mcp-code-indexer 1.6.5__py3-none-any.whl → 1.8.0__py3-none-any.whl

mcp_code_indexer/git_hook_handler.py

@@ -49,17 +49,18 @@ class GitHookHandler:
     OPENROUTER_API_URL = "https://openrouter.ai/api/v1/chat/completions"
     OPENROUTER_MODEL = "anthropic/claude-sonnet-4"
     
-    def __init__(self, db_manager: DatabaseManager, cache_dir: Path):
+    def __init__(self, db_manager: DatabaseManager, cache_dir: Path, logger: Optional[logging.Logger] = None):
         """
         Initialize GitHookHandler.
         
         Args:
             db_manager: Database manager instance
             cache_dir: Cache directory for temporary files
+            logger: Logger instance to use (optional, creates default if not provided)
         """
         self.db_manager = db_manager
         self.cache_dir = cache_dir
-        self.logger = logging.getLogger(__name__)
+        self.logger = logger if logger is not None else logging.getLogger(__name__)
         self.token_counter = TokenCounter()
         
         # Git hook specific settings
@@ -136,32 +137,26 @@ class GitHookHandler:
             self.logger.info(f"Current overview length: {len(current_overview) if current_overview else 0} characters")
             self.logger.info(f"Current descriptions count: {len(current_descriptions)}")
             
-            # Build prompt for OpenRouter
-            self.logger.info("Building analysis prompt...")
-            prompt = self._build_githook_prompt(
-                git_diff, 
-                commit_message,
-                current_overview, 
-                current_descriptions,
-                changed_files
-            )
-            
-            # Log prompt details
-            prompt_chars = len(prompt)
-            prompt_tokens = self.token_counter.count_tokens(prompt)
-            self.logger.info(f"Analysis prompt: {prompt_chars} characters, {prompt_tokens} tokens")
+            # Use two-stage approach for large codebases
+            self.logger.info("Starting two-stage analysis approach...")
             
-            # Check total prompt token count
-            if prompt_tokens > self.config["max_diff_tokens"]:
-                self.logger.info(f"Skipping git hook update - prompt too large ({prompt_tokens} tokens > {self.config['max_diff_tokens']} limit)")
-                return
+            # Stage 1: Check if overview needs updating
+            overview_updates = await self._analyze_overview_updates(
+                git_diff, commit_message, current_overview, changed_files
+            )
             
-            self.logger.info(f"Prompt size OK ({prompt_tokens} <= {self.config['max_diff_tokens']} tokens), calling OpenRouter...")
+            # Stage 2: Update file descriptions 
+            file_updates = await self._analyze_file_updates(
+                git_diff, commit_message, current_descriptions, changed_files
+            )
             
-            # Call OpenRouter API
-            updates = await self._call_openrouter(prompt)
+            # Combine updates
+            updates = {
+                "file_updates": file_updates.get("file_updates", {}),
+                "overview_update": overview_updates.get("overview_update")
+            }
             
-            self.logger.info(f"OpenRouter response received, processing updates...")
+            self.logger.info(f"Two-stage analysis completed, processing updates...")
             
             # Apply updates to database
             await self._apply_updates(project_info, updates)
@@ -440,101 +435,143 @@ class GitHookHandler:
             self.logger.warning(f"Failed to get file descriptions: {e}")
             return {}
     
-    def _build_githook_prompt(
-        self, 
-        git_diff: str, 
-        commit_message: str,
-        overview: str, 
-        descriptions: Dict[str, str], 
+    async def _analyze_overview_updates(
+        self,
+        git_diff: str,
+        commit_message: str, 
+        current_overview: str,
         changed_files: List[str]
-    ) -> str:
+    ) -> Dict[str, Any]:
         """
-        Build prompt for OpenRouter API to analyze git changes.
+        Stage 1: Analyze if project overview needs updating.
         
         Args:
             git_diff: Git diff content
             commit_message: Commit message explaining the changes
-            overview: Current project overview
-            descriptions: Current file descriptions
+            current_overview: Current project overview
             changed_files: List of changed file paths
             
         Returns:
-            Formatted prompt for the API
+            Dict with overview_update key
         """
-        return f"""Analyze this git commit and update the file descriptions and project overview as needed.
+        self.logger.info("Stage 1: Analyzing overview updates...")
+        
+        prompt = f"""Analyze this git commit to determine if the project overview needs updating.
 
