gfp-mcp 0.2.1__tar.gz → 0.2.4__tar.gz

{gfp_mcp-0.2.1 → gfp_mcp-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gfp-mcp
-Version: 0.2.1
+Version: 0.2.4
 Summary: Model Context Protocol (MCP) server for GDSFactory+ photonic IC design
 Author: GDSFactory+ Team
 License: MIT
@@ -33,13 +33,15 @@ Requires-Dist: pytest>=7.0.0; extra == "dev"
 Requires-Dist: pytest-cov>=4.0.0; extra == "dev"
 Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
 Requires-Dist: ruff>=0.1.0; extra == "dev"
+Requires-Dist: bump-my-version>=0.26.0; extra == "dev"
+Requires-Dist: build>=1.4.0; extra == "dev"
 Dynamic: license-file
 
 # GDSFactory+ MCP Server
 
 [![PyPI version](https://img.shields.io/pypi/v/gfp-mcp.svg)](https://pypi.org/project/gfp-mcp/)
 [![Python versions](https://img.shields.io/pypi/pyversions/gfp-mcp.svg)](https://pypi.org/project/gfp-mcp/)
-[![Tests](https://github.com/doplaydo/gfp-mcp/workflows/Tests/badge.svg)](https://github.com/doplaydo/gfp-mcp/actions)
+[![Tests](https://github.com/doplaydo/gfp-mcp/actions/workflows/test.yml/badge.svg)](https://github.com/doplaydo/gfp-mcp/actions)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 
 Model Context Protocol (MCP) server for GDSFactory+ that enables AI assistants like Claude to design and build photonic integrated circuits.
@@ -57,15 +59,32 @@ This MCP server connects AI assistants to [GDSFactory+](https://gdsfactory.com),
 
 ### 2. Install the MCP Server
 
+**With uv (recommended):**
+
+If you don't have `uv` installed, see the [uv installation guide](https://docs.astral.sh/uv/getting-started/installation/).
+
 ```bash
-pip install gfp-mcp
+uv tool install gfp-mcp
 ```
 
-Or with uv:
+<details>
+<summary>Ephemeral approach</summary>
+
+Run without installing:
 
 ```bash
 uvx --from gfp-mcp gfp-mcp-serve
 ```
+</details>
+
+<details>
+<summary><strong>Alternative: pip install</strong></summary>
+
+```bash
+pip install gfp-mcp
+```
+
+</details>
 
 ### 3. Connect to Your AI Assistant
 
@@ -143,14 +162,12 @@ Try these commands with your AI assistant:
 
 ## Available Tools
 
-- **build_cell** - Build a single GDS cell by name
-- **build_cells** - Build multiple GDS cells in batch
+- **build_cells** - Build one or more GDS cells by name (pass a list, can be single-item)
 - **list_cells** - List all available photonic components
 - **get_cell_info** - Get detailed component metadata
-- **download_gds** - Download built GDS files
 - **list_projects** - List all running GDSFactory+ server instances
 - **get_project_info** - Get detailed information about a specific project
-- **check_drc** - Run Design Rule Check verification
+- **check_drc** - Run Design Rule Check verification (returns structured format with all violations including simplified location data for LLM-friendly troubleshooting)
 - **check_connectivity** - Run connectivity verification
 - **check_lvs** - Run Layout vs. Schematic verification
 

{gfp_mcp-0.2.1 → gfp_mcp-0.2.4}/README.md

@@ -2,7 +2,7 @@
 
 [![PyPI version](https://img.shields.io/pypi/v/gfp-mcp.svg)](https://pypi.org/project/gfp-mcp/)
 [![Python versions](https://img.shields.io/pypi/pyversions/gfp-mcp.svg)](https://pypi.org/project/gfp-mcp/)
-[![Tests](https://github.com/doplaydo/gfp-mcp/workflows/Tests/badge.svg)](https://github.com/doplaydo/gfp-mcp/actions)
+[![Tests](https://github.com/doplaydo/gfp-mcp/actions/workflows/test.yml/badge.svg)](https://github.com/doplaydo/gfp-mcp/actions)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 
 Model Context Protocol (MCP) server for GDSFactory+ that enables AI assistants like Claude to design and build photonic integrated circuits.
@@ -20,15 +20,32 @@ This MCP server connects AI assistants to [GDSFactory+](https://gdsfactory.com),
 
