mirror of
https://github.com/diegosouzapw/OmniRoute.git
synced 2026-04-28 14:29:54 +00:00
docs(i18n): sync documentation updates to 32 languages
This commit is contained in:
diegosouzapw
parent
857b692aac
commit
e5c4e450c0
2062 changed files with 257153 additions and 64249 deletions
|
|
@ -4,15 +4,19 @@
|
|||
|
||||
---
|
||||
|
||||
> Un guide complet et convivial pour les débutants sur le routeur proxy IA multifournisseur**omniroute**.---
|
||||
> A comprehensive, beginner-friendly guide to the **omniroute** multi-provider AI proxy router.
|
||||
|
||||
---
|
||||
|
||||
## 1. What Is omniroute?
|
||||
|
||||
omniroute est un**routeur proxy**qui se situe entre les clients IA (Claude CLI, Codex, Cursor IDE, etc.) et les fournisseurs d'IA (Anthropic, Google, OpenAI, AWS, GitHub, etc.). Cela résout un gros problème :
|
||||
omniroute is a **proxy router** that sits between AI clients (Claude CLI, Codex, Cursor IDE, etc.) and AI providers (Anthropic, Google, OpenAI, AWS, GitHub, etc.). It solves one big problem:
|
||||
|
||||
> **Different AI clients speak different "languages" (API formats), and different AI providers expect different "languages" too.**omniroute translates between them automatically.
|
||||
> **Different AI clients speak different "languages" (API formats), and different AI providers expect different "languages" too.** omniroute translates between them automatically.
|
||||
|
||||
Considérez-le comme un traducteur universel aux Nations Unies : n'importe quel délégué peut parler n'importe quelle langue, et le traducteur la convertit pour n'importe quel autre délégué.---
|
||||
Think of it like a universal translator at the United Nations — any delegate can speak any language, and the translator converts it for any other delegate.
|
||||
|
||||
---
|
||||
|
||||
## 2. Architecture Overview
|
||||
|
||||
|
|
@ -61,43 +65,44 @@ graph LR
|
|||
|
||||
### Core Principle: Hub-and-Spoke Translation
|
||||
|
||||
Toutes les traductions de format passent par le**format OpenAI comme hub** :```
|
||||
Client Format → [OpenAI Hub] → Provider Format (request)
|
||||
Provider Format → [OpenAI Hub] → Client Format (response)
|
||||
All format translation passes through **OpenAI format as the hub**:
|
||||
|
||||
```
|
||||
Client Format → [OpenAI Hub] → Provider Format (request)
|
||||
Provider Format → [OpenAI Hub] → Client Format (response)
|
||||
```
|
||||
|
||||
Cela signifie que vous n'avez besoin que de**N traducteurs**(un par format) au lieu de**N²**(chaque paire).---
|
||||
This means you only need **N translators** (one per format) instead of **N²** (every pair).
|
||||
|
||||
---
|
||||
|
||||
## 3. Project Structure
|
||||
|
||||
```
|
||||
|
||||
omniroute/
|
||||
├── open-sse/ ← Core proxy library (portable, framework-agnostic)
|
||||
│ ├── index.js ← Main entry point, exports everything
|
||||
│ ├── config/ ← Configuration & constants
|
||||
│ ├── executors/ ← Provider-specific request execution
|
||||
│ ├── handlers/ ← Request handling orchestration
|
||||
│ ├── services/ ← Business logic (auth, models, fallback, usage)
|
||||
│ ├── translator/ ← Format translation engine
|
||||
│ │ ├── request/ ← Request translators (8 files)
|
||||
│ │ ├── response/ ← Response translators (7 files)
|
||||
│ │ └── helpers/ ← Shared translation utilities (6 files)
|
||||
│ └── utils/ ← Utility functions
|
||||
├── src/ ← Application layer (Express/Worker runtime)
|
||||
│ ├── app/ ← Web UI, API routes, middleware
|
||||
│ ├── lib/ ← Database, auth, and shared library code
|
||||
│ ├── mitm/ ← Man-in-the-middle proxy utilities
|
||||
│ ├── models/ ← Database models
|
||||
│ ├── shared/ ← Shared utilities (wrappers around open-sse)
|
||||
│ ├── sse/ ← SSE endpoint handlers
|
||||
│ └── store/ ← State management
|
||||
├── data/ ← Runtime data (credentials, logs)
|
||||
│ └── provider-credentials.json (external credentials override, gitignored)
|
||||
└── tester/ ← Test utilities
|
||||
|
||||
````
|
||||
├── open-sse/ ← Core proxy library (portable, framework-agnostic)
|
||||
│ ├── index.js ← Main entry point, exports everything
|
||||
│ ├── config/ ← Configuration & constants
|
||||
│ ├── executors/ ← Provider-specific request execution
|
||||
│ ├── handlers/ ← Request handling orchestration
|
||||
│ ├── services/ ← Business logic (auth, models, fallback, usage)
|
||||
│ ├── translator/ ← Format translation engine
|
||||
│ │ ├── request/ ← Request translators (8 files)
|
||||
│ │ ├── response/ ← Response translators (7 files)
|
||||
│ │ └── helpers/ ← Shared translation utilities (6 files)
|
||||
│ └── utils/ ← Utility functions
|
||||
├── src/ ← Application layer (Express/Worker runtime)
|
||||
│ ├── app/ ← Web UI, API routes, middleware
|
||||
│ ├── lib/ ← Database, auth, and shared library code
|
||||
│ ├── mitm/ ← Man-in-the-middle proxy utilities
|
||||
│ ├── models/ ← Database models
|
||||
│ ├── shared/ ← Shared utilities (wrappers around open-sse)
|
||||
│ ├── sse/ ← SSE endpoint handlers
|
||||
│ └── store/ ← State management
|
||||
├── data/ ← Runtime data (credentials, logs)
|
||||
│ └── provider-credentials.json (external credentials override, gitignored)
|
||||
└── tester/ ← Test utilities
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
|
|
@ -105,16 +110,18 @@ omniroute/
|
|||
|
||||
### 4.1 Config (`open-sse/config/`)
|
||||
|
||||
La**source unique de vérité**pour toutes les configurations de fournisseurs.
