Cloud Shapes is an event-driven database designed to dynamically generate, update, and delete projections based on CloudEvents. It supports real-time relationship management, embedded data synchronization, and write-time updates for efficient, event-driven materialization of data.
✅ Projection-Based Data Model – Store and query data as projections, materialized from events.
✅ Event-Driven Writes – Automatically create, update, or delete projections when CloudEvents occur.
✅ Schema-Enforced Storage – Validate projection states using JSON Schema (JsonSchema).
✅ Real-Time Relationship Updates – Maintain data consistency with embedded relationships.
✅ Write-Time Data Embedding – Merge related projection data at write time for optimized queries.
✅ Indexing & Querying – Define indexes for fast lookups on projection properties.
✅ Multi-Tenant & Scalable – Designed for distributed event-driven architectures.
-
Define a
ProjectionType- Each projection type is defined by a schema (
JsonSchema). - Triggers determine when new projections should be created from incoming CloudEvents.
- Each projection type is defined by a schema (
-
CloudEvents Ingestion
- Events trigger projections updates, ensuring real-time data synchronization.
-
Write-Time Relationship Management
- Projections can embed related projection data and update it dynamically when relationships change.
-
Indexing & Querying
- Projections can be indexed for efficient lookups and optimized queries.
A projection type defines how CloudEvents affect projections:
name: Order
description: >
The `Order` projection represents an **order entity**, materialized from CloudEvents.
It includes **embedded customer and item data** to ensure real-time consistency and optimized query performance.
schema:
type: object
indexes:
- name: store
fields: [ store ]
unique: false
descending: false
text: false
triggers:
create:
- event:
type: io.petstore.events.order.created.v1
correlation:
attribute: subject
state: '${ .data }'
update:
- event:
type: io.petstore.events.order.updated.v1
correlation:
attribute: subject
state: '${ .data }'
delete:
- event:
type: io.petstore.events.order.deleted.v1
correlation:
attribute: subject
relationships:
- type: one-to-one
target: Customer
key: customerId
path: customer
- type: one-to-many
target: OrderItem
key: orderId
path: itemsWhen an order is processed, the system materializes the following JSON projection:
```json
{
"id": "order-123",
"customer": {
"id": "customer-456",
"name": "Alice Smith"
},
"items": [
{ "id": "item-1", "product": "Laptop", "price": 1000 },
{ "id": "item-2", "product": "Mouse", "price": 50 }
],
"status": "Processing"
}
```
🔹 Embedded relationships (customer, items) ensure real-time updates when related projections change.
Cloud Shapes dynamically updates projections when CloudEvents are received:
order.created→ Creates a new projectionorder.updated→ Modifies the projectionorder.deleted→ Removes the projection
🔹 Embedded relationships are updated dynamically at write time, ensuring consistency.
✅ Fully Event-Driven – No manual updates, all changes are event-driven.
✅ Real-Time Consistency – Embedded data updates instantly on related projection changes.
✅ Write-Time Optimizations – Data embedding reduces query-time overhead.
✅ Schema-Enforced Validation – Ensures data integrity using JSON Schema (JsonSchema).
✅ Scalable & Distributed – Built for high-throughput event-driven systems.
1️⃣ Clone the Cloud Shapes repository:
git clone https://github.com/neuroglia-io/cloud-shapes.git2️⃣ Start Cloud Shapes using Docker:
docker-compose -f deployments/docker-compose/docker-compose.yml up -d3️⃣ Define Projection Types and their Triggers
4️⃣ Start ingesting CloudEvents into the system.
5️⃣ Query real-time projections with pre-embedded data.
- 📖 Documentation (Coming Soon)
- 🚀 API Reference (Coming Soon)
Cloud Shapes is open-source and available under the Apache-2.0 License.