 ### 2. Install the MCP Server
 
+**With uv (recommended):**
+
+If you don't have `uv` installed, see the [uv installation guide](https://docs.astral.sh/uv/getting-started/installation/).
+
 ```bash
-pip install gfp-mcp
+uv tool install gfp-mcp
 ```
 
-Or with uv:
+<details>
+<summary>Ephemeral approach</summary>
+
+Run without installing:
 
 ```bash
 uvx --from gfp-mcp gfp-mcp-serve
 ```
+</details>
+
+<details>
+<summary><strong>Alternative: pip install</strong></summary>
+
+```bash
+pip install gfp-mcp
+```
+
+</details>
 
 ### 3. Connect to Your AI Assistant
 
@@ -106,14 +123,12 @@ Try these commands with your AI assistant:
 
 ## Available Tools
 
-- **build_cell** - Build a single GDS cell by name
-- **build_cells** - Build multiple GDS cells in batch
+- **build_cells** - Build one or more GDS cells by name (pass a list, can be single-item)
 - **list_cells** - List all available photonic components
 - **get_cell_info** - Get detailed component metadata
-- **download_gds** - Download built GDS files
 - **list_projects** - List all running GDSFactory+ server instances
 - **get_project_info** - Get detailed information about a specific project
-- **check_drc** - Run Design Rule Check verification
+- **check_drc** - Run Design Rule Check verification (returns structured format with all violations including simplified location data for LLM-friendly troubleshooting)
 - **check_connectivity** - Run connectivity verification
 - **check_lvs** - Run Layout vs. Schematic verification
 

{gfp_mcp-0.2.1 → gfp_mcp-0.2.4}/gfp_mcp.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gfp-mcp
-Version: 0.2.1
+Version: 0.2.4
 Summary: Model Context Protocol (MCP) server for GDSFactory+ photonic IC design
 Author: GDSFactory+ Team
 License: MIT
@@ -33,13 +33,15 @@ Requires-Dist: pytest>=7.0.0; extra == "dev"
 Requires-Dist: pytest-cov>=4.0.0; extra == "dev"
 Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
 Requires-Dist: ruff>=0.1.0; extra == "dev"
+Requires-Dist: bump-my-version>=0.26.0; extra == "dev"
+Requires-Dist: build>=1.4.0; extra == "dev"
 Dynamic: license-file
 
 # GDSFactory+ MCP Server
 
 [![PyPI version](https://img.shields.io/pypi/v/gfp-mcp.svg)](https://pypi.org/project/gfp-mcp/)
 [![Python versions](https://img.shields.io/pypi/pyversions/gfp-mcp.svg)](https://pypi.org/project/gfp-mcp/)
-[![Tests](https://github.com/doplaydo/gfp-mcp/workflows/Tests/badge.svg)](https://github.com/doplaydo/gfp-mcp/actions)
+[![Tests](https://github.com/doplaydo/gfp-mcp/actions/workflows/test.yml/badge.svg)](https://github.com/doplaydo/gfp-mcp/actions)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 
 Model Context Protocol (MCP) server for GDSFactory+ that enables AI assistants like Claude to design and build photonic integrated circuits.
@@ -57,15 +59,32 @@ This MCP server connects AI assistants to [GDSFactory+](https://gdsfactory.com),
 
 ### 2. Install the MCP Server
 
+**With uv (recommended):**
+
+If you don't have `uv` installed, see the [uv installation guide](https://docs.astral.sh/uv/getting-started/installation/).
+
 ```bash
-pip install gfp-mcp
+uv tool install gfp-mcp
 ```
 
