Architecture

MTS1B is a layered ecosystem. Each repo has exactly one role; dependencies flow downward only.

Dependency layers

┌────────────────────────────────────────────────────────────────────┐
│ Layer 7 — Frontends + Community                                    │
│   mts1b-frontends    mts1b-githubbot    mts1b-discordbot           │
└────────────────────────────────────────────────────────────────────┘
                              ▲
┌────────────────────────────────────────────────────────────────────┐
│ Layer 6 — Asset-class services                                     │
│   mts1b-treasury    mts1b-sports    mts1b-prediction-markets       │
└────────────────────────────────────────────────────────────────────┘
                              ▲
┌────────────────────────────────────────────────────────────────────┐
│ Layer 5 — Strategy + execution                                     │
│   mts1b-research    mts1b-oms    mts1b-tradingview                 │
└────────────────────────────────────────────────────────────────────┘
                              ▲
┌────────────────────────────────────────────────────────────────────┐
│ Layer 4 — Engines + ops                                            │
│   mts1b-GPUbacktester  mts1b-datalake  mts1b-llm  mts1b-operations │
│   mts1b-cloudburst     mts1b-reportslibrary                        │
└────────────────────────────────────────────────────────────────────┘
                              ▲
┌────────────────────────────────────────────────────────────────────┐
│ Layer 3 — Risk + portfolio + execution algos                       │
│   mts1b-riskengine    mts1b-portfolio    mts1b-oms-algos           │
└────────────────────────────────────────────────────────────────────┘
                              ▲
┌────────────────────────────────────────────────────────────────────┐
│ Layer 2 — Adapters + quant library                                 │
│   mts1b-brokers   mts1b-marketdata   mts1b-altdata                 │
│   mts1b-cryptodata  mts1b-macrodata  mts1b-quantkit                │
└────────────────────────────────────────────────────────────────────┘
                              ▲
┌────────────────────────────────────────────────────────────────────┐
│ Layer 1 — Platform                                                 │
│   mts1b-platform   mts1b-deploy   mts1b-pluginsdk                  │
└────────────────────────────────────────────────────────────────────┘
                              ▲
┌────────────────────────────────────────────────────────────────────┐
│ Layer 0 — Foundation                                               │
│   mts1b-foundation  ← types + schemas, zero runtime deps           │
└────────────────────────────────────────────────────────────────────┘

Hard rules

1. No upward dependencies. A higher layer can depend on a lower layer, never the reverse.
2. No circular dependencies. Enforced by tests/contracts/test_dependency_direction.py in every repo's CI.
3. mts1b-foundation has zero runtime deps (except pydantic + typing-extensions).
4. All types cross the boundary in mts1b-foundation form. No direct service-to-service custom types.
5. AST-based deduplication — 17 protected symbols (HRP, BL, Kelly, Sharpe, ...) are defined in exactly one repo each. CI fails on duplicates.

Data flow at runtime

marketdata    ──┐
altdata       ──┼─→ datalake ──→ research ──→ portfolio ──→ riskengine ──→ oms ──→ brokers
cryptodata    ──┤                                            ▲                ▲
macrodata     ──┘                              quantkit ─────┘                │
                                                                              │
                                              GPUbacktester (offline)         │
                                                                              │
                                              treasury ─── allocates NAV ─────┘

• operations observes everything via the NATS event bus and enforces compliance + halts.
• reportslibrary reads result parquets and generates trade reports + postmortems.
• llm is available to any service that wants AI assistance.

Why this layering?

• Replaceability. You should be able to swap any repo for an alternative implementation as long as it speaks the mts1b-foundation interface.
• Clarity. Reading layer 5 code, you know exactly which lower-layer primitives are available.
• Testability. Lower layers have stricter test coverage requirements (foundation = 100%, platform = 90%, ...).
• Community contribution. A contributor working on mts1b-marketdata doesn't need to understand mts1b-research to make a useful PR.

See dependency-graph for the full edge list.