|
||||
The **single source of truth** for all provider configuration.
|
||||
|
||||
| Fichier | Objectif |
|
||||
| File | Purpose |
|
||||
| ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| `constantes.ts` | Objet `PROVIDERS` avec les URL de base, les informations d'identification OAuth (par défaut), les en-têtes et les invites système par défaut pour chaque fournisseur. Définit également `HTTP_STATUS`, `ERROR_TYPES`, `COOLDOWN_MS`, `BACKOFF_CONFIG` et `SKIP_PATTERNS`. |
|
||||
| `credentialLoader.ts` | Charge les informations d'identification externes à partir de « data/provider-credentials.json » et les fusionne avec les valeurs par défaut codées en dur dans « PROVIDERS ». Garde les secrets hors du contrôle des sources tout en conservant la compatibilité ascendante. |
|
||||
| `providerModels.ts` | Registre central des modèles : mappe les alias des fournisseurs → les ID de modèle. Des fonctions comme `getModels()`, `getProviderByAlias()`. |
|
||||
| `codexInstructions.ts` | Instructions système injectées dans les requêtes Codex (contraintes d'édition, règles sandbox, politiques d'approbation). |
|
||||
| `defaultThinkingSignature.ts` | Signatures « pensées » par défaut pour les modèles Claude et Gemini. |
|
||||
| `ollamaModels.ts` | Définition de schéma pour les modèles Ollama locaux (nom, taille, famille, quantification). |#### Credential Loading Flow
|
||||
| `constants.ts` | `PROVIDERS` object with base URLs, OAuth credentials (defaults), headers, and default system prompts for every provider. Also defines `HTTP_STATUS`, `ERROR_TYPES`, `COOLDOWN_MS`, `BACKOFF_CONFIG`, and `SKIP_PATTERNS`. |
|
||||
| `credentialLoader.ts` | Loads external credentials from `data/provider-credentials.json` and merges them over the hardcoded defaults in `PROVIDERS`. Keeps secrets out of source control while maintaining backwards compatibility. |
|
||||
| `providerModels.ts` | Central model registry: maps provider aliases → model IDs. Functions like `getModels()`, `getProviderByAlias()`. |
|
||||
| `codexInstructions.ts` | System instructions injected into Codex requests (editing constraints, sandbox rules, approval policies). |
|
||||
| `defaultThinkingSignature.ts` | Default "thinking" signatures for Claude and Gemini models. |
|
||||
| `ollamaModels.ts` | Schema definition for local Ollama models (name, size, family, quantization). |
|
||||
|
||||
#### Credential Loading Flow
|
||||
|
||||
```mermaid
|
||||
flowchart TD
|
||||
|
|
@ -133,22 +140,24 @@ flowchart TD
|
|||
J --> F
|
||||
F -->|Done| L["PROVIDERS ready with\nmerged credentials"]
|
||||
E --> L
|
||||
````
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 4.2 Executors (`open-sse/executors/`)
|
||||
|
||||
Les exécuteurs encapsulent la**logique spécifique au fournisseur**à l'aide du**Modèle de stratégie**. Chaque exécuteur remplace les méthodes de base selon les besoins.```mermaid
|
||||
Executors encapsulate **provider-specific logic** using the **Strategy Pattern**. Each executor overrides base methods as needed.