-Or with uv:
+<details>
+<summary>Ephemeral approach</summary>
+
+Run without installing:
 
 ```bash
 uvx --from gfp-mcp gfp-mcp-serve
 ```
+</details>
+
+<details>
+<summary><strong>Alternative: pip install</strong></summary>
+
+```bash
+pip install gfp-mcp
+```
+
+</details>
 
 ### 3. Connect to Your AI Assistant
 
@@ -143,14 +162,12 @@ Try these commands with your AI assistant:
 
 ## Available Tools
 
-- **build_cell** - Build a single GDS cell by name
-- **build_cells** - Build multiple GDS cells in batch
+- **build_cells** - Build one or more GDS cells by name (pass a list, can be single-item)
 - **list_cells** - List all available photonic components
 - **get_cell_info** - Get detailed component metadata
-- **download_gds** - Download built GDS files
 - **list_projects** - List all running GDSFactory+ server instances
 - **get_project_info** - Get detailed information about a specific project
-- **check_drc** - Run Design Rule Check verification
+- **check_drc** - Run Design Rule Check verification (returns structured format with all violations including simplified location data for LLM-friendly troubleshooting)
 - **check_connectivity** - Run connectivity verification
 - **check_lvs** - Run Layout vs. Schematic verification
 

{gfp_mcp-0.2.1 → gfp_mcp-0.2.4}/gfp_mcp.egg-info/requires.txt

@@ -10,3 +10,5 @@ pytest>=7.0.0
 pytest-cov>=4.0.0
 pytest-asyncio>=0.21.0
 ruff>=0.1.0
+bump-my-version>=0.26.0
+build>=1.4.0

{gfp_mcp-0.2.1 → gfp_mcp-0.2.4}/mcp_standalone/__init__.py

@@ -39,4 +39,4 @@ __all__ = [
     "get_resource_content",
 ]
 
-__version__ = "0.2.1"
+__version__ = "0.2.4"

{gfp_mcp-0.2.1 → gfp_mcp-0.2.4}/mcp_standalone/client.py

@@ -43,6 +43,21 @@ class FastAPIClient:
         self._client: httpx.AsyncClient | None = None
         self._registry = ServerRegistry()
 
+    def _has_available_servers(self) -> bool:
+        """Check if any servers are available in the registry."""
+        return len(self._registry.list_servers()) > 0
+
+    def _get_default_server_url(self) -> str | None:
+        """Get the first available server URL from registry if no base_url configured."""
+        if self.base_url:
+            return self.base_url
+
+        servers = self._registry.list_servers()
+        if servers:
+            return f"http://localhost:{servers[0].port}"
+
+        return None
+
     async def __aenter__(self) -> Self:
         """Enter async context."""
         await self.start()
@@ -55,12 +70,17 @@ class FastAPIClient:
     async def start(self) -> None:
         """Start the HTTP client with connection pooling."""
         if self._client is None:
+            base_url = self.base_url or "http://localhost"
+
             self._client = httpx.AsyncClient(
-                base_url=self.base_url,
+                base_url=base_url,
                 timeout=httpx.Timeout(self.timeout),
                 limits=httpx.Limits(max_keepalive_connections=5, max_connections=10),
             )
-            logger.debug("HTTP client started with base URL: %s", self.base_url)
+            logger.debug(
+                "HTTP client started with base URL: %s (resolved per-request if needed)",
+                self.base_url or "from registry",
+            )
 
     async def close(self) -> None:
         """Close the HTTP client."""
@@ -79,22 +99,46 @@ class FastAPIClient:
             Base URL for the request
 
         Raises:
-            ValueError: If project not found in registry
+            ValueError: If base URL cannot be resolved
         """
-        # If no project specified, use default base_url
-        if project is None:
+        if project is not None:
+            server_info = self._registry.get_server_by_project(project)
+            if server_info is None:
+                available = self._registry.list_servers()
+                if available:
+                    project_list = ", ".join([s.project_name for s in available[:3]])
+                    if len(available) > 3:
+                        project_list += f", ... and {len(available) - 3} more"
+                    msg = (
+                        f"Project '{project}' not found in registry. "
+                        f"Available projects: {project_list}. "
+                        "Use list_projects tool to see all running servers."
+                    )
+                else:
+                    msg = (
+                        f"Project '{project}' not found. No GDSFactory+ servers are running. "
+                        "Please open a GDSFactory+ project in VSCode with the extension installed."
+                    )
+                raise ValueError(msg)
+
+            return f"http://localhost:{server_info.port}"
+
+        if self.base_url:
             return self.base_url
 
