Collected Medications with Dose Text and Timing References

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

EditUnder reviewWhat this benchmark means

README.md

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

Goal: build the CM record for each collected medication from the concomitant-medication form, carrying the reported name, indication, dose, frequency, route, and dates as collected, and the form's "taken before study" and "ongoing" boxes as timing references instead of invented dates.

Input: the medication log for two subjects: each medication is one repeat of the IG.CM item group, and the repeat number becomes CMSEQ. The dose is collected as free text, and the frequency and route as the labels printed on the form (Once daily, Twice daily, By mouth). One subject reports a single medication with complete dates and a single dose; the other reports a medication started before the study with only year and month known and the "ongoing" box checked, plus a medication whose dose was collected as a range.

Variables:

CMTRT is the medication name as reported, kept exactly as written.
CMINDC is the indication as reported.
CMDOSE is the collected dose when it is a plain number, such as 100; empty otherwise.
CMDOSTXT is the collected dose when it is not a plain number, such as the range 200-400; empty when CMDOSE carries the dose.
CMDOSU is the unit of the collected dose.
CMDOSFRQ is the collected frequency label mapped to controlled terminology: Once daily becomes QD and Twice daily becomes BID.
CMROUTE is the collected route label mapped to controlled terminology: By mouth becomes ORAL.
CMSTDTC is the start of the medication as collected, keeping partial dates at their collected precision: year and month alone stay 2023-09.
CMENDTC is the end of the medication as collected; empty when no end date was collected.
CMSTRTPT is BEFORE when the "taken before study" box is checked, and CMSTTPT names the reference point, SCREENING; both empty otherwise.
CMENRTPT is ONGOING when the "ongoing" box is checked, and CMENTPT names the reference point, END OF STUDY; both empty otherwise.

Note: a frequency or route label the mappings do not list stops the run rather than passing through unmapped. An ongoing medication has no end date, and CMENRTPT records why, so an empty CMENDTC is never left to stand for "ongoing" on its own.