|
||||
|
||||
```mermaid
|
||||
classDiagram
|
||||
class BaseExecutor {
|
||||
+buildUrl(model, stream, options)
|
||||
+buildHeaders(credentials, stream, body)
|
||||
+transformRequest(body, model, stream, credentials)
|
||||
+execute(url, options)
|
||||
+shouldRetry(status, error)
|
||||
+refreshCredentials(credentials, log)
|
||||
}
|
||||
class BaseExecutor {
|
||||
+buildUrl(model, stream, options)
|
||||
+buildHeaders(credentials, stream, body)
|
||||
+transformRequest(body, model, stream, credentials)
|
||||
+execute(url, options)
|
||||
+shouldRetry(status, error)
|
||||
+refreshCredentials(credentials, log)
|
||||
}
|
||||
|
||||
class DefaultExecutor {
|
||||
+refreshCredentials()
|
||||
|
|
@ -185,31 +194,34 @@ class BaseExecutor {
|
|||
BaseExecutor <|-- CodexExecutor
|
||||
BaseExecutor <|-- GeminiCLIExecutor
|
||||
BaseExecutor <|-- GithubExecutor
|
||||
```
|
||||
|
||||
````
|
||||
|
||||
| Exécuteur testamentaire | Fournisseur | Spécialisations clés |
|
||||
| Executor | Provider | Key Specializations |
|
||||
| ---------------- | ------------------------------------------ | ------------------------------------------------------------------------------------------------------------------- |
|
||||
| `base.ts` | — | Base abstraite : création d'URL, en-têtes, logique de nouvelle tentative, actualisation des informations d'identification |
|
||||
| `par défaut.ts` | Claude, Gémeaux, OpenAI, GLM, Kimi, MiniMax | Actualisation du jeton OAuth générique pour les fournisseurs standards |
|
||||
| `antigravité.ts` | Code Google Cloud | Génération d'ID de projet/session, secours multi-URL, nouvelle tentative d'analyse personnalisée à partir des messages d'erreur ("réinitialisation après 2h7m23s") |
|
||||
| `curseur.ts` | Curseur IDE |**Le plus complexe** : authentification par somme de contrôle SHA-256, encodage de requête Protobuf, EventStream binaire → analyse de réponse SSE |
|
||||
| `codex.ts` | Codex OpenAI | Injecte les instructions système, gère les niveaux de réflexion, supprime les paramètres non pris en charge |
|
||||
| `gemini-cli.ts` | CLI Google Gemini | Création d'URL personnalisées (`streamGenerateContent`), actualisation du jeton Google OAuth |
|
||||
| `github.ts` | Copilote GitHub | Système à double jeton (GitHub OAuth + jeton Copilot), imitation d'en-tête VSCode |
|
||||
| `kiro.ts` | AWS CodeWhisperer | Analyse binaire AWS EventStream, cadres d'événements AMZN, estimation de jetons |
|
||||
| `index.ts` | — | Factory : nom du fournisseur de cartes → classe d'exécuteur, avec solution de secours par défaut |---
|
||||
| `base.ts` | — | Abstract base: URL building, headers, retry logic, credential refresh |
|
||||
| `default.ts` | Claude, Gemini, OpenAI, GLM, Kimi, MiniMax | Generic OAuth token refresh for standard providers |
|
||||
| `antigravity.ts` | Google Cloud Code | Project/session ID generation, multi-URL fallback, custom retry parsing from error messages ("reset after 2h7m23s") |
|
||||
| `cursor.ts` | Cursor IDE | **Most complex**: SHA-256 checksum auth, Protobuf request encoding, binary EventStream → SSE response parsing |
|
||||
| `codex.ts` | OpenAI Codex | Injects system instructions, manages thinking levels, removes unsupported parameters |
|
||||
| `gemini-cli.ts` | Google Gemini CLI | Custom URL building (`streamGenerateContent`), Google OAuth token refresh |
|
||||
| `github.ts` | GitHub Copilot | Dual token system (GitHub OAuth + Copilot token), VSCode header mimicking |
|
||||
| `kiro.ts` | AWS CodeWhisperer | AWS EventStream binary parsing, AMZN event frames, token estimation |
|
||||
| `index.ts` | — | Factory: maps provider name → executor class, with default fallback |
|
||||
|
||||
---
|
||||
|
||||
### 4.3 Handlers (`open-sse/handlers/`)
|
||||
|
||||
La**couche d'orchestration** : coordonne la traduction, l'exécution, le streaming et la gestion des erreurs.
|
||||
The **orchestration layer** — coordinates translation, execution, streaming, and error handling.
|
||||
|
||||
| Fichier | Objectif |
|
||||
| ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
|
||||
| `chatCore.ts` |**Orchestrateur central**(~600 lignes). Gère le cycle de vie complet de la demande : détection du format → traduction → répartition de l'exécuteur → réponse en streaming/non-streaming → actualisation du jeton → gestion des erreurs → journalisation de l'utilisation. |
|
||||
| `responsesHandler.ts` | Adaptateur pour l'API Responses d'OpenAI : convertit le format des réponses → Fins de discussion → envoie à `chatCore` → reconvertit SSE au format de réponses. |
|
||||
| `embeddings.ts` | Gestionnaire de génération d'intégration : résout le modèle d'intégration → fournisseur, envoi à l'API du fournisseur, renvoie la réponse d'intégration compatible OpenAI. Prend en charge plus de 6 fournisseurs. |
|
||||
| `imageGeneration.ts` | Gestionnaire de génération d'images : résout le modèle d'image → fournisseur, prend en charge les modes compatibles OpenAI, Gemini-image (Antigravity) et de secours (Nebius). Renvoie des images base64 ou URL. |#### Request Lifecycle (chatCore.ts)
|
||||
| File | Purpose |
|
||||
| --------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| `chatCore.ts` | **Central orchestrator** (~600 lines). Handles the complete request lifecycle: format detection → translation → executor dispatch → streaming/non-streaming response → token refresh → error handling → usage logging. |
|
||||
| `responsesHandler.ts` | Adapter for OpenAI's Responses API: converts Responses format → Chat Completions → sends to `chatCore` → converts SSE back to Responses format. |
|
||||
| `embeddings.ts` | Embedding generation handler: resolves embedding model → provider, dispatches to provider API, returns OpenAI-compatible embedding response. Supports 6+ providers. |
|
||||
| `imageGeneration.ts` | Image generation handler: resolves image model → provider, supports OpenAI-compatible, Gemini-image (Antigravity), and fallback (Nebius) modes. Returns base64 or URL images. |
|
||||
|
||||
#### Request Lifecycle (chatCore.ts)
|
||||
|
||||
```mermaid
|
||||
sequenceDiagram
|
||||
|
|
@ -244,28 +256,30 @@ sequenceDiagram
|
|||
chatCore->>Executor: Retry with credential refresh
|
||||
chatCore->>chatCore: Account fallback logic
|
||||
end
|
||||
````
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 4.4 Services (`open-sse/services/`)
|
||||
|
||||
| Logique métier qui prend en charge les gestionnaires et les exécuteurs. | File | Purpose |
|
||||
| ----------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------- |
|
