router-maestro 0.1.3__tar.gz → 0.1.5__tar.gz

router_maestro-0.1.5/.markdownlint.json

@@ -0,0 +1,4 @@
+{
+  "MD013": false,
+  "MD060": false
+}

{router_maestro-0.1.3 → router_maestro-0.1.5}/CLAUDE.md

@@ -57,7 +57,7 @@ Router-Maestro is a multi-model routing system that exposes both OpenAI-compatib
 - `schemas/` - Pydantic models for both API formats
 
 **CLI (`src/router_maestro/cli/`)**
-- Typer-based CLI with subcommands: `server`, `auth`, `model`, `context`, `config`, `stats`
+- Typer-based CLI with subcommands: `server`, `auth`, `model`, `context`, `config`
 - Each subcommand in its own module registered in `main.py`
 
 ### Data Flow
@@ -72,7 +72,7 @@ Router-Maestro is a multi-model routing system that exposes both OpenAI-compatib
 
 Configuration and data files follow XDG conventions:
 - **Config** (`~/.config/router-maestro/`): `providers.json`, `priorities.json`, `contexts.json`
-- **Data** (`~/.local/share/router-maestro/`): `auth.json`, `server.json`, `stats.db`
+- **Data** (`~/.local/share/router-maestro/`): `auth.json`, `server.json`
 
 ### Model Identification
 

{router_maestro-0.1.3 → router_maestro-0.1.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: router-maestro
-Version: 0.1.3
+Version: 0.1.5
 Summary: Multi-model routing and load balancing system with OpenAI-compatible API
 Author-email: Kanwen Li <likanwen@icloud.com>
 License-Expression: MIT
@@ -132,8 +132,8 @@ router-maestro config claude-code
 
 Models are identified using the format `{provider}/{model-id}`:
 
-| Example | Description |
-|---------|-------------|
+| Example                           | Description                         |
+| --------------------------------- | ----------------------------------- |
 | `github-copilot/gpt-4o` | GPT-4o via GitHub Copilot |
 | `github-copilot/claude-sonnet-4` | Claude Sonnet 4 via GitHub Copilot |
 | `openai/gpt-4-turbo` | GPT-4 Turbo via OpenAI |
@@ -164,8 +164,8 @@ router-maestro model priority list
 
 **Fallback** triggers when a request fails with a retryable error (429, 5xx):
 
-| Strategy | Behavior |
-|----------|----------|
+| Strategy     | Behavior                             |
+| ------------ | ------------------------------------ |
 | `priority` | Try next model in priorities list |
 | `same-model` | Try same model on different provider |
 | `none` | Fail immediately |
@@ -195,8 +195,8 @@ POST /v1/chat/completions  {"model": "anthropic/claude-3-5-sonnet", ...}
 
 A **context** is a named connection profile that stores an endpoint URL and API key. Contexts let you manage multiple Router-Maestro deployments from a single CLI.
 
-| Context | Use Case |
-|---------|----------|
+| Context  | Use Case                                   |
+| -------- | ------------------------------------------ |
 | `local` | Default context for `router-maestro server start` |
 | `docker` | Connect to a local Docker container |
 | `my-vps` | Connect to a remote VPS deployment |
@@ -216,25 +216,25 @@ router-maestro model list
 
 ### Server
 
-| Command | Description |
-|---------|-------------|
-| `server start --port 8080` | Start the server |
+| Command                    | Description        |
+| -------------------------- | ------------------ |
+| `server start --port 8080` | Start the server   |
 | `server stop` | Stop the server |
 | `server info` | Show server status |
 
 ### Authentication
 
-| Command | Description |
-|---------|-------------|
-| `auth login [provider]` | Authenticate with a provider |
+| Command                 | Description                    |
+| ----------------------- | ------------------------------ |
+| `auth login [provider]` | Authenticate with a provider   |
 | `auth logout <provider>` | Remove authentication |
 | `auth list` | List authenticated providers |
 