YAML specification

1 spec file

spec.yaml

Edit229 lines

1yamaa: "1.0"2type: spec3model: SDTM4domain: CM5keys: [STUDYID, USUBJID, CMSEQ]6input:7  ODM:8    path: input/odm.csv9    types: {StudyEventRepeatKey: int, ItemGroupRepeatKey: int}1011output:12  path: cm.csv13  columns: [DOMAIN, STUDYID, USUBJID, CMSEQ, CMTRT, CMINDC, CMDOSE, CMDOSTXT,14            CMDOSU, CMDOSFRQ, CMROUTE, CMSTDTC, CMENDTC, CMSTRTPT, CMSTTPT,15            CMENRTPT, CMENTPT]16  order_by: [USUBJID, CMSEQ]1718columns:19  - name: DOMAIN20    type: str21    label: Domain Abbreviation22    derivation: "'CM'"2324  - name: STUDYID25    type: str26    label: Study Identifier27    derivation: ODM.StudyOID2829  - name: USUBJID30    type: str31    label: Unique Subject Identifier32    derivation: ODM.SubjectKey3334  - name: CMSEQ35    type: int36    label: Sequence Number37    derivation: ODM.ItemGroupRepeatKey3839  - name: CMTRT40    type: str41    label: Reported Name of Drug, Med, or Therapy42    verifications:43      - not_missing4445  - name: CMINDC46    type: str47    label: Indication4849  - name: DSTXT50    type: str51    label: Collected dose text (intermediate, not in output)5253  # A collected dose that is a plain number is the numeric dose.54  - name: CMDOSE55    type: float56    label: Dose per Administration57    derivation:58      str_extract:59        source: DSTXT60        pattern: '^[0-9]+(\.[0-9]+)?$'61        no_match: null6263  # Any other collected dose, such as the range "200-400", stays as text.64  - name: CMDOSTXT65    type: str66    label: Dose Description67    derivation:68      case:69        - when: "DSTXT IS NOT NULL AND CMDOSE IS NULL"70          then: DSTXT7172  - name: CMDOSU73    type: str74    label: Dose Units75    verifications:76      - allowed_values:77          values: [mg]7879  - name: FRQLBL80    type: str81    label: Collected frequency label (intermediate, not in output)8283  - name: CMDOSFRQ84    type: str85    label: Dosing Frequency per Interval86    derivation:87      mapping:88        source: FRQLBL89        dict:90          Once daily: QD91          Twice daily: BID9293  - name: ROUTELBL94    type: str95    label: Collected route label (intermediate, not in output)9697  - name: CMROUTE98    type: str99    label: Route of Administration100    derivation:101      mapping:102        source: ROUTELBL103        dict:104          By mouth: ORAL105106  - name: CMSTDTC107    type: str108    label: Start Date/Time of Medication109    # As collected, including partial ISO 8601 dates such as "2023-09".110111  - name: CMENDTC112    type: str113    label: End Date/Time of Medication114115  - name: PRIOR116    type: str117    label: Collected taken-before-study box (intermediate, not in output)118119  - name: CMSTRTPT120    type: str121    label: Start Relative to Reference Time Point122    derivation:123      case:124        - when: "PRIOR = 'Y'"125          then: "'BEFORE'"126127  - name: CMSTTPT128    type: str129    label: Start Reference Time Point130    derivation:131      case:132        - when: "CMSTRTPT IS NOT NULL"133          then: "'SCREENING'"134135  - name: ONGO136    type: str137    label: Collected ongoing box (intermediate, not in output)138139  - name: CMENRTPT140    type: str141    label: End Relative to Reference Time Point142    derivation:143      case:144        - when: "ONGO = 'Y'"145          then: "'ONGOING'"146147  - name: CMENTPT148    type: str149    label: End Reference Time Point150    derivation:151      case:152        - when: "CMENRTPT IS NOT NULL"153          then: "'END OF STUDY'"154155rows:156  - id: medication157    dataset: ODM158    group_by: [ODM.StudyOID, ODM.SubjectKey, ODM.ItemGroupOID,159               ODM.ItemGroupRepeatKey]160    derivations:161      CMTRT:162        aggregate:163          filter: "ODM.ItemOID = 'IT.CM.CMTRT'"164          expr: "ONLY(ODM.Value)"165      CMINDC:166        aggregate:167          filter: "ODM.ItemOID = 'IT.CM.CMINDC'"168          expr: "ONLY(ODM.Value)"169      DSTXT:170        aggregate:171          filter: "ODM.ItemOID = 'IT.CM.CMDSTXT'"172          expr: "ONLY(ODM.Value)"173      CMDOSU:174        aggregate:175          filter: "ODM.ItemOID = 'IT.CM.CMDOSU'"176          expr: "ONLY(ODM.Value)"177      FRQLBL:178        aggregate:179          filter: "ODM.ItemOID = 'IT.CM.CMDOSFRQ'"180          expr: "ONLY(ODM.Value)"181      ROUTELBL:182        aggregate:183          filter: "ODM.ItemOID = 'IT.CM.CMROUTE'"184          expr: "ONLY(ODM.Value)"185      CMSTDTC:186        aggregate:187          filter: "ODM.ItemOID = 'IT.CM.CMSTDTC'"188          expr: "ONLY(ODM.Value)"189      CMENDTC:190        aggregate:191          filter: "ODM.ItemOID = 'IT.CM.CMENDTC'"192          expr: "ONLY(ODM.Value)"193      PRIOR:194        aggregate:195          filter: "ODM.ItemOID = 'IT.CM.CMPRIOR'"196          expr: "ONLY(ODM.Value)"197      ONGO:198        aggregate:199          filter: "ODM.ItemOID = 'IT.CM.CMONGO'"200          expr: "ONLY(ODM.Value)"201202verifications:203  - row_count:204      min: 3205      max: 3206  - assert:207      id: numeric-dose-has-no-dose-text208      when: "CMDOSE IS NOT NULL"209      require: "CMDOSTXT IS NULL"210  - assert:211      id: dose-range-kept-as-text212      when: "CMTRT = 'IBUPROFEN'"213      require: "CMDOSTXT = '200-400' AND CMDOSE IS NULL"214  - assert:215      id: partial-start-date-kept-at-collected-precision216      when: "CMTRT = 'LISINOPRIL'"217      require: "CMSTDTC = '2023-09'"218  - assert:219      id: taken-before-study-names-screening220      when: "CMSTRTPT = 'BEFORE'"221      require: "CMSTTPT = 'SCREENING'"222  - assert:223      id: ongoing-has-no-end-date224      when: "CMENRTPT = 'ONGOING'"225      require: "CMENDTC IS NULL AND CMENTPT = 'END OF STUDY'"226  - assert:227      id: blank-end-date-is-ongoing228      when: "CMENDTC IS NULL"229      require: "CMENRTPT = 'ONGOING'"

