feat(data-access): GeoExperiment impact-measurement phases + insights/impactTaskId fields

[ssilare-adobe]

What

Adds the shared-model changes for the impact-measurement workflow (see the Impact Validation Engine | Impact Measurement Workflow wiki):

• Two new GeoExperiment.PHASES: impact_measurement_started, impact_measurement_done.
• New nullable insights attribute — the measurement output (ExperimentInsights schema), stored beside metadata.
• New nullable impactTaskId attribute — the Mystique task handle, a first-class field like pre/postScheduleId, so the engine can resume polling across cron ticks.
• Updated index.d.ts types and unit tests.

Why

The experimentation engine keeps an experiment IN_PROGRESS after post-analysis, triggers a Mystique analysis, polls it, and writes the resulting insights back onto the experiment. These fields/phases are the data-model foundation for that.

Notes

• DynamoDB/electrodb — no SQL migration in this repo.
• Unit tests: 2370 passing; 100% coverage on the changed geo-experiment files; lint clean.

🤖 Generated with Claude Code

[github-actions]

This PR will trigger a minor release when merged.

[MysticatBot]

Hey @ssilare-adobe,

⚠ Degraded review - no spec document was found for this change (searched the PR links, the touched repos' docs, the architecture/guidelines docs, and linked Jira). This review covers code-level quality but could not validate the change against an agreed design, so confidence is reduced. Add a spec link (PR template section 4) and re-request review for a full-confidence pass.

Verdict: Approve - clean additive change following established entity patterns exactly.
Complexity: LOW - small diff, single service.
Changes: Adds impact-measurement lifecycle phases and two new nullable fields (impactTaskId, insights) to the GeoExperiment data model (4 files).

Non-blocking (3): minor issues and suggestions

• suggestion: getInsights() return type could use a named ExperimentInsights interface instead of bare object, since the PR description documents a known schema shape (version, generatedAt, analyses) - packages/spacecat-shared-data-access/src/models/geo-experiment/index.d.ts:38. Matches existing metadata/error convention, so optional.
• suggestion: Add a test that sets insights to null to confirm the nullable path works through the model layer - packages/spacecat-shared-data-access/test/unit/models/geo-experiment/geo-experiment.model.test.js:178
• suggestion: Add a negative validation test for impactTaskId with empty string to confirm the hasText validator rejects it - packages/spacecat-shared-data-access/test/unit/models/geo-experiment/geo-experiment.model.test.js:71

---

_{Skill: pr-review | Model: us.anthropic.claude-opus-4-6-v1[1m] | Duration: 1m 8s | Cost: $4.00 | Commit: e8ed764f95c6fd77795015cafeabb3f171b4f398
If this code review was useful, please react with 👍. Otherwise, react with 👎.}

[dipratap]

@ssilare-adobe Please also create DB migration PR in mysticat and validate on dev

[MysticatBot]

Hey @ssilare-adobe,

⚠ Degraded review - no spec document was found for this change (searched the PR links, the touched repos' docs, the architecture/guidelines docs, and linked Jira). This review covers code-level quality but could not validate the change against an agreed design, so confidence is reduced. Add a spec link (PR template section 4) and re-request review for a full-confidence pass.

Verdict: Approve - clean additive change addressing all reviewer feedback, following established patterns.
Complexity: LOW - small diff, single service.
Changes: Adds impact-measurement lifecycle phases, a metadata key for the task ID, and an insightsLocation string field to the GeoExperiment data model (4 files).

Non-blocking (2): minor issues and suggestions

• suggestion: Update the PR description to match the final implementation - it still references impactTaskId as a "first-class field like pre/postScheduleId" and insights as "the measurement output", but the code now stores the task ID in metadata (METADATA_KEYS.IMPACT_MEASUREMENT_TASK_ID) and uses insightsLocation (an S3 pointer, mirroring promptsLocation). The inline discussion makes this clear, but the description will confuse future readers landing via git blame.
• nit: insightsLocation validator (value) => !value || hasText(value) accepts empty string because !'' is truthy - consistent with promptsLocation and the rest of the codebase, so not a regression, but worth knowing if downstream consumers assume non-empty.

Previously flagged, now resolved

• impactTaskId moved from schema attribute to METADATA_KEYS constant (per dipratap's request)
• insights object replaced with insightsLocation S3 pointer string (per dipratap's request)
• TypeScript declarations updated to match

---

_{Skill: pr-review | Model: us.anthropic.claude-opus-4-6-v1[1m] | Duration: 1m 11s | Cost: $2.50 | Commit: dfc8ec7842ca169a567b02066f3a84a847eee636
If this code review was useful, please react with 👍. Otherwise, react with 👎.}

[solaris007]

🎉 This PR is included in version @adobe/spacecat-shared-data-access-v4.2.0 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