||||
| `provider.ts` | **Format detection** (`detectFormat`): analyzes request body structure to identify Claude/OpenAI/Gemini/Antigravity/Responses formats (includes `max_tokens` heuristic for Claude). Also: URL building, header building, thinking config normalization. Supports `openai-compatible-*` and `anthropic-compatible-*` dynamic providers. |
|
||||
| `model.ts` | Model string parsing (`claude/model-name` → `{provider: "claude", model: "model-name"}`), alias resolution with collision detection, input sanitization (rejects path traversal/control chars), and model info resolution with async alias getter support. |
|
||||
| `accountFallback.ts` | Rate-limit handling: exponential backoff (1s → 2s → 4s → max 2min), account cooldown management, error classification (which errors trigger fallback vs. not). |
|
||||
| `tokenRefresh.ts` | OAuth token refresh for **every provider**: Google (Gemini, Antigravity), Claude, Codex, Qwen, Qoder, GitHub (OAuth + Copilot dual-token), Kiro (AWS SSO OIDC + Social Auth). Includes in-flight promise deduplication cache and retry with exponential backoff. |
|
||||
| `combo.ts` | **Combo models**: chains of fallback models. If model A fails with a fallback-eligible error, try model B, then C, etc. Returns actual upstream status codes. |
|
||||
| `usage.ts` | Fetches quota/usage data from provider APIs (GitHub Copilot quotas, Antigravity model quotas, Codex rate limits, Kiro usage breakdowns, Claude settings). |
|
||||
| `accountSelector.ts` | Smart account selection with scoring algorithm: considers priority, health status, round-robin position, and cooldown state to pick the optimal account for each request. |
|
||||
| `contextManager.ts` | Request context lifecycle management: creates and tracks per-request context objects with metadata (request ID, timestamps, provider info) for debugging and logging. |
|
||||
| `ipFilter.ts` | IP-based access control: supports allowlist and blocklist modes. Validates client IP against configured rules before processing API requests. |
|
||||
| `sessionManager.ts` | Session tracking with client fingerprinting: tracks active sessions using hashed client identifiers, monitors request counts, and provides session metrics. |
|
||||
| `signatureCache.ts` | Request signature-based deduplication cache: prevents duplicate requests by caching recent request signatures and returning cached responses for identical requests within a time window. |
|
||||
| `systemPrompt.ts` | Global system prompt injection: prepends or appends a configurable system prompt to all requests, with per-provider compatibility handling. |
|
||||
| `thinkingBudget.ts` | Reasoning token budget management: supports passthrough, auto (strip thinking config), custom (fixed budget), and adaptive (complexity-scaled) modes for controlling thinking/reasoning tokens. |
|
||||
| `wildcardRouter.ts` | Wildcard model pattern routing: resolves wildcard patterns (e.g., `*/claude-*`) to concrete provider/model pairs based on availability and priority. |
|
||||
Business logic that supports the handlers and executors.
|
||||
|
||||
| File | Purpose |
|
||||
| -------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| `provider.ts` | **Format detection** (`detectFormat`): analyzes request body structure to identify Claude/OpenAI/Gemini/Antigravity/Responses formats (includes `max_tokens` heuristic for Claude). Also: URL building, header building, thinking config normalization. Supports `openai-compatible-*` and `anthropic-compatible-*` dynamic providers. |
|
||||
| `model.ts` | Model string parsing (`claude/model-name` → `{provider: "claude", model: "model-name"}`), alias resolution with collision detection, input sanitization (rejects path traversal/control chars), and model info resolution with async alias getter support. |
|
||||
| `accountFallback.ts` | Rate-limit handling: exponential backoff (1s → 2s → 4s → max 2min), account cooldown management, error classification (which errors trigger fallback vs. not). |
|
||||
| `tokenRefresh.ts` | OAuth token refresh for **every provider**: Google (Gemini, Antigravity), Claude, Codex, Qwen, Qoder, GitHub (OAuth + Copilot dual-token), Kiro (AWS SSO OIDC + Social Auth). Includes in-flight promise deduplication cache and retry with exponential backoff. |
|
||||
| `combo.ts` | **Combo models**: chains of fallback models. If model A fails with a fallback-eligible error, try model B, then C, etc. Returns actual upstream status codes. |
|
||||
| `usage.ts` | Fetches quota/usage data from provider APIs (GitHub Copilot quotas, Antigravity model quotas, Codex rate limits, Kiro usage breakdowns, Claude settings). |
|
||||
| `accountSelector.ts` | Smart account selection with scoring algorithm: considers priority, health status, round-robin position, and cooldown state to pick the optimal account for each request. |
|
||||
| `contextManager.ts` | Request context lifecycle management: creates and tracks per-request context objects with metadata (request ID, timestamps, provider info) for debugging and logging. |
|
||||
| `ipFilter.ts` | IP-based access control: supports allowlist and blocklist modes. Validates client IP against configured rules before processing API requests. |
|
||||
| `sessionManager.ts` | Session tracking with client fingerprinting: tracks active sessions using hashed client identifiers, monitors request counts, and provides session metrics. |
|
||||
| `signatureCache.ts` | Request signature-based deduplication cache: prevents duplicate requests by caching recent request signatures and returning cached responses for identical requests within a time window. |
|
||||
| `systemPrompt.ts` | Global system prompt injection: prepends or appends a configurable system prompt to all requests, with per-provider compatibility handling. |
|
||||
| `thinkingBudget.ts` | Reasoning token budget management: supports passthrough, auto (strip thinking config), custom (fixed budget), and adaptive (complexity-scaled) modes for controlling thinking/reasoning tokens. |
|
||||
| `wildcardRouter.ts` | Wildcard model pattern routing: resolves wildcard patterns (e.g., `*/claude-*`) to concrete provider/model pairs based on availability and priority. |
|
||||
|
||||
#### Token Refresh Deduplication
|
||||
|
||||
|
|
@ -334,7 +348,9 @@ flowchart LR
|
|||
|
||||
### 4.5 Translator (`open-sse/translator/`)
|
||||
|
||||
Le**moteur de traduction de format**utilisant un système de plugin d'auto-enregistrement.#### Architecture
|
||||
The **format translation engine** using a self-registering plugin system.