 COMMIT MESSAGE:
 {commit_message or "No commit message available"}
 
 CURRENT PROJECT OVERVIEW:
-{overview or "No overview available"}
+{current_overview or "No overview available"}
 
-CURRENT FILE DESCRIPTIONS:
-{json.dumps(descriptions, indent=2)}
+CHANGED FILES:
+{', '.join(changed_files)}
 
 GIT DIFF:
 {git_diff}
 
-CHANGED FILES:
-{', '.join(changed_files)}
-
 INSTRUCTIONS:
 
-Use the COMMIT MESSAGE to understand the intent and context of the changes. The commit message explains what the developer was trying to accomplish.
+Update project overview ONLY if there are major structural changes like:
+- New major features or components (indicated by commit message or new directories)
+- Architectural changes (new patterns, frameworks, or approaches)
+- Significant dependency additions (Cargo.toml, package.json, requirements.txt changes)
+- New API endpoints or workflows
+- Changes to build/deployment processes
+
+Do NOT update for: bug fixes, small refactors, documentation updates, version bumps.
+
+If updating, provide comprehensive narrative (10-20 pages of text) with directory structure, architecture, components, and workflows.
 
-1. **File Descriptions**: Update descriptions for any files that have changed significantly. Consider both the diff content and the commit message context. Only include files that need actual description updates.
+Return ONLY a JSON object:
+{{
+  "overview_update": "Updated overview text" or null
+}}"""
+
+        # Log prompt details
+        prompt_chars = len(prompt)
+        prompt_tokens = self.token_counter.count_tokens(prompt)
+        self.logger.info(f"Stage 1 prompt: {prompt_chars} characters, {prompt_tokens} tokens")
+        
+        if prompt_tokens > self.config["max_diff_tokens"]:
+            self.logger.warning(f"Stage 1 prompt too large ({prompt_tokens} tokens), skipping overview analysis")
+            return {"overview_update": None}
+            
+        # Call OpenRouter API
+        result = await self._call_openrouter(prompt)
+        self.logger.info("Stage 1 completed: overview analysis")
+        
+        return result
 
-2. **Project Overview**: Update ONLY if there are major structural changes like:
-   - New major features or components (which may be indicated by commit message)
-   - Architectural changes (new patterns, frameworks, or approaches)
-   - Significant dependency additions
-   - New API endpoints or workflows
-   - Changes to build/deployment processes
-   
-   Do NOT update overview for minor changes like bug fixes, small refactors, or documentation updates.
+    async def _analyze_file_updates(
+        self,
+        git_diff: str,
+        commit_message: str,
+        current_descriptions: Dict[str, str],
+        changed_files: List[str]
+    ) -> Dict[str, Any]:
+        """
+        Stage 2: Analyze file description updates.
+        
+        Args:
+            git_diff: Git diff content
+            commit_message: Commit message explaining the changes
+            current_descriptions: Current file descriptions for changed files only
+            changed_files: List of changed file paths
+            
+        Returns:
+            Dict with file_updates key
+        """
+        self.logger.info("Stage 2: Analyzing file description updates...")
+        
+        # Only include descriptions for changed files to reduce token usage
+        relevant_descriptions = {
+            path: desc for path, desc in current_descriptions.items() 
+            if path in changed_files
+        }
+        
+        prompt = f"""Analyze this git commit and update file descriptions for changed files.
 
-3. **Overview Format**: If updating the overview, follow this structure with comprehensive narrative (10-20 pages of text):
+COMMIT MESSAGE:
+{commit_message or "No commit message available"}
 
-````
-## Directory Structure
-```
-src/
-├── api/          # REST API endpoints and middleware
-├── models/       # Database models and business logic  
-├── services/     # External service integrations
-├── utils/        # Shared utilities and helpers
-└── tests/        # Test suites
-```
+CURRENT FILE DESCRIPTIONS (for changed files only):
+{json.dumps(relevant_descriptions, indent=2)}
 
-## Architecture Overview
-[Describe how components interact, data flow, key design decisions]
+CHANGED FILES:
+{', '.join(changed_files)}
 
-## Core Components
-### API Layer
-[Details about API structure, authentication, routing]
+GIT DIFF:
+{git_diff}
 
-### Data Model
-[Key entities, relationships, database design]
+INSTRUCTIONS:
 
-## Key Workflows
-1. User Authentication Flow
-   [Step-by-step description]
-2. Data Processing Pipeline
-   [How data moves through the system]
+Use the COMMIT MESSAGE to understand the intent and context of the changes.
 
