Architecture Case Study
Search & Information Flow
How product information is indexed and how customer search requests flow through the architecture.
Indexing flow
sequenceDiagram
participant Source as Travel source
participant Transform as Transformation
participant Search as Search index
Source->>Transform: Product records
Transform->>Transform: Normalise searchable fields
Transform->>Search: Add / update documents
Search-->>Transform: Indexing task status
The transformation step owns the mapping between operational product data and the search contract.
This keeps search-specific concerns out of the source system.
Customer query flow
sequenceDiagram
participant Customer
participant Webflow
participant Search as Meilisearch
Customer->>Webflow: Search or change filters
Webflow->>Search: Query + facets + sort + page
Search-->>Webflow: Hits + facet distribution
Webflow-->>Customer: Updated cards and controls
Filter semantics
Within one facet, selected values can represent alternatives.
Across different dimensions, constraints combine.
That distinction belongs in the search contract rather than being reimplemented inconsistently across UI components.
Information ownership
The index does not become the system of record.
Its contents are a search projection derived from authoritative travel data and can be rebuilt when required.
That ownership model simplifies recovery, schema evolution and governance.