Inputs

Under review1 source file

input/odm.csv

Edit29 rows

input/odm.csv
StudyOID	MetaDataVersionOID	SubjectKey	StudyEventOID	StudyEventRepeatKey	ItemGroupOID	ItemGroupRepeatKey	ItemOID	Value
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMTRT	ASPIRIN
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMINDC	Headache
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMDSTXT	100
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMDOSU	mg
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMDOSFRQ	Once daily
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMROUTE	By mouth
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMSTDTC	2024-03-05
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMENDTC	2024-03-12
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMPRIOR	N
YAMAA-01	MDV.1	YAMAA-01-001	TREATMENT	1	IG.CM	1	IT.CM.CMONGO	N
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMTRT	LISINOPRIL
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMINDC	Hypertension
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMDSTXT	10
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMDOSU	mg
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMDOSFRQ	Once daily
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMROUTE	By mouth
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMSTDTC	2023-09
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMPRIOR	Y
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	1	IT.CM.CMONGO	Y
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMTRT	IBUPROFEN
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMINDC	Back pain
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMDSTXT	200-400
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMDOSU	mg
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMDOSFRQ	Twice daily
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMROUTE	By mouth
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMSTDTC	2024-04-02
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMENDTC	2024-04-09
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMPRIOR	N
YAMAA-01	MDV.1	YAMAA-01-002	TREATMENT	1	IG.CM	2	IT.CM.CMONGO	N

Expected output

Under review3 expected rows

expected/cm.csv

Edit3 rows

expected/cm.csv
DOMAIN	STUDYID	CMSEQ	CMTRT	CMINDC	CMDOSE	CMDOSTXT	CMDOSU	CMDOSFRQ	CMROUTE	CMSTDTC	CMENDTC	CMSTRTPT	CMSTTPT	CMENRTPT	CMENTPT
CM	YAMAA-01	1	ASPIRIN	Headache	100	—	mg	QD	ORAL	2024-03-05	2024-03-12	—	—	—	—
CM	YAMAA-01	1	LISINOPRIL	Hypertension	10	—	mg	QD	ORAL	2023-09	—	BEFORE	SCREENING	ONGOING	END OF STUDY
CM	YAMAA-01	2	IBUPROFEN	Back pain	—	200-400	mg	BID	ORAL	2024-04-02	2024-04-09	—	—	—	—

