跳到主要内容

Copy provider marketplace (Phase 4)

Browsable directory of copy strategies. Provider publishes copy profile as listing with verified-live badge (strategy source account trade real money, not demo) plus performance fee. Followers browse marketplace, ranked by performance score projected from execution-transparency data.

Model​

  • CopyProviderListing = aggregate: UserId, ProfileId, display name, description, performance fee, VerifiedLive, Published + PublishedAt. One listing per profile (unique index).
  • Verified-live derived at publish time from profile source TradingAccount.IsLive — provider can't self-assert.
  • Performance stats not stored on listing — read-model projection over CopyExecution transparency log (fill rate, avg latency, avg realized slippage), so marketplace always reflect live execution quality.

Ranking​

CopyEndpoints.MarketplaceScore(fillRate, avgLatencyMs, avgSlippagePoints, verifiedLive) → 0–100 score: fill rate dominates (×60), low latency + low slippage add (×20 each), verified-live badge add small trust bonus. Deterministic + monotonic, so ordering stable.

API​

  • POST /api/copy/profiles/{id}/publish — publish/update profile listing (DisplayName, Description, PerformanceFeePercent); verified-live set from source account.
  • DELETE /api/copy/profiles/{id}/publish — unpublish.
  • GET /api/copy/marketplace — all published listings, ranked, each with performance summary (executions, fill rate, avg latency, avg slippage, score) + verified-live badge.

Tests​

  • Unit (CopyProviderListingTests) — aggregate invariants: display name required; publish set timestamp; unpublish hide; update replace display fields + fee + badge.
  • Integration (CopyMarketplaceTests, real Postgres) — published listing persist with badge; one listing per profile (unique index); ranking score prefer verified/high-fill providers.

Copy host untouched (listings + read model only), so copy DST stress suite unaffected.