pjandro/OmniRoute
1
0
Fork 0
mirror of https://github.com/diegosouzapw/OmniRoute.git synced 2026-08-26 09:02:11 +03:00
Code Issues Packages Projects Releases Wiki Activity

merge release/v3.8.51 into feat/subscription-first-routing (baseline reconciled)

Browse Source
This commit is contained in:
Markus Hartung
2026-08-25 01:43:43 -03:00
parent 0a6cdbe0f6 a0ceccc6f0
commit 4caecc8d48
800 changed files with 47188 additions and 16600 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
docs/routing/AUTO-COMBO.md
View File

@@ -183,30 +183,31 @@ See [#7992](https://github.com/diegosouzapw/OmniRoute/issues/7992) and [#7111](h
## How It Works (Persisted Auto-Combos)
The Auto-Combo Engine dynamically selects the best provider/model for each request using a **14-factor scoring function** (defined in `open-sse/services/autoCombo/scoring.ts` → `DEFAULT_WEIGHTS`). Weights form a normalized distribution (custom weights are renormalized by `normalizeScoringWeights()`).
The Auto-Combo Engine dynamically selects the best provider/model for each request using a **15-factor scoring function** (defined in `open-sse/services/autoCombo/scoring.ts` → `DEFAULT_WEIGHTS`). The default weights sum to `1.0`; custom weights are renormalized by `normalizeScoringWeights()`.
![Auto-Combo 14-factor scoring](../diagrams/exported/auto-combo-12factor.svg)
![Auto-Combo 15-factor scoring](../diagrams/exported/auto-combo-12factor.svg)
> Source: [diagrams/auto-combo-12factor.mmd](../diagrams/auto-combo-12factor.mmd) (regenerate via `npm run docs:render-diagrams`). The filename predates the current factor set; the diagram shows 13 of the 14 factors (missing `sessionAvailability`).
> Source: [diagrams/auto-combo-12factor.mmd](../diagrams/auto-combo-12factor.mmd) (regenerate via `npm run docs:render-diagrams`). The filename is historical; the source and rendered diagram show all 15 factors declared in `DEFAULT_WEIGHTS`.
| Factor | Default Weight | Description |
| :-------------------- | :------------- | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `health` | 0.20 | Health score from circuit breaker (CLOSED=1.0, HALF_OPEN=0.5, OPEN=0.0) |
| `quota` | 0.15 | Remaining quota / rate-limit headroom [0..1] |
| `costInv` | 0.15 | Inverse **blended** cost (60% input + 40% output token price, normalized) — cheaper = higher score |
| `latencyInv` | 0.12 | Inverse p95 latency normalized to pool — faster = higher score |
| `taskFit` | 0.08 | Task-type fitness (coding, review, planning, analysis, debugging, docs) |
| `stability` | 0.05 | Variance-based stability (low latency stdDev / error rate) |
| `tierPriority` | 0.05 | Account-tier priority — Ultra=1.0, Pro=0.67, Standard=0.33, Free=0.0 |
| `tierAffinity` | 0.05 | Affinity between the candidate's tier and the manifest-recommended tier |
| `specificityMatch` | 0.05 | Match between request specificity (manifest hint) and model tier |
| `contextAffinity` | 0.05 | Affinity between the request's context-window need and the model's context window |
| `sessionAvailability` | 0.05 | OAuth session availability of the candidate connection for this session (`getOAuthSessionAvailability()`; non-OAuth connections score 1.0) |
| `connectionDensity` | 0.05 | Spreads load across connections of the same provider (anti-concentration) |
| `quota` | 0.1429 | Remaining quota / rate-limit headroom [0..1] |
| `health` | 0.1605 | Health score from circuit breaker (CLOSED=1.0, HALF_OPEN=0.5, OPEN=0.0) |
| `costInv` | 0.1429 | Inverse **blended** cost (60% input + 40% output token price, normalized) — cheaper = higher score |
| `latencyInv` | 0.1143 | Inverse p95 latency normalized to pool — faster = higher score |
| `taskFit` | 0.0762 | Task-type fitness (coding, review, planning, analysis, debugging, docs) |
| `stability` | 0.0476 | Variance-based stability (low latency stdDev / error rate) |
| `tierPriority` | 0.0476 | Account-tier priority — Ultra=1.0, Pro=0.67, Standard=0.33, Free=0.0 |
| `tierAffinity` | 0.0476 | Affinity between the candidate's tier and the manifest-recommended tier |
| `specificityMatch` | 0.0476 | Match between request specificity (manifest hint) and model tier |
| `contextAffinity` | 0.0476 | Affinity between the request's context-window need and the model's context window |
| `sessionAvailability` | 0.0476 | OAuth session availability of the candidate connection for this session (`getOAuthSessionAvailability()`; non-OAuth connections score 1.0) |
| `connectionDensity` | 0.0476 | Spreads load across connections of the same provider (anti-concentration) |
| `cacheAffinity` | 0.00 | Rendezvous-hash affinity toward the connection likeliest to already hold this request's prompt-cache prefix (`open-sse/services/combo/promptCacheAffinity.ts`); disabled by default (#8008) |
| `resetWindowAffinity` | 0.00 | Bias toward connections whose quota reset window is favorable (disabled by default) |
| `quality` | 0.03 | Feedback-driven output-quality signal from the routing-event quality tracker; candidates without observations receive a neutral 0.5 |
**Sum:** `0.20 + 0.15 + 0.15 + 0.12 + 0.08 + 0.05 + 0.05 + 0.05 + 0.05 + 0.05 + 0.05 + 0.05 + 0.00 + 0.00 = 1.05` as literally declared in `DEFAULT_WEIGHTS`; user-configured weights are renormalized into a distribution by `normalizeScoringWeights()` before scoring.
**Sum:** `0.1429 + 0.1605 + 0.1429 + 0.1143 + 0.0762 + (7 × 0.0476) + 0.00 + 0.00 + 0.03 = 1.0` as declared in `DEFAULT_WEIGHTS`; user-configured weights are renormalized into a distribution by `normalizeScoringWeights()` before scoring.
## Mode Packs
@@ -677,8 +678,8 @@ Including the bare `auto` (default) plus the 6 `AutoVariant` values declared in
## How tiers fit Auto-Combo
The 14-factor scoring function (`open-sse/services/autoCombo/scoring.ts`) treats tier
membership as two signals: `tierPriority` (0.05) and `tierAffinity` (0.05). See the
The 15-factor scoring function (`open-sse/services/autoCombo/scoring.ts`) treats tier
membership as two signals: `tierPriority` (0.0476) and `tierAffinity` (0.0476). See the
canonical [scoring factor table](#how-it-works-persisted-auto-combos) above for the full
`DEFAULT_WEIGHTS` set — the per-pack overrides (ship-fast/cost-saver/quality-first/
offline-friendly) are listed in the "Weight profiles per pack" table.