dataflow¶
Haive Dataflow - Registry and Discovery System.
The haive-dataflow package provides a comprehensive system for component discovery, registration, and management in the Haive ecosystem.
It enables automatic discovery of components like agents, tools, engines, and games, handling their registration, dependency management, configuration, and persistence.
Typical usage example:
```python from haive.dataflow import registry_system, EntityType, discover_agents
# Discover and register agents discovered_agents = discover_agents() print(f”Discovered {len(discovered_agents)} agents”)
# Register a custom component registry_system.register_entity(
name=”CustomAgent”, type=EntityType.AGENT, description=”A custom agent implementation”, module_path=”my_module.agents”, class_name=”CustomAgent”
)
# Query components agents = registry_system.get_entities_by_type(EntityType.AGENT) ```
This package consists of several modules:
registry: Core registry functionality for component management discovery: Component discovery mechanisms serialization: Tools for serializing/deserializing complex objects db: Database integration for persistence api: API endpoints for accessing registry data providers: Provider implementations for various services
Submodules¶
- dataflow.api
- dataflow.auth
- dataflow.bin
- dataflow.config
- dataflow.conversations
- dataflow.core
- dataflow.db
- dataflow.discovery
- dataflow.fetchers
- dataflow.importers
- dataflow.internal_websockets
- dataflow.llms
- dataflow.mcp
- dataflow.models
- dataflow.persistence
- dataflow.platform
- dataflow.providers
- dataflow.registries
- dataflow.registry
- dataflow.serialization
- dataflow.server_management
- dataflow.utils