-[Continue with other sections...]
-````
+Update descriptions for files that have changed significantly. Consider both the diff content and commit message context. Only include files that need actual description updates.
 
-Return ONLY a JSON object in this exact format:
+Return ONLY a JSON object:
 {{
   "file_updates": {{
     "path/to/file1.py": "Updated description for file1",
     "path/to/file2.js": "Updated description for file2"
-  }},
-  "overview_update": "Updated project overview text (or null if no update needed)"
-}}
+  }}
+}}"""
 
-Return ONLY the JSON, no other text."""
+        # Log prompt details  
+        prompt_chars = len(prompt)
+        prompt_tokens = self.token_counter.count_tokens(prompt)
+        self.logger.info(f"Stage 2 prompt: {prompt_chars} characters, {prompt_tokens} tokens")
+        
+        if prompt_tokens > self.config["max_diff_tokens"]:
+            self.logger.warning(f"Stage 2 prompt too large ({prompt_tokens} tokens), skipping file analysis")
+            return {"file_updates": {}}
+            
+        # Call OpenRouter API
+        result = await self._call_openrouter(prompt)
+        self.logger.info("Stage 2 completed: file description analysis")
+        
+        return result
     
     @retry(
         wait=wait_exponential(multiplier=1, min=4, max=60),
@@ -633,19 +670,32 @@ Return ONLY the JSON, no other text."""
         try:
             data = json.loads(response_text.strip())
             
-            # Validate structure
-            if "file_updates" not in data:
-                raise ValueError("Missing 'file_updates' field")
-            if "overview_update" not in data:
-                raise ValueError("Missing 'overview_update' field")
-            
-            if not isinstance(data["file_updates"], dict):
-                raise ValueError("'file_updates' must be a dictionary")
-            
-            # Validate descriptions
-            for path, desc in data["file_updates"].items():
-                if not isinstance(desc, str) or not desc.strip():
-                    raise ValueError(f"Invalid description for {path}")
+            # Handle both single-stage and two-stage responses
+            if "file_updates" in data and "overview_update" in data:
+                # Original single-stage format
+                if not isinstance(data["file_updates"], dict):
+                    raise ValueError("'file_updates' must be a dictionary")
+                
+                # Validate descriptions
+                for path, desc in data["file_updates"].items():
+                    if not isinstance(desc, str) or not desc.strip():
+                        raise ValueError(f"Invalid description for {path}")
+                        
+            elif "file_updates" in data:
+                # Stage 2 format (file updates only)
+                if not isinstance(data["file_updates"], dict):
+                    raise ValueError("'file_updates' must be a dictionary")
+                    
+                # Validate descriptions
+                for path, desc in data["file_updates"].items():
+                    if not isinstance(desc, str) or not desc.strip():
+                        raise ValueError(f"Invalid description for {path}")
+                        
+            elif "overview_update" in data:
+                # Stage 1 format (overview only) - overview_update can be null
+                pass
+            else:
+                raise ValueError("Response must contain 'file_updates' and/or 'overview_update'")
             
             return data
             

mcp_code_indexer/logging_config.py

@@ -163,6 +163,9 @@ def setup_command_logger(
         
         logger.addHandler(file_handler)
         
+        # Set up component loggers to also log to this command's log file
+        _setup_component_loggers_for_command(command_name, file_handler, structured_formatter)
+        
         logger.info(f"=== {command_name.upper()} SESSION STARTED ===")
         
     except (OSError, PermissionError) as e:
@@ -175,6 +178,56 @@ def setup_command_logger(
     return logger
 
 
+def _setup_component_loggers_for_command(
+    command_name: str,
+    file_handler: logging.Handler,
+    formatter: logging.Formatter
+) -> None:
+    """
+    Set up component loggers to also send logs to the command's log file.
+    
+    Args:
+        command_name: Name of the command
+        file_handler: File handler to add to component loggers
+        formatter: Formatter to use for the handler
+    """
+    # List of component logger names that should also log to command files
+    component_loggers = [
+        "mcp_code_indexer.database.database",
+        "mcp_code_indexer.server.mcp_server", 
+        "mcp_code_indexer.token_counter",
+        "mcp_code_indexer.file_scanner",
+        "mcp_code_indexer.error_handler",
+        "mcp_code_indexer.merge_handler"
+    ]
+    
+    for component_logger_name in component_loggers:
+        component_logger = logging.getLogger(component_logger_name)
+        
+        # Create a separate handler for this command to avoid interference
+        command_handler = logging.handlers.RotatingFileHandler(
+            file_handler.baseFilename,
+            maxBytes=file_handler.maxBytes,
+            backupCount=file_handler.backupCount,
+            encoding='utf-8'
+        )
+        command_handler.setLevel(logging.DEBUG)
+        command_handler.setFormatter(formatter)
+        
+        # Add a marker to identify which command this handler belongs to
+        command_handler._command_name = command_name
+        
+        # Remove any existing handlers for this command (in case of multiple calls)
+        existing_handlers = [h for h in component_logger.handlers if hasattr(h, '_command_name') and h._command_name == command_name]
+        for handler in existing_handlers:
+            component_logger.removeHandler(handler)
+            handler.close()
+        
+        # Add the new handler
+        component_logger.addHandler(command_handler)
+        component_logger.setLevel(logging.DEBUG)  # Ensure component loggers capture all levels
+
+
 def log_performance_metrics(
     logger: logging.Logger,
     operation: str,