Shaded columns are derived— = missing value

Human review and assessment

Variable Name	Variable Label	Type	Controlled Terms or Format	Origin	Conversion Definition	Variable Type	Variable Order
DOMAIN	Domain Abbreviation	Char		Assigned	Set constant value "CM".	SDTM	1
STUDYID	Study Identifier	Char		Collected	Copy value from ODM.StudyOID.	SDTM	2
USUBJID	Unique Subject Identifier	Char		Collected	Copy value from ODM.SubjectKey.	SDTM	3
CMSEQ	Sequence Number	Num		Collected	Copy value from ODM.ItemGroupRepeatKey.	SDTM	4
CMTRT	Reported Name of Drug, Med, or Therapy	Char		Derived	Set per constructed row; see the Row construction sheet.	SDTM	5
CMINDC	Indication	Char		Derived	Set per constructed row; see the Row construction sheet.	SDTM	6
CMDOSE	Dose per Administration	Num		Derived	Extract the whole match matching ^[0-9]+(\.[0-9]+)?$ from DSTXT; no match → blank.	SDTM	7
CMDOSTXT	Dose Description	Char		Derived	If DSTXT IS NOT NULL AND CMDOSE IS NULL then "DSTXT"; otherwise blank.	SDTM	8
CMDOSU	Dose Units	Char	mg	Derived	Set per constructed row; see the Row construction sheet.	SDTM	9
CMDOSFRQ	Dosing Frequency per Interval	Char		Derived	Recode FRQLBL: "Once daily" → "QD", "Twice daily" → "BID"; missing values stay missing; unlisted values are errors.	SDTM	10
CMROUTE	Route of Administration	Char		Derived	Recode ROUTELBL: "By mouth" → "ORAL"; missing values stay missing; unlisted values are errors.	SDTM	11
CMSTDTC	Start Date/Time of Medication	Char		Derived	Set per constructed row; see the Row construction sheet.	SDTM	12
CMENDTC	End Date/Time of Medication	Char		Derived	Set per constructed row; see the Row construction sheet.	SDTM	13
CMSTRTPT	Start Relative to Reference Time Point	Char		Derived	If PRIOR = 'Y' then "BEFORE"; otherwise blank.	SDTM	14
CMSTTPT	Start Reference Time Point	Char		Derived	If CMSTRTPT IS NOT NULL then "SCREENING"; otherwise blank.	SDTM	15
CMENRTPT	End Relative to Reference Time Point	Char		Derived	If ONGO = 'Y' then "ONGOING"; otherwise blank.	SDTM	16
CMENTPT	End Reference Time Point	Char		Derived	If CMENRTPT IS NOT NULL then "END OF STUDY"; otherwise blank.	SDTM	17

Template	Include when	CMTRT	CMINDC	DSTXT	CMDOSU	FRQLBL	ROUTELBL	CMSTDTC	CMENDTC	PRIOR	ONGO
medication		Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMTRT'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMINDC'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMDSTXT'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMDOSU'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMDOSFRQ'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMROUTE'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMSTDTC'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMENDTC'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMPRIOR'	Aggregate ONLY(ODM.Value) over records matching on the shared output keys, where ODM.ItemOID = 'IT.CM.CMONGO'

Variable	Permitted values
CMDOSU	mg

Version	Date	Author	Description
1.7	2026-10-20	yamaa docs build	Read 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
1.6	2026-10-17	yamaa docs build	Describe collected date parts written as ISO 8601 date text and times of day counted in seconds since midnight
1.5	2026-10-17	yamaa docs build	Name 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
1.4	2026-10-15	yamaa docs build	State datetime precision codes through the hour and minute, and complete only the time parts a collected datetime lacks
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/sdtm-cm-collected-medications/full.md

Edit30 lines

Following CDISC SDTM standards, use the provided ODM dataset to create a CM dataset with one record per medication.

The output dataset should contain the following columns in this order: DOMAIN, STUDYID, USUBJID, CMSEQ, CMTRT, CMINDC, CMDOSE, CMDOSTXT, CMDOSU, CMDOSFRQ, CMROUTE, CMSTDTC, CMENDTC, CMSTRTPT, CMSTTPT, CMENRTPT, CMENTPT

STUDYID is the study OID and USUBJID is the subject key, each as given.
CMSEQ is the repeat number of the medication for the subject.
CMDOSE is the collected dose as a number when the dose text is a plain number, such as "100"; CMDOSE has no value otherwise.
CMDOSTXT is the collected dose text when it is not a plain number, such as the range "200-400"; CMDOSTXT has no value when CMDOSE carries the dose.
CMDOSFRQ is the controlled term for the collected frequency label: "Once daily" becomes "QD" and "Twice daily" becomes "BID".
CMROUTE is the controlled term for the collected route label: "By mouth" becomes "ORAL".
CMSTDTC is the start of the medication as collected, keeping partial dates at their collected precision: a start known only to year and month stays "2023-09".
CMENDTC is the end of the medication as collected; CMENDTC has no value when no end date was collected.
CMSTRTPT is "BEFORE" when the "taken before study" box is checked, and CMSTTPT is "SCREENING" then; both have no value otherwise.
CMENRTPT is "ONGOING" when the "ongoing" box is checked, and CMENTPT is "END OF STUDY" then; both have no value otherwise.