|
||||
|
||||
#### Architecture
|
||||
|
||||
```mermaid
|
||||
graph TD
|
||||
|
|
@ -360,13 +376,15 @@ graph TD
|
|||
end
|
||||
```
|
||||
|
||||
| Annuaire | Fichiers | Descriptif |
|
||||
| ------------ | ------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------- |
|
||||
| `demande/` | 8 traducteurs | Convertissez les corps de requête entre les formats. Chaque fichier s'auto-enregistre via `register(from, to, fn)` lors de l'importation. |
|
||||
| `réponse/` | 7 traducteurs | Convertissez les morceaux de réponse en streaming entre les formats. Gère les types d’événements SSE, les blocs de réflexion et les appels d’outils. |
|
||||
| `helpers/` | 6 aides | Utilitaires partagés : `claudeHelper` (extraction d'invite système, configuration de réflexion), `geminiHelper` (mapping parties/contenu), `openaiHelper` (filtrage de format), `toolCallHelper` (génération d'ID, injection de réponse manquante), `maxTokensHelper`, `responsesApiHelper`. |
|
||||
| `index.ts` | — | Moteur de traduction : `translateRequest()`, `translateResponse()`, gestion des états, registre. |
|
||||
| `formats.ts` | — | Constantes de format : `OPENAI`, `CLAUDE`, `GEMINI`, `ANTIGRAVITY`, `KIRO`, `CURSOR`, `OPENAI_RESPONSES`. | #### Key Design: Self-Registering Plugins |
|
||||
| Directory | Files | Description |
|
||||
| ------------ | ------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| `request/` | 8 translators | Convert request bodies between formats. Each file self-registers via `register(from, to, fn)` on import. |
|
||||
| `response/` | 7 translators | Convert streaming response chunks between formats. Handles SSE event types, thinking blocks, tool calls. |
|
||||
| `helpers/` | 6 helpers | Shared utilities: `claudeHelper` (system prompt extraction, thinking config), `geminiHelper` (parts/contents mapping), `openaiHelper` (format filtering), `toolCallHelper` (ID generation, missing response injection), `maxTokensHelper`, `responsesApiHelper`. |
|
||||
| `index.ts` | — | Translation engine: `translateRequest()`, `translateResponse()`, state management, registry. |
|
||||
| `formats.ts` | — | Format constants: `OPENAI`, `CLAUDE`, `GEMINI`, `ANTIGRAVITY`, `KIRO`, `CURSOR`, `OPENAI_RESPONSES`. |
|
||||
|
||||
#### Key Design: Self-Registering Plugins
|
||||
|
||||
```javascript
|
||||
// Each translator file calls register() on import:
|
||||
|
|
@ -381,15 +399,17 @@ import "./request/claude-to-openai.js"; // ← self-registers
|
|||
|
||||
### 4.6 Utils (`open-sse/utils/`)
|
||||
|
||||
| Fichier | Objectif |
|
||||
| ------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------- |
|
||||
| `erreur.ts` | Création de réponses aux erreurs (format compatible OpenAI), analyse des erreurs en amont, extraction du temps de nouvelle tentative Antigravity à partir des messages d'erreur, streaming d'erreurs SSE. |
|
||||
| `stream.ts` | **SSE Transform Stream** : le pipeline de streaming principal. Deux modes : `TRANSLATE` (traduction plein format) et `PASSTHROUGH` (normaliser + extraire l'utilisation). Gère la mise en mémoire tampon des blocs, l'estimation de l'utilisation et le suivi de la longueur du contenu. Les instances d'encodeur/décodeur par flux évitent l'état partagé. |
|
||||
| `streamHelpers.ts` | Utilitaires SSE de bas niveau : `parseSSELine` (tolérant les espaces), `hasValuableContent` (filtre les morceaux vides pour OpenAI/Claude/Gemini), `fixInvalidId`, `formatSSE` (sérialisation SSE sensible au format avec nettoyage `perf_metrics`). |
|
||||
| `usageTracking.ts` | Extraction de l'utilisation des jetons à partir de n'importe quel format (Claude/OpenAI/Gemini/Responses), estimation avec des ratios outil/message séparés par jeton, ajout de tampon (marge de sécurité de 2000 jetons), filtrage de champs spécifiques au format, journalisation de la console avec couleurs ANSI. |
|
||||
| `requestLogger.ts` | Legacy file-based request logging helper kept for compatibility. Current deployments should prefer `APP_LOG_TO_FILE` for application logs and the call log pipeline for persisted request artifacts. |
|
||||
| `bypassHandler.ts` | Intercepte les modèles spécifiques de Claude CLI (extraction de titre, échauffement, décompte) et renvoie de fausses réponses sans appeler aucun fournisseur. Prend en charge le streaming et le non-streaming. Intentionnellement limité à la portée Claude CLI. |
|
||||
| `networkProxy.ts` | Résout l'URL du proxy sortant pour un fournisseur donné avec la priorité : configuration spécifique au fournisseur → configuration globale → variables d'environnement (`HTTPS_PROXY`/`HTTP_PROXY`/`ALL_PROXY`). Prend en charge les exclusions `NO_PROXY`. Met en cache la configuration pendant 30 s. | #### SSE Streaming Pipeline |
|
||||
| File | Purpose |
|
||||
| ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
