python health mcp tool (Azure#41406)

* initial commit

* add health tool to joint mcp server

* remove separate health mcp server

* formatting

* fix whitespace

* doc

Author: kristapratico

.github/copilot-instructions.md

@@ -223,4 +223,42 @@ tox -e pylint --c <path_to_tox.ini> --root .
 **REQUIREMENTS:**
 - Use Python 3.9 compatible environment
 - Follow official fixing guidelines
-- Use tox mcp tool for running MyPy
+- Use tox mcp tool for running MyPy
+
+---
+
+## Python SDK Health tool
+
+- Use the azure-sdk-python-mcp mcp tool to lookup a library's health status.
+- Always include the date of last update based on the Last Refresh date.
+- Explanation of statuses can be found here: https://github.com/Azure/azure-sdk-for-python/blob/main/doc/repo_health_status.md
+- Release blocking checks are MyPy, Pylint, Sphinx, and Tests - CI. These checks should all PASS. If not PASS, mention that the library is blocked for release.
+- If links are available in the table, make the statuses (e.g. PASS, WARNING, etc) you report linked. Avoid telling the user to check the links in the report themselves.
+- Don't share information like SDK Owned
+
+### Example
+
+As of <Last Refresh date>, here is the health status for azure-ai-projects:
+
+Overall Status: ⚠️ NEEDS_ACTION
+
+✅ Passing Checks:
+
+Pyright: PASS
+Sphinx: PASS
+Type Checked Samples: ENABLED
+SLA Questions and Bugs: 0
+
+⚠️ Areas Needing Attention:
+
+Pylint: WARNING
+Tests - Live: ❓ UNKNOWN
+Tests - Samples: ❌ DISABLED
+Customer-reported issues: 🔴 5 open issues
+
+❌ Release blocking
+
+Mypy: FAIL
+Tests - CI: FAIL
+
+This library is failing two release blocking checks - Mypy and Tests - CI. The library needs attention primarily due to Pylint warnings, disabled sample tests, and open customer-reported issues.

.vscode/mcp.json

@@ -8,6 +8,6 @@
                 "run",
                 "main.py"
             ],
-        },
+        }
     }
-  }
+}

doc/dev/ai/typespec_generation.md

@@ -136,3 +136,12 @@ Copilot will guide you through these steps. It will inform you of expected durat
 *   **Your action:** Copy the PR URL and use the `azure-rest-api-specs` agent as instructed.
 
 By following this guide and interacting effectively with Copilot, you can significantly accelerate the SDK generation process while ensuring adherence to Azure SDK standards.
+
+
+### Checking the health status of a library
+
+To get a quick glance of a library's health and whether it can pass CI checks to release the package, you can ask VS Code Copilot to check its health status. For example, while in `Agent` mode, you can say:
+
+`What is the health status of azure-ai-projects?`
+
+This will report the library's status from [aka.ms/azsdk/python/health](https://www.aka.ms/azsdk/python/health) and help identify any blockers for release.

tools/mcp/azure-sdk-python-mcp/main.py

@@ -1,11 +1,15 @@
-from mcp.server.fastmcp import FastMCP
-import subprocess
-from typing import Dict, List, Optional, Any
 import logging
 import sys
 import os
 import re
+import csv
+import subprocess
+from io import StringIO
+from typing import Dict, List, Optional, Any
+
+import httpx
 from github import Github
+from mcp.server.fastmcp import FastMCP
 
 # Create FastMCP instance
 mcp = FastMCP("azure-sdk-python-mcp")
@@ -242,6 +246,51 @@ def init_local_tool(tsp_config_path: str, repo_path:str) -> Dict[str, Any]:
             "code": 1
         }
 
+
+@mcp.tool("check_library_health")
+def check_library_health(library_name: str) -> Dict[str, Any]:
+    """Checks the health status of a client library.
+
+    :param str library_name: The name of the library to check.
+    :returns: A dictionary containing the result of the command.
+    """
+
+    url = "https://raw.githubusercontent.com/Azure/azure-sdk-for-python/python-sdk-health-report/scripts/repo_health_status_report/health_report.csv"
+    response = httpx.get(url)
+
+    try:
+        response.raise_for_status()
+
+        csv_data = StringIO(response.text)
+        reader = csv.reader(csv_data)
+        
+        headers = next(reader)
+        column_headers = headers[1:]
+
+        result = {}
+        for row in reader:
+            if row:
+                key = row[0]
+                values_dict = {header: value for header, value in zip(column_headers, row[1:])}
+                result[key] = values_dict
+        
+    except httpx.HTTPError as e:
+        logger.error(f"Error downloading health report: {e}")
+        return {
+            "success": False,
+            "stderr": f"Failed to fetch health report for {library_name}. Status code: {response.status_code}",
+            "stdout": "",
+            "code": response.status_code
+        }
+
+    return {
+        "success": True,
+        "stdout": result[library_name] if library_name in result else f"No health data found for {library_name}",
+        "stderr": "",
+        "code": response.status_code
+    }
+
+
 # Run the MCP server
 if __name__ == "__main__":
-    mcp.run(transport='stdio')
+    mcp.run(transport='stdio')

tools/mcp/azure-sdk-python-mcp/pyproject.toml

@@ -13,4 +13,5 @@ dependencies = [
     "pygithub>=2.6.1",
     "pylint>=3.3.7",
     "tox>=4.26.0",
+    "httpx>=0.24.0",
 ]