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

prompts/sdtm-cm-collected-medications/conventions.md

Edit16 lines

Following CDISC SDTM standards, use the provided ODM dataset to create a CM dataset with one record per medication.

CMSTDTC keeps a partial date at its collected precision: a start known only to year and month stays "2023-09".
CMSTRTPT is "BEFORE" or has no value, and CMSTTPT is "SCREENING" or has no value.
CMENRTPT is "ONGOING" or has no value, and CMENTPT is "END OF STUDY" or has no value.

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

prompts/sdtm-cm-collected-medications/brief.md

Edit9 lines

Following CDISC SDTM standards, use the provided ODM dataset to create a CM dataset with one record per medication.

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

R

Edit3 lines

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

Python

Edit3 lines

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

solutions/sdtm-cm-collected-medications/result.R

Edit76 lines

1# Reference solution for the yamaa benchmark sdtm-cm-collected-medications (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)10library(stringr)1112odm <- read_csv(13  "/app/input/odm.csv",14  col_types = cols(.default = col_character())15)1617# A plain number carries the dose; anything else (like a range 50-75) stays18# as text. Frequency and route labels map to controlled terminology.19cm <- odm |>20  filter(ItemGroupOID == "IG.CM") |>21  pivot_wider(22    id_cols = c(StudyOID, SubjectKey, ItemGroupRepeatKey),23    names_from = ItemOID,24    values_from = Value,25    values_fn = first26  ) |>27  mutate(28    CMSEQ = as.integer(ItemGroupRepeatKey),29    DOSE_RAW = str_trim(`IT.CM.CMDSTXT`),30    DOSE_PLAIN = str_detect(DOSE_RAW, "^\\d+(\\.\\d+)?$"),31    # BEFORE/SCREENING when the taken-before-study box is checked; ONGOING/32    # END OF STUDY when the ongoing box is checked. An ongoing medication33    # has no end date.34    DOMAIN = "CM",35    STUDYID = StudyOID,36    USUBJID = SubjectKey,37    CMTRT = na_if(`IT.CM.CMTRT`, ""),38    CMINDC = na_if(`IT.CM.CMINDC`, ""),39    CMDOSE = if_else(40      DOSE_PLAIN, suppressWarnings(as.numeric(DOSE_RAW)), NA_real_,41      missing = NA_real_42    ),43    CMDOSTXT = if_else(44      DOSE_PLAIN, NA_character_, na_if(DOSE_RAW, ""), missing = na_if(DOSE_RAW, "")45    ),46    CMDOSU = na_if(`IT.CM.CMDOSU`, ""),47    CMDOSFRQ = case_when(48      `IT.CM.CMDOSFRQ` == "Once daily" ~ "QD",49      `IT.CM.CMDOSFRQ` == "Twice daily" ~ "BID",50      .default = na_if(`IT.CM.CMDOSFRQ`, "")51    ),52    CMROUTE = case_when(53      `IT.CM.CMROUTE` == "By mouth" ~ "ORAL",54      .default = na_if(`IT.CM.CMROUTE`, "")55    ),56    CMSTDTC = na_if(`IT.CM.CMSTDTC`, ""),57    CMENDTC = if_else(58      `IT.CM.CMONGO` == "Y", NA_character_, na_if(`IT.CM.CMENDTC`, ""),59      missing = na_if(`IT.CM.CMENDTC`, "")60    ),61    CMSTRTPT = if_else(`IT.CM.CMPRIOR` == "Y", "BEFORE", NA_character_, missing = NA_character_),62    CMSTTPT = if_else(`IT.CM.CMPRIOR` == "Y", "SCREENING", NA_character_, missing = NA_character_),63    CMENRTPT = if_else(`IT.CM.CMONGO` == "Y", "ONGOING", NA_character_, missing = NA_character_),64    CMENTPT = if_else(65      `IT.CM.CMONGO` == "Y", "END OF STUDY", NA_character_, missing = NA_character_66    )67  ) |>68  arrange(USUBJID, CMSEQ) |>69  select(70    DOMAIN, STUDYID, USUBJID, CMSEQ, CMTRT, CMINDC, CMDOSE, CMDOSTXT,71    CMDOSU, CMDOSFRQ, CMROUTE, CMSTDTC, CMENDTC, CMSTRTPT, CMSTTPT,72    CMENRTPT, CMENTPT73  )7475dir.create("/app/output", showWarnings = FALSE)76write_csv(cm, "/app/output/cm.csv", na = "")