 ### Models
 
-| Command | Description |
-|---------|-------------|
-| `model list` | List available models |
+| Command                            | Description            |
+| ---------------------------------- | ---------------------- |
+| `model list`                       | List available models  |
 | `model refresh` | Refresh models cache |
 | `model priority list` | Show priorities |
 | `model priority <model> --position <n>` | Set priority |
@@ -242,9 +242,9 @@ router-maestro model list
 
 ### Contexts (Remote Management)
 
-| Command | Description |
-|---------|-------------|
-| `context show` | Show current context |
+| Command                                              | Description          |
+| ---------------------------------------------------- | -------------------- |
+| `context show`                                       | Show current context |
 | `context list` | List all contexts |
 | `context set <name>` | Switch context |
 | `context add <name> --endpoint <url> --api-key <key>` | Add remote context |
@@ -252,8 +252,8 @@ router-maestro model list
 
 ### Other
 
-| Command | Description |
-|---------|-------------|
+| Command              | Description                   |
+| -------------------- | ----------------------------- |
 | `config claude-code` | Generate Claude Code settings |
 
 ## API Reference
@@ -301,8 +301,8 @@ POST /api/admin/models/refresh   # Refresh model cache
 
 Following XDG Base Directory specification:
 
-| Type | Path | Contents |
-|------|------|----------|
+| Type       | Path                               | Contents                     |
+| ---------- | ---------------------------------- | ---------------------------- |
 | **Config** | `~/.config/router-maestro/` | |
 | | `providers.json` | Custom provider definitions |
 | | `priorities.json` | Model priorities and fallback |
@@ -340,8 +340,8 @@ export OLLAMA_API_KEY="sk-..."
 
 Configuration files are automatically reloaded every 5 minutes:
 
-| File | Auto-Reload |
-|------|-------------|
+| File               | Auto-Reload      |
+| ------------------ | ---------------- |
 | `priorities.json` | ✓ (5 min) |
 | `providers.json` | ✓ (5 min) |
 | `auth.json` | Requires restart |
@@ -405,6 +405,7 @@ router-maestro model list
 The Docker Compose setup includes Traefik for automatic HTTPS via Let's Encrypt with DNS challenge.
 
 For detailed configuration options including:
+
 - Other DNS providers (Route53, DigitalOcean, etc.)
 - HTTP challenge setup
 - Traefik dashboard configuration

{router_maestro-0.1.3 → router_maestro-0.1.5}/README.md

@@ -95,8 +95,8 @@ router-maestro config claude-code
 
 Models are identified using the format `{provider}/{model-id}`:
 
-| Example | Description |
-|---------|-------------|
+| Example                           | Description                         |
+| --------------------------------- | ----------------------------------- |
 | `github-copilot/gpt-4o` | GPT-4o via GitHub Copilot |
 | `github-copilot/claude-sonnet-4` | Claude Sonnet 4 via GitHub Copilot |
 | `openai/gpt-4-turbo` | GPT-4 Turbo via OpenAI |
@@ -127,8 +127,8 @@ router-maestro model priority list
 
 **Fallback** triggers when a request fails with a retryable error (429, 5xx):
 
-| Strategy | Behavior |
-|----------|----------|
+| Strategy     | Behavior                             |
+| ------------ | ------------------------------------ |
 | `priority` | Try next model in priorities list |
 | `same-model` | Try same model on different provider |
 | `none` | Fail immediately |
@@ -158,8 +158,8 @@ POST /v1/chat/completions  {"model": "anthropic/claude-3-5-sonnet", ...}
 
 A **context** is a named connection profile that stores an endpoint URL and API key. Contexts let you manage multiple Router-Maestro deployments from a single CLI.
 