|
||||
| `error.ts` | Error response building (OpenAI-compatible format), upstream error parsing, Antigravity retry-time extraction from error messages, SSE error streaming. |
|
||||
| `stream.ts` | **SSE Transform Stream** — the core streaming pipeline. Two modes: `TRANSLATE` (full format translation) and `PASSTHROUGH` (normalize + extract usage). Handles chunk buffering, usage estimation, content length tracking. Per-stream encoder/decoder instances avoid shared state. |
|
||||
| `streamHelpers.ts` | Low-level SSE utilities: `parseSSELine` (whitespace-tolerant), `hasValuableContent` (filters empty chunks for OpenAI/Claude/Gemini), `fixInvalidId`, `formatSSE` (format-aware SSE serialization with `perf_metrics` cleanup). |
|
||||
| `usageTracking.ts` | Token usage extraction from any format (Claude/OpenAI/Gemini/Responses), estimation with separate tool/message char-per-token ratios, buffer addition (2000 tokens safety margin), format-specific field filtering, console logging with ANSI colors. |
|
||||
| `requestLogger.ts` | Legacy file-based request logging helper kept for compatibility. Current deployments should prefer `APP_LOG_TO_FILE` for application logs and the call log pipeline for persisted request artifacts. |
|
||||
| `bypassHandler.ts` | Intercepts specific patterns from Claude CLI (title extraction, warmup, count) and returns fake responses without calling any provider. Supports both streaming and non-streaming. Intentionally limited to Claude CLI scope. |
|
||||
| `networkProxy.ts` | Resolves outbound proxy URL for a given provider with precedence: provider-specific config → global config → environment variables (`HTTPS_PROXY`/`HTTP_PROXY`/`ALL_PROXY`). Supports `NO_PROXY` exclusions. Caches config for 30s. |
|
||||
|
||||
#### SSE Streaming Pipeline
|
||||
|
||||
```mermaid
|
||||
flowchart TD
|
||||
|
|
@ -431,81 +451,103 @@ logs/
|
|||
|
||||
### 4.7 Application Layer (`src/`)
|
||||
|
||||
| Annuaire | Objectif |
|
||||
| -------------- | ----------------------------------------------------------------------------------------------------- | ----------------------- |
|
||||
| `src/app/` | Interface utilisateur Web, routes API, middleware express, gestionnaires de rappel OAuth |
|
||||
| `src/lib/` | Accès à la base de données (`localDb.ts`, `usageDb.ts`), authentification, partagé |
|
||||
| `src/mitm/` | Utilitaires proxy Man-in-the-middle pour intercepter le trafic des fournisseurs |
|
||||
| `src/modèles/` | Définitions du modèle de base de données |
|
||||
| `src/partagé/` | Wrappers autour des fonctions open-sse (fournisseur, flux, erreur, etc.) |
|
||||
| `src/sse/` | Gestionnaires de points de terminaison SSE qui connectent la bibliothèque open-sse aux routes Express |
|
||||
| `src/magasin/` | Gestion de l'état des applications | #### Notable API Routes |
|
||||
| Directory | Purpose |
|
||||
| ------------- | ---------------------------------------------------------------------- |
|
||||
| `src/app/` | Web UI, API routes, Express middleware, OAuth callback handlers |
|
||||
| `src/lib/` | Database access (`localDb.ts`, `usageDb.ts`), authentication, shared |
|
||||
| `src/mitm/` | Man-in-the-middle proxy utilities for intercepting provider traffic |
|
||||
| `src/models/` | Database model definitions |
|
||||
| `src/shared/` | Wrappers around open-sse functions (provider, stream, error, etc.) |
|
||||
| `src/sse/` | SSE endpoint handlers that wire the open-sse library to Express routes |
|
||||
| `src/store/` | Application state management |
|
||||
|
||||
| Itinéraire | Méthodes | Objectif |
|
||||
| --------------------------------------------- | ------------------------ | ------------------------------------------------------------------------------------------------------- | --- |
|
||||
| `/api/provider-models` | OBTENIR/POST/DELETE | CRUD pour les modèles personnalisés par fournisseur |
|
||||
| `/api/models/catalogue` | OBTENIR | Catalogue agrégé de tous les modèles (chat, intégration, image, personnalisé) regroupés par fournisseur |
|
||||
| `/api/settings/proxy` | OBTENIR/METTRE/SUPPRIMER | Configuration du proxy sortant hiérarchique (`global/providers/combos/keys`) |
|
||||
| `/api/settings/proxy/test` | POSTER | Valide la connectivité proxy et renvoie l'adresse IP/latence publique |
|
||||
| `/v1/providers/[provider]/chat/completions` | POSTER | Compléments de chat dédiés par fournisseur avec validation du modèle |
|
||||
| `/v1/providers/[provider]/embeddings` | POSTER | Intégrations dédiées par fournisseur avec validation du modèle |
|
||||
| `/v1/providers/[provider]/images/générations` | POSTER | Génération d'images dédiée par fournisseur avec validation du modèle |
|
||||
| `/api/settings/ip-filter` | OBTENIR/METTRE | Gestion des listes autorisées/bloquées IP |
