Issue-2843/compatible wizard [help needed] by seolhwa-kim-data4life · Pull Request #2889 · OHDSI/Data2Evidence

seolhwa-kim-data4life · 2026-07-13T14:31:08Z

Description #2843

When a condition is selected in the Wizards dashboards, the concept name (e.g. "Acute bronchitis") was being sent to the endpoint as CONCEPT_CODE, but the SQL joins on concept_code, so it never matched.
Per the ticket, non-HANA (duckdb/postgres) datasets should send the concept_id and HANA datasets should send the concept_code.

Here is following changes in wizard-dashboard (https://github.com/data2evidence/wizards-dashboard/pull/24)

Verified locally

Incidence rate, prevalence rate, mortality rate, and demographics dashboards are visible for synpuf5pct dataset in the D2E develop portal

Help needed

I wish a careful review / helping me test on these — this isn't my main area (nor ts / vue / wizards test).
How can i also check Incidence rate, prevalence rate, mortality rate, and demographics dashboards are visible for HANA_CDMSYNPUF1K dataset?

…oards Wizards dashboards sent the condition concept NAME to parquet-export as CONCEPT_CODE, so the SQL join on concept_code never matched (and names with spaces were rejected by the endpoint). Non-HANA (duckdb/postgres) datasets need concept_id and HANA datasets need concept_code. Frontend + config layer (Tasks 1-3): - 03_Config seed: add condition attributes that return the identifier as `value` and the concept name as `text` (useRefValue + useRefText) - condition_occ_concept_id (@REF.CONCEPT_ID) for duckdb, conditionconceptcode (@REF.CONCEPT_CODE) for HANA and HANA-lean, in both CDW and PA configs. - Add knex migration to patch these attributes into already-seeded configs (the seed is insert-ignore on existing environments). - Point the wizards condition fields at the new attributes and set allowFreeText:false (identifiers are not free-typed); sync dev mocks/docs. - Plumb the dataset dialect from the portal /dataset endpoint through useDashboardFlow -> ChartToolbar -> ShinyDashboardModal -> ShinyDashboardIframe into the postMessage AUTH_TOKEN context so the dashboard can pick the right key. - Update shinyDashboardContext tests for the new dialect field. Backend (parquet-export CONCEPT_ID support), SQL templates and Python dashboard changes are follow-ups. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Non-HANA (duckdb/postgres) SQL templates now match conditions on concept_id, so the endpoint must accept CONCEPT_ID1..5 in the request body alongside the existing CONCEPT_CODE1..5 (HANA). - Add CONCEPT_ID1..5 to RESERVED_PLACEHOLDERS and substitute them unquoted, validated as digits only (isValidConceptId), defaulting empty slots to 0 so the SQL stays valid (templates filter concept_id = 0). - Add a dialect guard: reject with a clear 400 when a HANA dataset gets a concept_id template or a non-HANA dataset gets a concept_code template, instead of silently returning an empty result. Dialect is read from the portal /dataset response (added `dialect` to DatasetMetadata). - Export substituteTemplateParams/isValidConceptId and guard app.listen behind import.meta.main so the module is unit-testable; add index.test.ts. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

The wizard e2e spec embeds the canonical cross-sectional-demographics dashboard (Python + SQL) that it pastes into Manage Dashboard for the duckdb Demo dataset. Update it to match the new dialect-aware contract: - Forward context.dialect from the AUTH_TOKEN postMessage into a d2e_dialect Shiny input and read it in get_context(). - Build the parquet-export payload with CONCEPT_ID{n} (int) for non-HANA and CONCEPT_CODE{n} (string) for HANA, keyed off the dialect. - Rewrite the entry CTE to join {{VOCAB_SCHEMA}}.concept on concept_id ({{CONCEPT_IDn}}) and drop empty (0) slots, sourcing concept_code from the joined concept row. - Show the condition displayName (not the raw id) in the dashboard description. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Copilot

Pull request overview

This PR updates the Patient Analytics “wizards” flow to correctly pass a condition identifier that matches the target database dialect (concept_id for non-HANA; concept_code for HANA), and propagates the dataset dialect into the Shiny dashboard context so parquet export SQL templates can be substituted safely.