-| Context | Use Case |
-|---------|----------|
+| Context  | Use Case                                   |
+| -------- | ------------------------------------------ |
 | `local` | Default context for `router-maestro server start` |
 | `docker` | Connect to a local Docker container |
 | `my-vps` | Connect to a remote VPS deployment |
@@ -179,25 +179,25 @@ router-maestro model list
 
 ### Server
 
-| Command | Description |
-|---------|-------------|
-| `server start --port 8080` | Start the server |
+| Command                    | Description        |
+| -------------------------- | ------------------ |
+| `server start --port 8080` | Start the server   |
 | `server stop` | Stop the server |
 | `server info` | Show server status |
 
 ### Authentication
 
-| Command | Description |
-|---------|-------------|
-| `auth login [provider]` | Authenticate with a provider |
+| Command                 | Description                    |
+| ----------------------- | ------------------------------ |
+| `auth login [provider]` | Authenticate with a provider   |
 | `auth logout <provider>` | Remove authentication |
 | `auth list` | List authenticated providers |
 
 ### Models
 
-| Command | Description |
-|---------|-------------|
-| `model list` | List available models |
+| Command                            | Description            |
+| ---------------------------------- | ---------------------- |
+| `model list`                       | List available models  |
 | `model refresh` | Refresh models cache |
 | `model priority list` | Show priorities |
 | `model priority <model> --position <n>` | Set priority |
@@ -205,9 +205,9 @@ router-maestro model list
 
 ### Contexts (Remote Management)
 
-| Command | Description |
-|---------|-------------|
-| `context show` | Show current context |
+| Command                                              | Description          |
+| ---------------------------------------------------- | -------------------- |
+| `context show`                                       | Show current context |
 | `context list` | List all contexts |
 | `context set <name>` | Switch context |
 | `context add <name> --endpoint <url> --api-key <key>` | Add remote context |
@@ -215,8 +215,8 @@ router-maestro model list
 
 ### Other
 
-| Command | Description |
-|---------|-------------|
+| Command              | Description                   |
+| -------------------- | ----------------------------- |
 | `config claude-code` | Generate Claude Code settings |
 
 ## API Reference
@@ -264,8 +264,8 @@ POST /api/admin/models/refresh   # Refresh model cache
 
 Following XDG Base Directory specification:
 
-| Type | Path | Contents |
-|------|------|----------|
+| Type       | Path                               | Contents                     |
+| ---------- | ---------------------------------- | ---------------------------- |
 | **Config** | `~/.config/router-maestro/` | |
 | | `providers.json` | Custom provider definitions |
 | | `priorities.json` | Model priorities and fallback |
@@ -303,8 +303,8 @@ export OLLAMA_API_KEY="sk-..."
 
 Configuration files are automatically reloaded every 5 minutes:
 
-| File | Auto-Reload |
-|------|-------------|
+| File               | Auto-Reload      |
+| ------------------ | ---------------- |
 | `priorities.json` | ✓ (5 min) |
 | `providers.json` | ✓ (5 min) |
 | `auth.json` | Requires restart |
@@ -368,6 +368,7 @@ router-maestro model list
 The Docker Compose setup includes Traefik for automatic HTTPS via Let's Encrypt with DNS challenge.
 
 For detailed configuration options including:
+
 - Other DNS providers (Route53, DigitalOcean, etc.)
 - HTTP challenge setup
 - Traefik dashboard configuration

{router_maestro-0.1.3 → router_maestro-0.1.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "router-maestro"
-version = "0.1.3"
+version = "0.1.5"
 description = "Multi-model routing and load balancing system with OpenAI-compatible API"
 readme = "README.md"
 license = "MIT"

{router_maestro-0.1.3 → router_maestro-0.1.5}/src/router_maestro/__init__.py

@@ -1,3 +1,3 @@
 """Router-Maestro: Multi-model routing and load balancing system."""
 