solutions/sdtm-cm-collected-medications/result.py

Edit90 lines

1# Reference solution for the yamaa benchmark sdtm-cm-collected-medications (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 1.67"""Build the SDTM CM dataset from the long-format ODM extract.89One record per medication: a medication is one ItemGroupRepeatKey group10within a subject. Reads /app/input/odm.csv and writes /app/output/cm.csv.11"""1213import csv14import re15from collections import OrderedDict1617INPUT = "/app/input/odm.csv"18OUTPUT = "/app/output/cm.csv"1920COLUMNS = [21    "DOMAIN", "STUDYID", "USUBJID", "CMSEQ", "CMTRT", "CMINDC", "CMDOSE",22    "CMDOSTXT", "CMDOSU", "CMDOSFRQ", "CMROUTE", "CMSTDTC", "CMENDTC",23    "CMSTRTPT", "CMSTTPT", "CMENRTPT", "CMENTPT",24]2526FREQ_MAP = {"Once daily": "QD", "Twice daily": "BID"}27ROUTE_MAP = {"By mouth": "ORAL"}2829PLAIN_NUMBER = re.compile(r"^[+-]?(\d+(\.\d*)?|\.\d+)$")303132def dose_fields(text):33    """Split collected dose text into (CMDOSE, CMDOSTXT)."""34    if text is None or text == "":35        return "", ""36    if PLAIN_NUMBER.match(text.strip()):37        value = float(text)38        number = str(int(value)) if value == int(value) else repr(value)39        return number, ""40    return "", text414243def main():44    with open(INPUT, newline="", encoding="utf-8") as fh:45        rows = list(csv.DictReader(fh))4647    # Group item rows into medication records, preserving collected order.48    records = OrderedDict()49    for row in rows:50        key = (row["StudyOID"], row["SubjectKey"], row["ItemGroupRepeatKey"])51        record = records.setdefault(key, {})52        record[row["ItemOID"]] = row["Value"]5354    out_rows = []55    for (study_oid, subject_key, repeat), items in records.items():56        cmdose, cmdostxt = dose_fields(items.get("IT.CM.CMDSTXT"))5758        prior = items.get("IT.CM.CMPRIOR") == "Y"59        ongoing = items.get("IT.CM.CMONGO") == "Y"6061        out_rows.append({62            "DOMAIN": "CM",63            "STUDYID": study_oid,64            "USUBJID": subject_key,65            "CMSEQ": repeat,66            "CMTRT": items.get("IT.CM.CMTRT", ""),67            "CMINDC": items.get("IT.CM.CMINDC", ""),68            "CMDOSE": cmdose,69            "CMDOSTXT": cmdostxt,70            "CMDOSU": items.get("IT.CM.CMDOSU", ""),71            "CMDOSFRQ": FREQ_MAP.get(items.get("IT.CM.CMDOSFRQ", ""),72                                     items.get("IT.CM.CMDOSFRQ", "")),73            "CMROUTE": ROUTE_MAP.get(items.get("IT.CM.CMROUTE", ""),74                                     items.get("IT.CM.CMROUTE", "")),75            "CMSTDTC": items.get("IT.CM.CMSTDTC", ""),76            "CMENDTC": items.get("IT.CM.CMENDTC", ""),77            "CMSTRTPT": "BEFORE" if prior else "",78            "CMSTTPT": "SCREENING" if prior else "",79            "CMENRTPT": "ONGOING" if ongoing else "",80            "CMENTPT": "END OF STUDY" if ongoing else "",81        })8283    with open(OUTPUT, "w", newline="", encoding="utf-8") as fh:84        writer = csv.DictWriter(fh, fieldnames=COLUMNS)85        writer.writeheader()86        writer.writerows(out_rows)878889if __name__ == "__main__":90    main()

