mirror of
https://github.com/diegosouzapw/OmniRoute.git
synced 2026-08-26 17:12:27 +03:00
b2b568b08e
* fix routed target request parameters * chore: rerun CI * test(chatcore): align PR #7323 Codex-routing test with #7533 verbosity gating The "Codex Responses routing keeps reasoning effort while dropping GPT-only verbosity" test translated a Responses-shape request with credentials=null and only the positional `provider` arg set to "opencode-go". #7533's verbosity carry-over (Responses `text.verbosity` -> Chat `verbosity`) reads the destination from `credentials.provider`, which the source->openai translation step never threads from the positional provider arg — so `translated.verbosity` came back undefined instead of "low", failing before prepareUpstreamBody's sanitizer was even reached. The test's intent (per its own name/comment) is a combo/fallback reroute: translate while still addressed at Codex (an #7533-allowlisted OpenAI-param destination, so verbosity legitimately survives translateRequest), then resolve the final upstream target to opencode-go/GLM so prepareUpstreamBody's sanitizeRequestForResolvedTarget (#7050/#7533) strips the GPT-only verbosity for that concrete target while preserving reasoning_effort. Fixed by passing `credentials: { provider: "codex" }` to the first translateRequest call to match how production actually carries the destination provider, instead of relying on the provider positional argument. No production code changed — #7050 and #7533's sanitization are intentional and protected; only the test's setup was misaligned with them. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com> --------- Co-authored-by: Diego Rodrigues de Sa e Souza <diegosouza.pw@gmail.com> Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
80 lines
4.3 KiB
Markdown
80 lines
4.3 KiB
Markdown
# open-sse/services/ — Routing Engine & Cross-Cutting Services
|
|
|
|
**Purpose**: 134 service modules (top-level) powering request routing, rate limiting, quota management, token refresh, fallback strategies, and runtime state. The combo routing engine (`combo.ts`) is the core; supporting services handle resilience, accounting, and decision-making.
|
|
|
|
Live count: `ls open-sse/services/*.ts | wc -l` (currently 134). More including sub-dirs like `autoCombo/` and `compression/`.
|
|
|
|
---
|
|
|
|
## Combo Routing Engine
|
|
|
|
- **`combo.ts`** — Entry point for multi-model routing. **`handleComboChat()`** iterates through targets in order until success or all fail. **`resolveComboTargets()`** expands combo config into ordered `ResolvedComboTarget[]` (provider + model + account + credentials).
|
|
- **Strategies** (17): `priority`, `weighted`, `fill-first`, `round-robin`, `P2C`, `random`, `least-used`, `reset-aware`, `reset-window`, `cost-optimized`, `strict-random`, `auto`, `lkgp`, `context-optimized`, `context-relay`, `headroom`, `fusion`. Source: `ROUTING_STRATEGY_VALUES` in `src/shared/constants/routingStrategies.ts`.
|
|
- Each target calls **`handleSingleModel()`** which wraps `handleChatCore()` with per-target error handling and circuit breaker checks.
|
|
|
|
## Key Services
|
|
|
|
### Quota & Rate Limiting
|
|
|
|
- **`rateLimitManager.ts`** — Token bucket per API key + provider combo. Rejects before dispatch.
|
|
- **`usage.ts`** — Per-request token/cost consumption tracking.
|
|
- **`quotaCache.ts`** — In-memory quota snapshots, pre-loaded at startup.
|
|
|
|
### Account & Token Management
|
|
|
|
- **`tokenRefresh.ts`** — OAuth token expiration detection and refresh.
|
|
- **`accountFallback.ts`** — Account switching on quota/rate-limit. Also houses model lockout.
|
|
- **`sessionManager.ts`** — Request session state across retries.
|
|
|
|
### Request Routing & Intelligence
|
|
|
|
- **`wildcardRouter.ts`** — Wildcard route matching in combo configs.
|
|
- **`intentClassifier.ts`** — Request intent classification for intelligent routing.
|
|
- **`taskAwareRouter.ts`** — Task-type-based routing (reasoning → o1, code-gen → Cursor).
|
|
- **`targetRequestSanitizer.ts`** — Final provider/model-aware parameter sanitation after routing resolution and before executor dispatch.
|
|
- **`thinkingBudget.ts`** — Thinking token allocation for o1/o3 models.
|
|
- **`contextManager.ts`** — Routing context injection (system prompts, memory).
|
|
|
|
### Model Lifecycle & Fallback
|
|
|
|
- **`modelDeprecation.ts`** — Deprecated model detection and successor routing.
|
|
- **`modelFamilyFallback.ts`** — T5 intra-family fallback chains.
|
|
- **`emergencyFallback.ts`** — Last-resort fallback to stable free providers.
|
|
|
|
### State & Detection
|
|
|
|
- **`workflowFSM.ts`** — Multi-turn workflow state machine.
|
|
- **`backgroundTaskDetector.ts`** — Long-running task detection for batch routing.
|
|
- **`ipFilter.ts`** — IP-based routing rules.
|
|
- **`signatureCache.ts`** — Request signature caching for deduplication.
|
|
- **`volumeDetector.ts`** — Volume spike detection for rate-limit escalation.
|
|
- **`contextHandoff.ts`** — Session context serialization for A2A handoff.
|
|
|
|
### Prompt Compression Pipeline (`compression/`)
|
|
|
|
- **`strategySelector.ts`** — Compression mode selection (off/lite/standard/aggressive/ultra/rtk/stacked).
|
|
- **`lite.ts`** — 5 lite techniques (whitespace, dedup, tool results, redundant removal, image URLs).
|
|
- **`caveman.ts` / `cavemanRules.ts`** — Caveman-style semantic condensation with rule packs.
|
|
- **`engines/rtk/`** — RTK tool-output compression (command detection, JSON filters, dedup, truncation).
|
|
- **`engines/registry.ts`** — Engine registry for standalone and stacked pipelines.
|
|
- **`stats.ts`** — Per-request compression stats.
|
|
- **`types.ts`** — Shared types (`CompressionMode`, `CompressionConfig`, `CompressionStats`).
|
|
|
|
---
|
|
|
|
## Adding a New Service
|
|
|
|
1. Create `open-sse/services/[serviceName].ts`
|
|
2. Export main handler function
|
|
3. Add unit tests in `tests/unit/services/`
|
|
4. Integrate into `handlers/chatCore.ts` (if routing-related) or `combo.ts`
|
|
5. Document in this file
|
|
|
|
## Anti-Patterns
|
|
|
|
- Synchronous DB calls in `combo.ts` hot path — pre-compute and cache
|
|
- Retry logic in handlers — use `retry()` from resilience service
|
|
- Direct provider config access — use `providerRegistry` getter functions
|
|
- Hardcoded fallback chains — define in `modelFamilyFallback.ts`
|
|
- State mutations across concurrent requests — use request-scoped context only
|