ManagedCode Agent Lightning Migration Plan This plan tracks parity work between external/microsoft-agent-lightning (Python) and the C# port.
✅ Complete in C#
🚧 Planned / not yet ported
❓ Needs investigation / decide if we port
Component
Python Source
Status
Notes
Domain models
agentlightning/types/core.py✅
ManagedCode.AgentLightning.Core/Models – rollout/attempt/triplet + hooks
Tracing models
agentlightning/types/tracer.py✅
ManagedCode.AgentLightning.Core/Tracing – span DTOs & helpers with OpenTelemetry tests
Resources
agentlightning/types/resources.py✅
ManagedCode.AgentLightning.Core/Resources – LLM/proxy/prompt resources mirrored
LitAgent base
agentlightning/litagent/litagent.py✅
LitAgentBase<T> with hook lifecycle, LightningAgent derives from it
Adapter infrastructure
agentlightning/adapter/base.py✅
Adapters/Adapter, TraceAdapter, TraceToMessagesAdapter implemented with tests
Runner infrastructure
agentlightning/runner/base.py✅
LitAgentRunner processes rollouts via LightningAgent and stores spans
Store interface
agentlightning/store/base.py✅
ILightningStore contract + InMemoryLightningStore covering queue/attempt/span lifecycle
Trainer orchestration
agentlightning/trainer/trainer.py✅
Trainer orchestrates batches via store + runner (tested)
Adapter
Python Source
Status
Notes
Trace → messages
adapter/messages.py✅
TraceToMessagesAdapter translates GenAI spans into OpenAI chat payloads
Trace → triplets
adapter/triplet.py✅
TracerTraceToTripletAdapter exports triplets with reward policies
OTEL trace adapter
adapter/base.py🚧
Hook Activity -> SpanModel bridging
Component
Python Source
Status
Notes
LightningStore (async)
store/base.py✅
ILightningStore exposes start/enqueue/start-attempt, span sequencing, and wait semantics
In-memory store
store/memory.py✅
Expanded store handles attempts, spans, resources, and polling waits with thread-safe state
Client/server bridge
store/client_server.py❓
Decide ASP.NET hosting approach
Runner execution strategies
execution/*🚧
C# runner supports single-step execution, retry-aware polling, and resource hydration; parallel orchestration still pending
Algorithms & Training Pipelines
Component
Python Source
Status
Notes
Algorithm base class
algorithm/base.py🚧
Define async lifecycle (SetupAsync, TrainAsync, TeardownAsync) with dataset plumbing
APO (Automatic Prompt Optimization)
algorithm/apo/apo.py🚧
Requires prompt diffing, versioned templates, and evaluation harness
Trainer legacy compat
trainer/legacy.py🚧
Implement legacy hooks while aligning with new runner/store abstractions
Trainer orchestration
trainer/trainer.py🚧
Port training loop, scheduler, and algorithm/run coordination
Registry/config utilities
trainer/registry.py, trainer/init_utils.py🚧
Recreate component registration and config binding over Options
Component
Python Source
Status
Notes
Reward emitters
emitter/reward.py, reward.py🚧
Implement reward span helpers with OTEL integration
Message/object emitters
emitter/message.py, emitter/object.py, emitter/utils.py🚧
Required for parity in trace adapters
Instrumentation (AgentOps, LiteLLM, vLLM)
instrumentation/*❓
Determine .NET bindings and optionality
Logging utilities
logging.py✅
Replaced with Microsoft.Extensions.Logging configuration helpers
Area
Status
Notes
Python fixture import
🚧
Need harness to reuse JSON/SQLite fixtures from submodule
Integration test parity
🚧
Blocked until adapters, store, runner port complete
Docs & README updates
🚧
Document hosting, configuration, and migration progress
Packaging & CI
✅
.NET solution, format/test gates, and workflows in place
Component
Python Source
Status
Notes
Logging helpers
logging.py✅
Using Microsoft.Extensions.Logging
Legacy server/client
server.py, client.py❓
Decide on support for legacy flows
Area
Status
Notes
Core models & resources
✅
Unit tests in ManagedCode.AgentLightning.Tests
Span conversions
✅
Tracing/SpanModelTests
Resource helper coverage
✅
Resources/ResourceModelTests
Adapter tests
🚧
Need to mirror upstream fixtures
Runner/store/trainer integration
🚧
Blocked until components ported
.NET 9 solution scaffolding with central package management
Core rollout/attempt models and runtime scaffolding (LightningAgent + LocalChatClient)
CI/CodeQL/release workflows (ManagedCode templates)
Span/resource models with OpenTelemetry conversions and unit coverage
Expand runner execution strategies (parallel workers, retries, resource coordination).
Reproduce key Python fixtures/tests for adapters, store logic, and integration flows.
Stand up algorithm/trainer scaffolding (base class, APO components, legacy compat).
Implement reward/message emitter instrumentation and vendor integration bindings.
Update this document whenever a component moves between statuses or when new design decisions affect parity.