Skip to content

S28 · Pipeline skill

Unify and enrich records

Merge source records into one catalog, identify gaps and have reviewers accept additional facts.

Heterogeneous ingestion + entity resolution + quality checks and human enrichment

Pipeline: 4 steps
Human review: 1
Components: 10

Proposed composition; integration pendingSynthetic data

Download · ZIP · 19KB Pipeline

Unify and enrich records — Synthetic walkthrough of inputs, step handoffs, human review and the final result. Live services still need integration.

01Pipeline

Base pipeline

Ordered steps. Each consumes the previous output; review gates hold until a person accepts.

Starts from: Authorized sources, legacy records and target catalog fields
Delivers: Human-accepted catalog facts and references

Review contract

People decide at review gates; components supply data, drafts or review surfaces. Unaccepted outputs do not advance.

Integration note

Collectors target different sources; define field contracts before merging and replace hotel-specific taxonomies where needed.

01
Ingest and stage sources
Input
Authorized sources, legacy records and target catalog fields
Output
Staged records with provenance and missing-data notes
- Travel master and inventory synchronizationUpstream record staging + import audits and snapshots
- Cruise itinerary, fare and ship crawlingBrowser batch collection + structured details and progress
- Legacy data and provenance migrationLegacy record mapping and import + orphan and completeness checks
02
Normalize and resolve entities
Input
Staged records with provenance and missing-data notes
Output
Canonical entities, detail relations and aliases
- Multi-provider cruise catalog normalizationNested source data normalization + relational tables and error reports
- Hotel catalog and media ingestionQueued supplier catalog ingestion + media and price caches
- Canonical geography, hotels and POIsHeterogeneous name resolution + canonical entities and aliases
03
Classify and locate gaps
Input
Canonical entities, detail relations and aliases
Output
Classification results and missing-fact list
- Semantic hotel taxonomy assignmentVersioned taxonomy rules + source-linked tag assignments
- Grounding coverage and repair planningFact coverage checks + targeted enrichment tasks
04
Investigate and accept facts
Human review
Input
Classification results and missing-fact list
Output
Human-accepted catalog facts and references
Decided byAssigned reviewer or requesting user
- Web-grounded destination and hotel researchEntity fact research + structured provenance snapshots
- Destination and hotel grounding cardsEntity-based fact retrieval + fit reasons and evidence

02Industry versions

Industry versions

Travel & hotels Source-supported industry Merge destination, hotel and voyage records from suppliers and fill missing facts before adding them to the catalog.

Adaptation Each supplier needs a field contract and retained conflicts and provenance; catalog wiring remains to be implemented.

01
Collect and stage supplier records
Input
Permitted supplier data, legacy catalog and canonical fields
Output
Staged hotel and voyage data with provenance
- Travel master and inventory synchronizationUpstream record staging + import audits and snapshots
- Cruise itinerary, fare and ship crawlingBrowser batch collection + structured details and progress
- Legacy data and provenance migrationLegacy record mapping and import + orphan and completeness checks
02
Resolve places and accommodation
Input
Staged hotel and voyage data with provenance
Output
Canonical entities, details and aliases
- Multi-provider cruise catalog normalizationNested source data normalization + relational tables and error reports
- Hotel catalog and media ingestionQueued supplier catalog ingestion + media and price caches
- Canonical geography, hotels and POIsHeterogeneous name resolution + canonical entities and aliases
03
Check tags and fact gaps
Input
Canonical entities, details and aliases
Output
Classification results and missing amenity facts
- Semantic hotel taxonomy assignmentVersioned taxonomy rules + source-linked tag assignments
- Grounding coverage and repair planningFact coverage checks + targeted enrichment tasks
04
Editor reviews added facts
Human review
Input
Classification results and missing amenity facts
Output
Accepted catalog facts and sources
Decided byAssigned reviewer or requesting user
- Web-grounded destination and hotel researchEntity fact research + structured provenance snapshots
- Destination and hotel grounding cardsEntity-based fact retrieval + fit reasons and evidence

Travel & hotels · Solutions →

03Parameters

Parameters

From the pack’s params.schema.json. Choose values to build a synthetic request.

skills/_pipelines/unify-reference-catalog/params.schema.json

locale · required industry · required adapter_mode · required Human review · 4 Investigate and accept facts

Request (synthetic)

{
  "synthetic": true,
  "run_id": "synthetic-example",
  "locale": "en",
  "industry": "base",
  "adapter_mode": "local",
  "records": [
    {
      "id": "synthetic-record-1",
      "synthetic": true,
      "text": "Synthetic observation A: a draft needs review."
    },
    {
      "id": "synthetic-record-2",
      "synthetic": true,
      "text": "Synthetic observation B: retain the unresolved question."
    }
  ],
  "decisions": {}
}

04Member components

Member components

Travel master and inventory synchronization

Upstream record staging + import audits and snapshots
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Cruise itinerary, fare and ship crawling

Browser batch collection + structured details and progress
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Legacy data and provenance migration

Legacy record mapping and import + orphan and completeness checks
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Multi-provider cruise catalog normalization

Nested source data normalization + relational tables and error reports
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Hotel catalog and media ingestion

Queued supplier catalog ingestion + media and price caches
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Canonical geography, hotels and POIs

Heterogeneous name resolution + canonical entities and aliases
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Semantic hotel taxonomy assignment

Versioned taxonomy rules + source-linked tag assignments
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Grounding coverage and repair planning

Fact coverage checks + targeted enrichment tasks
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Web-grounded destination and hotel research

Entity fact research + structured provenance snapshots
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline
Destination and hotel grounding cards

Entity-based fact retrieval + fit reasons and evidence
- Travel & hotels · Source-supported industry
- Manufacturing & industry · Transferable use
Working 1 pipeline

05Skills pack

Skills pack

Runnable reference on synthetic fixtures with tests. Domain contracts, services and approvals still need integration.

unify-reference-catalog-1.0.0.zip

Contains SKILL.md, pipeline.json, params.schema.json, runner, tests, fixtures, integration checklist and sources.

Synthetic workflow reference; service integration pending

Download · 19KB

SHA-256 286bf3859a2cac8c81fba8bb3dcef2a100f55812cda2cf079628ab81c352b384