|
||||
| `/api/settings/thinking-budget` | OBTENIR/METTRE | Configuration du budget du jeton de raisonnement (passthrough/auto/custom/adaptatif) |
|
||||
| `/api/settings/system-prompt` | OBTENIR/METTRE | Injection rapide du système global pour toutes les demandes |
|
||||
| `/api/sessions` | OBTENIR | Suivi et métriques des sessions actives |
|
||||
| `/api/rate-limites` | OBTENIR | Statut de limite de débit par compte | --- |
|
||||
#### Notable API Routes
|
||||
|
||||
| Route | Methods | Purpose |
|
||||
| --------------------------------------------- | --------------- | ------------------------------------------------------------------------------------- |
|
||||
| `/api/provider-models` | GET/POST/DELETE | CRUD for custom models per provider |
|
||||
| `/api/models/catalog` | GET | Aggregated catalog of all models (chat, embedding, image, custom) grouped by provider |
|
||||
| `/api/settings/proxy` | GET/PUT/DELETE | Hierarchical outbound proxy configuration (`global/providers/combos/keys`) |
|
||||
| `/api/settings/proxy/test` | POST | Validates proxy connectivity and returns public IP/latency |
|
||||
| `/v1/providers/[provider]/chat/completions` | POST | Dedicated per-provider chat completions with model validation |
|
||||
| `/v1/providers/[provider]/embeddings` | POST | Dedicated per-provider embeddings with model validation |
|
||||
| `/v1/providers/[provider]/images/generations` | POST | Dedicated per-provider image generation with model validation |
|
||||
| `/api/settings/ip-filter` | GET/PUT | IP allowlist/blocklist management |
|
||||
| `/api/settings/thinking-budget` | GET/PUT | Reasoning token budget configuration (passthrough/auto/custom/adaptive) |
|
||||
| `/api/settings/system-prompt` | GET/PUT | Global system prompt injection for all requests |
|
||||
| `/api/sessions` | GET | Active session tracking and metrics |
|
||||
| `/api/rate-limits` | GET | Per-account rate limit status |
|
||||
|
||||
---
|
||||
|
||||
## 5. Key Design Patterns
|
||||
|
||||
### 5.1 Hub-and-Spoke Translation
|
||||
|
||||
Tous les formats sont traduits via le**format OpenAI comme hub**. L'ajout d'un nouveau fournisseur ne nécessite que l'écriture d'**une paire**de traducteurs (vers/depuis OpenAI), et non de N paires.### 5.2 Executor Strategy Pattern
|
||||
All formats translate through **OpenAI format as the hub**. Adding a new provider only requires writing **one pair** of translators (to/from OpenAI), not N pairs.
|
||||
|
||||
Chaque fournisseur dispose d'une classe d'exécuteur dédiée héritant de « BaseExecutor ». L'usine dans `executors/index.ts` sélectionne la bonne au moment de l'exécution.### 5.3 Self-Registering Plugin System
|
||||
### 5.2 Executor Strategy Pattern
|
||||
|
||||
Les modules de traduction s'enregistrent eux-mêmes lors de l'importation via `register()`. Ajouter un nouveau traducteur consiste simplement à créer un fichier et à l'importer.### 5.4 Account Fallback with Exponential Backoff
|
||||
Each provider has a dedicated executor class inheriting from `BaseExecutor`. The factory in `executors/index.ts` selects the right one at runtime.
|
||||
|
||||
Lorsqu'un fournisseur renvoie 429/401/500, le système peut passer au compte suivant, en appliquant des temps de recharge exponentiels (1s → 2s → 4s → max 2min).### 5.5 Combo Model Chains
|
||||
### 5.3 Self-Registering Plugin System
|
||||
|
||||
Un « combo » regroupe plusieurs chaînes « fournisseur/modèle ». Si le premier échoue, revenez automatiquement au suivant.### 5.6 Stateful Streaming Translation
|
||||
Translator modules register themselves on import via `register()`. Adding a new translator is just creating a file and importing it.
|
||||
|
||||
La traduction des réponses maintient l'état dans les morceaux SSE (suivi des blocs de réflexion, accumulation d'appels d'outils, indexation des blocs de contenu) via le mécanisme `initState()`.### 5.7 Usage Safety Buffer
|
||||
### 5.4 Account Fallback with Exponential Backoff
|
||||
|
||||
Un tampon de 2 000 jetons est ajouté à l'utilisation signalée pour empêcher les clients d'atteindre les limites de la fenêtre contextuelle en raison de la surcharge des invites système et de la traduction du format.---
|
||||
When a provider returns 429/401/500, the system can switch to the next account, applying exponential cooldowns (1s → 2s → 4s → max 2min).
|
||||
|
||||
### 5.5 Combo Model Chains
|
||||
|
||||
A "combo" groups multiple `provider/model` strings. If the first fails, fallback to the next automatically.
|
||||
|
||||
### 5.6 Stateful Streaming Translation
|
||||
|
||||
Response translation maintains state across SSE chunks (thinking block tracking, tool call accumulation, content block indexing) via the `initState()` mechanism.