-        # Look up project in registry
-        server_info = self._registry.get_server_by_project(project)
-        if server_info is None:
-            msg = (
-                f"Project '{project}' not found in registry. "
-                "Make sure the server is running for this project."
+        default_url = self._get_default_server_url()
+        if default_url:
+            logger.info(
+                "No project specified, using first available server: %s", default_url
             )
-            raise ValueError(msg)
+            return default_url
 
-        return f"http://localhost:{server_info.port}"
+        msg = (
+            "No project specified and no GDSFactory+ servers are running. "
+            "Either: (1) Start a server by opening a GDSFactory+ project in VSCode, "
+            "(2) Specify a project parameter, or (3) Set GFP_API_URL environment variable."
+        )
+        raise ValueError(msg)
 
     async def request(
         self,
@@ -125,7 +169,6 @@ class FastAPIClient:
         if self._client is None:
             await self.start()
 
-        # Resolve the base URL for this request
         base_url = self._resolve_base_url(project)
 
         last_error = None
@@ -143,7 +186,6 @@ class FastAPIClient:
                     base_url,
                 )
 
-                # Build full URL with the resolved base URL
                 full_url = f"{base_url}{path}"
 
                 response = await self._client.request(  # type: ignore[union-attr]
@@ -155,7 +197,6 @@ class FastAPIClient:
                 )
                 response.raise_for_status()
 
-                # Try to parse JSON, fall back to text
                 try:
                     return response.json()
                 except (ValueError, TypeError):
@@ -165,19 +206,16 @@ class FastAPIClient:
                 last_error = e
                 logger.warning("Request failed (attempt %d): %s", attempt + 1, e)
 
-                # Don't retry on client errors (4xx)
                 if (
                     isinstance(e, httpx.HTTPStatusError)
                     and 400 <= e.response.status_code < 500
                 ):
                     raise
 
-                # Exponential backoff for retries
                 if attempt < MCPConfig.MAX_RETRIES - 1:
                     await asyncio.sleep(backoff)
                     backoff *= 2
 
-        # All retries failed
         logger.error("All %d attempts failed", MCPConfig.MAX_RETRIES)
         raise last_error  # type: ignore[misc]
 

{gfp_mcp-0.2.1 → gfp_mcp-0.2.4}/mcp_standalone/config.py

@@ -15,34 +15,28 @@ class MCPConfig:
     that proxies requests to the FastAPI backend.
     """
 
-    # FastAPI base URL (default: http://localhost:8787)
-    # This default is primarily for backward compatibility.
-    # The MCP server automatically discovers running servers via the registry.
-    API_URL: Final[str] = os.getenv("GFP_API_URL", "http://localhost:8787")
+    API_URL: Final[str | None] = os.getenv("GFP_API_URL")
 
-    # Timeout for tool calls in seconds (default: 300 = 5 minutes)
     TIMEOUT: Final[int] = int(os.getenv("GFP_MCP_TIMEOUT", "300"))
 
-    # Enable debug logging
     DEBUG: Final[bool] = os.getenv("GFP_MCP_DEBUG", "false").lower() in (
         "true",
         "1",
         "yes",
     )
 
-    # Retry configuration
     MAX_RETRIES: Final[int] = 3
     RETRY_BACKOFF: Final[float] = 0.5  # Initial backoff in seconds
 
     @classmethod
-    def get_api_url(cls, override: str | None = None) -> str:
+    def get_api_url(cls, override: str | None = None) -> str | None:
         """Get the FastAPI base URL.
 
         Args:
             override: Optional URL to override the environment variable
 
         Returns:
-            The API base URL
+            The API base URL or None if not configured
         """
         return override or cls.API_URL