solutions/sdtm-cm-collected-medications/result.sas

Edit42 lines

1/* opensas reference solution for the yamaa benchmark sdtm-cm-collected-medications.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 ItemGroupRepeatKey; run;12data wide;13  length CMTRT CMINDC CMDOSU CMDOSFRQ CMROUTE CMSTDTC CMENDTC CMPRIOR CMONGO DOSETEXT $1024;14  retain CMTRT CMINDC CMDOSU CMDOSFRQ CMROUTE CMSTDTC CMENDTC CMPRIOR CMONGO DOSETEXT;15  set ordered;16  where ItemGroupOID='IG.CM';17  by StudyOID SubjectKey ItemGroupRepeatKey;18  if first.ItemGroupRepeatKey then do;19    CMTRT=''; CMINDC=''; CMDOSU=''; CMDOSFRQ=''; CMROUTE=''; CMSTDTC=''; CMENDTC=''; CMPRIOR=''; CMONGO=''; DOSETEXT='';20  end;21    if ItemOID='IT.CM.CMTRT' then CMTRT=Value;22  if ItemOID='IT.CM.CMINDC' then CMINDC=Value;23  if ItemOID='IT.CM.CMDOSU' then CMDOSU=Value;24  if ItemOID='IT.CM.CMDOSFRQ' then CMDOSFRQ=Value;25  if ItemOID='IT.CM.CMROUTE' then CMROUTE=Value;26  if ItemOID='IT.CM.CMSTDTC' then CMSTDTC=Value;27  if ItemOID='IT.CM.CMENDTC' then CMENDTC=Value;28  if ItemOID='IT.CM.CMPRIOR' then CMPRIOR=Value;29  if ItemOID='IT.CM.CMONGO' then CMONGO=Value;30  if ItemOID='IT.CM.CMDSTXT' then DOSETEXT=Value;31  if last.ItemGroupRepeatKey then output;32run;33data result; set wide; length DOMAIN $2 STUDYID USUBJID CMDOSTXT CMSTRTPT CMSTTPT CMENRTPT CMENTPT $1024;34DOMAIN='CM'; STUDYID=StudyOID; USUBJID=SubjectKey; CMSEQ=input(ItemGroupRepeatKey,best32.);35if prxmatch('/^[+-]?[0-9]+(\.[0-9]+)?$/',strip(DOSETEXT)) then CMDOSE=input(DOSETEXT,best32.); else CMDOSTXT=DOSETEXT;36if CMDOSFRQ='Once daily' then CMDOSFRQ='QD'; if CMDOSFRQ='Twice daily' then CMDOSFRQ='BID'; if CMROUTE='By mouth' then CMROUTE='ORAL';37if CMPRIOR='Y' then do; CMSTRTPT='BEFORE'; CMSTTPT='SCREENING'; end;38if CMONGO='Y' then do; CMENDTC=''; CMENRTPT='ONGOING'; CMENTPT='END OF STUDY'; end; run;39proc sql;40  create table final as select DOMAIN, STUDYID, USUBJID, CMSEQ, CMTRT, CMINDC, CMDOSE, CMDOSTXT, CMDOSU, CMDOSFRQ, CMROUTE, CMSTDTC, CMENDTC, CMSTRTPT, CMSTTPT, CMENRTPT, CMENTPT from result;41quit;42proc export data=final outfile="/app/output/cm.csv" dbms=csv replace; run;

input/odm.csv

expected/cm.csv

Your review

Mapping spec

Agent prompt

prompts/sdtm-cm-collected-medications/full.md

prompts/sdtm-cm-collected-medications/conventions.md

prompts/sdtm-cm-collected-medications/brief.md

yamaa code

R

Python

Solution without yamaa

solutions/sdtm-cm-collected-medications/result.R

solutions/sdtm-cm-collected-medications/result.py

solutions/sdtm-cm-collected-medications/result.sas