-__version__ = "0.1.3"
+__version__ = "0.1.5"

{router_maestro-0.1.3 → router_maestro-0.1.5}/src/router_maestro/cli/config.py

@@ -15,9 +15,17 @@ from rich.table import Table
 from router_maestro.cli.client import ServerNotRunningError, get_admin_client
 from router_maestro.config.server import get_current_context_api_key
 
-app = typer.Typer(no_args_is_help=True)
+app = typer.Typer(invoke_without_command=True)
 console = Console()
 
+# Available CLI tools for configuration
+CLI_TOOLS = {
+    "claude-code": {
+        "name": "Claude Code",
+        "description": "Generate settings.json for Claude Code CLI",
+    },
+}
+
 
 def get_claude_code_paths() -> dict[str, Path]:
     """Get Claude Code settings paths."""
@@ -27,6 +35,33 @@ def get_claude_code_paths() -> dict[str, Path]:
     }
 
 
+@app.callback(invoke_without_command=True)
+def config_callback(ctx: typer.Context) -> None:
+    """Generate configuration for CLI tools (interactive selection if not specified)."""
+    if ctx.invoked_subcommand is not None:
+        return
+
+    # Interactive selection
+    console.print("\n[bold]Available CLI tools:[/bold]")
+    tools = list(CLI_TOOLS.items())
+    for i, (key, info) in enumerate(tools, 1):
+        console.print(f"  {i}. {info['name']} - {info['description']}")
+
+    console.print()
+    choice = Prompt.ask(
+        "Select tool to configure",
+        choices=[str(i) for i in range(1, len(tools) + 1)],
+        default="1",
+    )
+
+    idx = int(choice) - 1
+    tool_key = tools[idx][0]
+
+    # Dispatch to the appropriate command
+    if tool_key == "claude-code":
+        claude_code_config()
+
+
 @app.command(name="claude-code")
 def claude_code_config() -> None:
     """Generate Claude Code CLI settings.json for router-maestro."""
@@ -104,19 +139,29 @@ def claude_code_config() -> None:
     )
     anthropic_url = f"{base_url}/api/anthropic"
 
-    config = {
-        "env": {
-            "ANTHROPIC_BASE_URL": anthropic_url,
-            "ANTHROPIC_AUTH_TOKEN": auth_token,
-            "ANTHROPIC_MODEL": main_model,
-            "ANTHROPIC_SMALL_FAST_MODEL": fast_model,
-            "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1",
-        }
+    env_config = {
+        "ANTHROPIC_BASE_URL": anthropic_url,
+        "ANTHROPIC_AUTH_TOKEN": auth_token,
+        "ANTHROPIC_MODEL": main_model,
+        "ANTHROPIC_SMALL_FAST_MODEL": fast_model,
+        "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1",
     }
 
+    # Load existing settings to preserve other sections (e.g., MCP servers)
+    existing_config: dict = {}
+    if settings_path.exists():
+        try:
+            with open(settings_path, encoding="utf-8") as f:
+                existing_config = json.load(f)
+        except (json.JSONDecodeError, OSError):
+            pass  # If file is corrupted, start fresh
+
+    # Merge: update env section while preserving other sections
+    existing_config["env"] = env_config
+
     settings_path.parent.mkdir(parents=True, exist_ok=True)
     with open(settings_path, "w", encoding="utf-8") as f:
-        json.dump(config, f, indent=2)
+        json.dump(existing_config, f, indent=2)
 
     console.print(
         Panel(

{router_maestro-0.1.3 → router_maestro-0.1.5}/src/router_maestro/routing/router.py

@@ -382,7 +382,7 @@ class Router:
         try:
             await provider.ensure_token()
             if is_stream:
-                stream = provider.chat_completion_stream(actual_request)
+                stream = await provider.chat_completion_stream(actual_request)
                 logger.info("Stream request routed to %s", provider_name)
                 return stream, provider_name
             else:
@@ -417,7 +417,7 @@ class Router:
                 try:
                     await other_provider.ensure_token()
                     if is_stream:
-                        stream = other_provider.chat_completion_stream(fallback_request)
+                        stream = await other_provider.chat_completion_stream(fallback_request)
                         logger.info("Stream fallback succeeded via %s", other_name)
                         return stream, other_name
                     else:

{router_maestro-0.1.3 → router_maestro-0.1.5}/src/router_maestro/server/routes/anthropic.py

@@ -26,6 +26,7 @@ from router_maestro.utils import (
     get_logger,
     map_openai_stop_reason_to_anthropic,
 )
+from router_maestro.utils.tokens import AnthropicStopReason
 
 logger = get_logger("server.routes.anthropic")
 
@@ -106,7 +107,7 @@ async def count_tokens(request: AnthropicCountTokensRequest):
 
     # Count messages
     for msg in request.messages:
-        content = msg.content if hasattr(msg, "content") else msg.get("content", "")
+        content = msg.content
         if isinstance(content, str):
             total_chars += len(content)
         elif isinstance(content, list):
@@ -115,12 +116,12 @@ async def count_tokens(request: AnthropicCountTokensRequest):
                     if block.get("type") == "text":
                         total_chars += len(block.get("text", ""))
                 elif hasattr(block, "text"):
-                    total_chars += len(block.text)
+                    total_chars += len(block.text)  # type: ignore[union-attr]
 
     return {"input_tokens": estimate_tokens_from_char_count(total_chars)}
 
 
-def _map_finish_reason(reason: str | None) -> str | None:
+def _map_finish_reason(reason: str | None) -> AnthropicStopReason | None:
     """Map OpenAI finish reason to Anthropic stop reason."""
     return map_openai_stop_reason_to_anthropic(reason)
 
@@ -144,7 +145,7 @@ def _estimate_input_tokens(request: AnthropicMessagesRequest) -> int:
 
     # Count messages
     for msg in request.messages:
-        content = msg.content if hasattr(msg, "content") else msg.get("content", "")
+        content = msg.content
         if isinstance(content, str):
             total_chars += len(content)
         elif isinstance(content, list):
@@ -161,7 +162,7 @@ def _estimate_input_tokens(request: AnthropicMessagesRequest) -> int:
                                 if isinstance(tc, dict) and tc.get("type") == "text":
                                     total_chars += len(tc.get("text", ""))
                 elif hasattr(block, "text"):
-                    total_chars += len(block.text)
+                    total_chars += len(block.text)  # type: ignore[union-attr]
 
     # Count tools definitions if present
     if request.tools:

{router_maestro-0.1.3 → router_maestro-0.1.5}/src/router_maestro/utils/tokens.py

@@ -1,8 +1,14 @@
 """Token estimation utilities."""
 
+from typing import Literal
+
 # Approximate characters per token for English text
 CHARS_PER_TOKEN = 4
 
+AnthropicStopReason = Literal[
+    "end_turn", "max_tokens", "stop_sequence", "tool_use", "pause_turn", "refusal"
+]
+
 
 def estimate_tokens(text: str) -> int:
     """Estimate token count from text.
@@ -31,7 +37,9 @@ def estimate_tokens_from_char_count(char_count: int) -> int:
     return char_count // CHARS_PER_TOKEN
 
 
-def map_openai_stop_reason_to_anthropic(openai_reason: str | None) -> str | None:
+def map_openai_stop_reason_to_anthropic(
+    openai_reason: str | None,
+) -> AnthropicStopReason | None:
     """Map OpenAI finish reason to Anthropic stop reason.
 
     Args:
@@ -42,7 +50,7 @@ def map_openai_stop_reason_to_anthropic(openai_reason: str | None) -> str | None
     """
     if openai_reason is None:
         return None
-    mapping = {
+    mapping: dict[str, AnthropicStopReason] = {
         "stop": "end_turn",
         "length": "max_tokens",
         "tool_calls": "tool_use",

{router_maestro-0.1.3 → router_maestro-0.1.5}/uv.lock

@@ -925,7 +925,7 @@ wheels = [
 
 [[package]]
 name = "router-maestro"
-version = "0.1.3"
+version = "0.1.5"
 source = { editable = "." }
 dependencies = [
     { name = "aiosqlite" },