|
||||
|
||||
### 5.7 Usage Safety Buffer
|
||||
|
||||
A 2000-token buffer is added to reported usage to prevent clients from hitting context window limits due to overhead from system prompts and format translation.
|
||||
|
||||
---
|
||||
|
||||
## 6. Supported Formats
|
||||
|
||||
| Formater | Itinéraire | Identifiant |
|
||||
| -------------------------- | ---------------- | ----------------- | --- |
|
||||
| Achèvements du chat OpenAI | source + cible | `openai` |
|
||||
| API de réponses OpenAI | source + cible | `openai-réponses` |
|
||||
| Claude Anthropique | source + cible | `claude` |
|
||||
| Google Gémeaux | source + cible | `Gémeaux` |
|
||||
| CLI Google Gemini | cible uniquement | `gemini-cli` |
|
||||
| Antigravité | source + cible | `antigravité` |
|
||||
| AWSKiro | cible uniquement | `kiro` |
|
||||
| Curseur | cible uniquement | `curseur` | --- |
|
||||
| Format | Direction | Identifier |
|
||||
| ----------------------- | --------------- | ------------------ |
|
||||
| OpenAI Chat Completions | source + target | `openai` |
|
||||
| OpenAI Responses API | source + target | `openai-responses` |
|
||||
| Anthropic Claude | source + target | `claude` |
|
||||
| Google Gemini | source + target | `gemini` |
|
||||
| Google Gemini CLI | target only | `gemini-cli` |
|
||||
| Antigravity | source + target | `antigravity` |
|
||||
| AWS Kiro | target only | `kiro` |
|
||||
| Cursor | target only | `cursor` |
|
||||
|
||||
---
|
||||
|
||||
## 7. Supported Providers
|
||||
|
||||
| Fournisseur | Méthode d'authentification | Exécuteur testamentaire | Notes clés |
|
||||
| -------------------------- | ---------------------------------------- | ----------------------- | ------------------------------------------------------------------------- | --- |
|
||||
| Claude Anthropique | Clé API ou OAuth | Par défaut | Utilise l'en-tête `x-api-key` |
|
||||
| Google Gémeaux | Clé API ou OAuth | Par défaut | Utilise l'en-tête `x-goog-api-key` |
|
||||
| CLI Google Gemini | OAuth | GémeauxCLI | Utilise le point de terminaison `streamGenerateContent` |
|
||||
| Antigravité | OAuth | Antigravité | Solution de secours multi-URL, nouvelle tentative d'analyse personnalisée |
|
||||
| OpenAI | Clé API | Par défaut | Authentification du porte-étendard |
|
||||
| Codex | OAuth | Codex | Injecte les instructions système, gère la réflexion |
|
||||
| Copilote GitHub | OAuth + jeton Copilot | GitHub | Double jeton, en-tête VSCode imitant |
|
||||
| Kiro (AWS) | AWS SSO OIDC ou Social | Kiro | Analyse binaire d'EventStream |
|
||||
| Curseur IDE | Authentification de la somme de contrôle | Curseur | Encodage Protobuf, sommes de contrôle SHA-256 |
|
||||
| Qwen | OAuth | Par défaut | Authentification standard |
|
||||
| Qoder | OAuth (Basique + Porteur) | Par défaut | En-tête à double authentification |
|
||||
| OuvrirRouter | Clé API | Par défaut | Authentification du porte-étendard |
|
||||
| GLM, Kimi, MiniMax | Clé API | Par défaut | Compatible avec Claude, utilisez `x-api-key` |
|
||||
| `openai-compatible-*` | Clé API | Par défaut | Dynamique : tout point de terminaison compatible OpenAI |
|
||||
| `anthropique-compatible-*` | Clé API | Par défaut | Dynamique : tout point de terminaison compatible Claude | --- |
|
||||
| Provider | Auth Method | Executor | Key Notes |
|
||||
| ------------------------ | ---------------------- | ----------- | --------------------------------------------- |
|
||||
| Anthropic Claude | API key or OAuth | Default | Uses `x-api-key` header |
|
||||
| Google Gemini | API key or OAuth | Default | Uses `x-goog-api-key` header |
|
||||
| Google Gemini CLI | OAuth | GeminiCLI | Uses `streamGenerateContent` endpoint |
|
||||
| Antigravity | OAuth | Antigravity | Multi-URL fallback, custom retry parsing |
|
||||
| OpenAI | API key | Default | Standard Bearer auth |
|
||||
| Codex | OAuth | Codex | Injects system instructions, manages thinking |
|
||||
| GitHub Copilot | OAuth + Copilot token | Github | Dual token, VSCode header mimicking |
|
||||
| Kiro (AWS) | AWS SSO OIDC or Social | Kiro | Binary EventStream parsing |
|
||||
| Cursor IDE | Checksum auth | Cursor | Protobuf encoding, SHA-256 checksums |
|
||||
| Qwen | OAuth | Default | Standard auth |
|
||||
| Qoder | OAuth (Basic + Bearer) | Default | Dual auth header |
|
||||
| OpenRouter | API key | Default | Standard Bearer auth |
|
||||
| GLM, Kimi, MiniMax | API key | Default | Claude-compatible, use `x-api-key` |
|
||||
| `openai-compatible-*` | API key | Default | Dynamic: any OpenAI-compatible endpoint |
|
||||
| `anthropic-compatible-*` | API key | Default | Dynamic: any Claude-compatible endpoint |
|
||||
|
||||
---
|
||||
|
||||
## 8. Data Flow Summary
|
||||
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue