | Classify ADaM copies from a declared input as Predecessor and recodes as Derived, and SDTM recodes of locally derived columns as Derived | ||
| 1.1 | 2026-09-29 | yamaa docs build | Resolve lookups into the donor record they select, and give every derivation shape a plain-language rule instead of its raw YAML | ||
| 1.0 | yamaa docs build | Initial generation from spec.yaml |
prompts/adam-adlbc-row-window/full.md
Edit27 linesFollowing CDISC ADaM standards, use the provided LB dataset to create an ADLBC dataset with one record per subject per parameter per analysis visit.
The output dataset should contain the following columns in this order: STUDYID, USUBJID, PARAMCD, AVISITN, AVAL, PREV_AVAL, CHG
Build the change parameters "_ALB" and "_BILI" from the albumin and bilirubin lab results; any other test produces no rows.
AVISITN is the analysis visit number, taken from the lab visit number. AVAL is the numeric lab result for the visit.
PREV_AVAL is the value of the visit just before in visit order, for the same subject and parameter, so an _ALB row never reads a _BILI value. It has no value on the subject's first visit for the parameter.
CHG is the change since that visit, not the change from baseline; it has no value whenever PREV_AVAL has none. A record with no numeric result gets no row, so the previous value and the change skip that visit and compare with the latest earlier visit that has a result. Equal consecutive values give CHG 0, and a single-visit subject leaves both empty.
Read the source datasets from /app/input and save the completed dataset as /app/output/adlbc.csv.
prompts/adam-adlbc-row-window/conventions.md
Edit11 linesFollowing CDISC ADaM standards, use the provided LB dataset to create an ADLBC dataset with one record per subject per parameter per analysis visit.
The output dataset should contain the following columns in this order: STUDYID, USUBJID, PARAMCD, AVISITN, AVAL, PREV_AVAL, CHG
PARAMCD is "_ALB" for albumin and "_BILI" for bilirubin.
Read the source datasets from /app/input and save the completed dataset as /app/output/adlbc.csv.
prompts/adam-adlbc-row-window/brief.md
Edit9 linesFollowing CDISC ADaM standards, use the provided LB dataset to create an ADLBC dataset with one record per subject per parameter per analysis visit.
The output dataset should contain the following columns in this order: STUDYID, USUBJID, PARAMCD, AVISITN, AVAL, PREV_AVAL, CHG
Read the source datasets from /app/input and save the completed dataset as /app/output/adlbc.csv.
run.R
Edit4 lineslibrary(yamaa)adlbc <- yamaa_domain("spec.yaml")$outputadlbcrun.py
Edit4 linesimport yamaaadlbc = yamaa.yamaa_domain("spec.yaml").outputadlbcsolutions/adam-adlbc-row-window/result.R
Edit41 lines# Reference solution for the yamaa benchmark adam-adlbc-row-window (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)lb <- read_csv( "/app/input/lb.csv", col_types = cols( LBSTRESN = col_double(), VISITNUM = col_integer(), .default = col_character() ))# Build the change parameters from albumin and bilirubin; any other test# produces no rows. A record with no numeric result gets no row.adbc <- lb |> filter(LBTESTCD %in% c("ALB", "BILI"), !is.na(LBSTRESN), !is.na(VISITNUM)) |> mutate( PARAMCD = case_when( LBTESTCD == "ALB" ~ "_ALB", LBTESTCD == "BILI" ~ "_BILI" ), AVISITN = VISITNUM, AVAL = LBSTRESN ) |> arrange(USUBJID, PARAMCD, AVISITN) |> group_by(USUBJID, PARAMCD) |> mutate( PREV_AVAL = lag(AVAL), CHG = AVAL - PREV_AVAL ) |> ungroup() |> select(STUDYID, USUBJID, PARAMCD, AVISITN, AVAL, PREV_AVAL, CHG)dir.create("/app/output", showWarnings = FALSE)write_csv(adbc, "/app/output/adlbc.csv", na = "")solutions/adam-adlbc-row-window/result.py
Edit45 lines# Reference solution for the yamaa benchmark adam-adlbc-row-window (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.from pathlib import Pathimport polars as pllb = pl.read_csv("/app/input/lb.csv", infer_schema=False).with_columns( pl.col("LBSTRESN").cast(pl.Float64, strict=False).alias("AVAL"), pl.col("VISITNUM").cast(pl.Int64, strict=False).alias("AVISITN"),)# Build the change parameters from albumin and bilirubin; any other test# produces no rows. A record with no numeric result gets no row.param_map = {"ALB": "_ALB", "BILI": "_BILI"}filtered = ( lb.filter( pl.col("LBTESTCD").is_in(list(param_map.keys())) & pl.col("AVAL").is_not_null() & pl.col("AVISITN").is_not_null() ) .with_columns( PARAMCD=pl.col("LBTESTCD").replace_strict(param_map), ) .select("STUDYID", "USUBJID", "PARAMCD", "AVISITN", "AVAL") .sort(["USUBJID", "PARAMCD", "AVISITN"]))# The previous visit's value for the same subject and parameter, and the# change since that visit.adbc = ( filtered.with_columns( PREV_AVAL=pl.col("AVAL").shift(1).over(["USUBJID", "PARAMCD"]), ) .with_columns( CHG=pl.col("AVAL") - pl.col("PREV_AVAL"), ) .select("STUDYID", "USUBJID", "PARAMCD", "AVISITN", "AVAL", "PREV_AVAL", "CHG"))Path("/app/output").mkdir(exist_ok=True)adbc.write_csv("/app/output/adlbc.csv")solutions/adam-adlbc-row-window/result.sas
Edit28 lines/* opensas reference solution for the yamaa benchmark adam-adlbc-row-window. Reads the benchmark inputs and derives the requested output. Execute with the existing opensas runtime: opensas /app/output/result.sas. */data lb; length STUDYID USUBJID LBTESTCD $1024; infile "/app/input/lb.csv" dsd dlm="," firstobs=2 truncover lrecl=32767; input STUDYID : $char1024. USUBJID : $char1024. LBTESTCD : $char1024. LBSTRESN VISITNUM;run;data collected; length PARAMCD $8; set lb; if LBTESTCD in ('ALB','BILI') and not missing(LBSTRESN); PARAMCD=cats('_',LBTESTCD); AVISITN=VISITNUM; AVAL=LBSTRESN;run;proc sort data=collected; by STUDYID USUBJID PARAMCD AVISITN; run;data result; retain previous priorchange; set collected; by STUDYID USUBJID PARAMCD; if first.PARAMCD then do; previous=.; priorchange=.; end; PREV_AVAL=previous; CHG=AVAL-PREV_AVAL; PREV2=priorchange; previous=AVAL; priorchange=CHG;run;proc sql; create table final as select STUDYID, USUBJID, PARAMCD, AVISITN, AVAL, PREV_AVAL, CHG from result;quit;proc export data=final outfile="/app/output/adlbc.csv" dbms=csv replace; run;
Comments
Signed in with GitHub; saved in repository Discussions