Changes:

Update wizard condition fields/config to use concept identifiers appropriate for non-HANA vs HANA (and disable free-text for these fields).
Add dataset dialect propagation from UI → Shiny context and use it to shape condition payloads.
Extend parquet-export to support {{CONCEPT_IDn}} placeholders (validation + substitution) and add unit tests + a server-start guard for import-based tests.

Reviewed changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/e2e/tests/09-patient-analytics/wizard.spec.ts	Pass dialect into Shiny context and build condition filters as concept_code vs concept_id depending on dialect; update one SQL template to use `{{CONCEPT_ID1}}`.
plugins/ui/apps/wizards/wizards-config.json	Switch standard wizard condition config paths to `condition_occ_concept_id` and disable free text.
plugins/ui/apps/wizards/wizards-config-hana-lean.json	Switch HANA-lean wizard condition config paths to `conditionconceptcode` and disable free text.
plugins/ui/apps/wizards/src/config/wizardDefinitions.ts	Update mock wizard field config paths to `condition_occ_concept_id`.
plugins/ui/apps/wizards/src/config/cdwConfig.ts	Update mock CDW config + mock attribute values to use `condition_occ_concept_id` with numeric values.
plugins/ui/apps/wizards/CONFIGURATION.md	Update documented attribute mapping to the new condition identifier fields.
plugins/ui/apps/vue-mri-ui-lib/src/utils/shinyDashboardContext.ts	Add `dialect` to Shiny dashboard auth message context.
plugins/ui/apps/vue-mri-ui-lib/src/utils/tests/shinyDashboardContext.test.ts	Extend tests to include dialect and verify defaulting to null.
plugins/ui/apps/vue-mri-ui-lib/src/composables/useDashboardFlow.ts	Fetch dataset dialect from portal and include it in computed dashboard context.
plugins/ui/apps/vue-mri-ui-lib/src/components/ShinyViewer/ShinyDashboardModal.vue	Thread `dialect` prop into the iframe component.
plugins/ui/apps/vue-mri-ui-lib/src/components/ShinyViewer/ShinyDashboardIframe.vue	Include `dialect` in the postMessage auth context sent to Shiny.
plugins/ui/apps/vue-mri-ui-lib/src/components/ChartToolbar.vue	Pass dialect from dashboardFlow context into the Shiny modal.
plugins/functions/parquet-export/index.ts	Add CONCEPT_ID placeholder support, dialect/template mismatch checks, and guard `app.listen` behind `import.meta.main`.
plugins/functions/parquet-export/index.test.ts	Add Deno unit tests for concept_id validation and template substitution behavior.
plugins/functions/mri-pg-config/src/db/seeds/03_Config.ts	Seed new condition identifier attributes for HANA and DuckDB configs (and PA config entries).
plugins/functions/mri-pg-config/src/db/migrations/20260710120000_add_wizard_condition_identifier_attributes.ts	Migration to patch existing stored config JSON with the new condition identifier attributes non-destructively.

 interface DatasetMetadata {
  id: string;
  databaseCode: string;
+  dialect?: string;
  schemaName: string;


    } catch (error) {
      dashboardSelectionError.value = (error as Error)?.message || 'Failed to load dashboard configuration'


jerome-ng · 2026-07-22T14:46:38Z

Thanks for your PR @seolhwa-kim-data4life. I see changes have been merged in https://github.com/data2evidence/wizards-dashboard/pull/24 and this PR focuses on

Providing the dialect to the shiny app
Modifying the configs for the data to be available (concept id) in this case

No 2 is less trivial as the suggested modifications will also change how cohort builder behaves. We will discuss this during the analyze workshop to figure out what should be used and create appropriate issues for both these points.

jerome-ng · 2026-07-27T07:08:10Z

@seolhwa-kim-data4life Hey. Are you able to change the use of concept id to use concept codes instead? The issue is that in order to use concept id, it must be available in the filter cards as an attribute, but we do not want to show the concept id. We can use the SNOMED ones which should be available in the demo dataset. we possibly might not need the dialect anymore if we use the code

The condition typeahead previously returned concept_id for duckdb/postgres datasets and concept_code for HANA, which required plumbing the dataset dialect down to the Shinylive dashboards and exposing a raw concept id as a cohort builder filter card attribute. Return the SNOMED concept_code everywhere instead, so the dashboards send a single CONCEPT_CODEn payload key and the dialect plumbing is unnecessary: - parquet-export: drop the CONCEPT_IDn placeholders, isValidConceptId and the dialect/template mismatch guard - mri-pg-config: the duckdb OMOP config gains a conditionconceptcode attribute (@REF.CONCEPT_CODE, restricted to standard SNOMED conditions); its PA entry stays hidden so no code column reaches the cohort builder - vue-mri-ui-lib: remove the dataset dialect fetch and the dialect field of the postMessage dashboard context - wizards configs and e2e wizard spec: point condition fields at conditionconceptcode and match on quoted {{CONCEPT_CODEn}} in SQL Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

seolhwa-kim-data4life · 2026-07-31T10:27:48Z

@seolhwa-kim-data4life Hey. Are you able to change the use of concept id to use concept codes instead? The issue is that in order to use concept id, it must be available in the filter cards as an attribute, but we do not want to show the concept id. We can use the SNOMED ones which should be available in the demo dataset. we possibly might not need the dialect anymore if we use the code

@jerome-ng now returns the SNOMED concept_code
One open question on point 2: the HANA configs already expose condition_concept_code with filtercard.visible: true, but I've kept the new duckdb entry hidden (visible: false, category: false). Should duckdb match HANA and show it, or stay hidden?

https://github.com/data2evidence/wizards-dashboard/pull/27

here it is now using concept_code back in the pr. idk what to do with vocabulary_id = 'SNOMED' tho. maybe will see with alicia.

The shipped dashboards send "abc123" for condition slots the user left empty; the stub was sending "". Both fail the concept join, but the stub should reproduce the real payload. With this the embedded SQL is byte-identical to sql-scripts/duckdb/cross-sectional-demographics.sql. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

seolhwa-kim-data4life and others added 4 commits July 10, 2026 15:38

Merge branch 'develop' into issue-2843/compatible-wizard

24b7da1

seolhwa-kim-data4life marked this pull request as ready for review July 13, 2026 14:41

Copilot AI review requested due to automatic review settings July 13, 2026 14:41

seolhwa-kim-data4life requested review from LSriragavan, SantanM, brandantck, csafreen, jerome-ng, khairul-syazwan, maggie-li-yd, p-hoffmann and suwarnoong as code owners July 13, 2026 14:41

seolhwa-kim-data4life changed the title ~~Issue 2843/compatible wizard~~ Issue-2843/compatible wizard [help needed] Jul 13, 2026

Copilot started reviewing on behalf of seolhwa-kim-data4life July 13, 2026 14:41 View session

Copilot AI reviewed Jul 13, 2026

View reviewed changes

seolhwa-kim-data4life added 2 commits July 14, 2026 11:20

Merge branch 'develop' into issue-2843/compatible-wizard

1deb73f

Merge branch 'develop' into issue-2843/compatible-wizard

6bc40f5

seolhwa-kim-data4life and others added 2 commits July 31, 2026 15:11

Merge branch 'develop' into issue-2843/compatible-wizard

c72b1d6

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Issue-2843/compatible wizard [help needed] - #2889

Issue-2843/compatible wizard [help needed]#2889
seolhwa-kim-data4life wants to merge 9 commits into
OHDSI:developfrom
seolhwa-kim-data4life:issue-2843/compatible-wizard

seolhwa-kim-data4life commented Jul 13, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

jerome-ng commented Jul 22, 2026 •

edited

Loading

Uh oh!

jerome-ng commented Jul 27, 2026

Uh oh!

seolhwa-kim-data4life commented Jul 31, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

		} catch (error) {
		dashboardSelectionError.value = (error as Error)?.message \|\| 'Failed to load dashboard configuration'

Uh oh!

Conversation

seolhwa-kim-data4life commented Jul 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description #2843

Verified locally

Help needed

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

jerome-ng commented Jul 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jerome-ng commented Jul 27, 2026

Uh oh!

seolhwa-kim-data4life commented Jul 31, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

seolhwa-kim-data4life commented Jul 13, 2026 •

edited

Loading

jerome-ng commented Jul 22, 2026 •

edited

Loading