mcp_code_indexer/main.py

@@ -499,7 +499,7 @@ async def handle_githook(args: argparse.Namespace) -> None:
         logger.debug("Database initialized successfully")
         
         # Initialize git hook handler
-        git_handler = GitHookHandler(db_manager, cache_dir)
+        git_handler = GitHookHandler(db_manager, cache_dir, logger)
         logger.debug("Git hook handler initialized")
         
         # Run git hook analysis

{mcp_code_indexer-1.6.5.dist-info → mcp_code_indexer-1.8.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mcp-code-indexer
-Version: 1.6.5
+Version: 1.8.0
 Summary: MCP server that tracks file descriptions across codebases, enabling AI agents to efficiently navigate and understand code through searchable summaries and token-aware overviews.
 Author: MCP Code Indexer Contributors
 Maintainer: MCP Code Indexer Contributors
@@ -59,8 +59,8 @@ Dynamic: requires-python
 
 # MCP Code Indexer 🚀
 
-[![PyPI version](https://badge.fury.io/py/mcp-code-indexer.svg?8)](https://badge.fury.io/py/mcp-code-indexer)
-[![Python](https://img.shields.io/pypi/pyversions/mcp-code-indexer.svg?8)](https://pypi.org/project/mcp-code-indexer/)
+[![PyPI version](https://badge.fury.io/py/mcp-code-indexer.svg?10)](https://badge.fury.io/py/mcp-code-indexer)
+[![Python](https://img.shields.io/pypi/pyversions/mcp-code-indexer.svg?10)](https://pypi.org/project/mcp-code-indexer/)
 [![License](https://img.shields.io/badge/License-MIT-blue.svg)](https://opensource.org/licenses/MIT)
 
 A production-ready **Model Context Protocol (MCP) server** that revolutionizes how AI agents navigate and understand codebases. Instead of repeatedly scanning files, agents get instant access to intelligent descriptions, semantic search, and context-aware recommendations.

{mcp_code_indexer-1.6.5.dist-info → mcp_code_indexer-1.8.0.dist-info}/RECORD

@@ -2,9 +2,9 @@ mcp_code_indexer/__init__.py,sha256=GhY2NLQ6lH3n5mxqw0t8T1gmZGKhM6KvjhZH8xW5O-A,
 mcp_code_indexer/__main__.py,sha256=4Edinoe0ug43hobuLYcjTmGp2YJnlFYN4_8iKvUBJ0Q,213
 mcp_code_indexer/error_handler.py,sha256=cNSUFFrGBMLDv4qa78c7495L1wSl_dXCRbzCJOidx-Q,11590
 mcp_code_indexer/file_scanner.py,sha256=ctXeZMROgDThEtjzsANTK9TbK-fhTScMBd4iyuleBT4,11734
-mcp_code_indexer/git_hook_handler.py,sha256=9duggl8FTi4wm_vE2elPP1OohaPgq1EhWUUbGT0ae5Q,28732
-mcp_code_indexer/logging_config.py,sha256=yCGQD-xx9oobS-YctOFcaE1Q3iiuOj2E6cTfKHbh_wc,7358
-mcp_code_indexer/main.py,sha256=7k00hj2C1CxTDDErbq2Ee072MbvqEAsRqrMHH-w1oM0,31538
+mcp_code_indexer/git_hook_handler.py,sha256=_gM7TAcZ_H6tXVfh_gX0RwV0cJVdR_jhYFP9pQikLrc,30959
+mcp_code_indexer/logging_config.py,sha256=_bd9XGCLQ2VHPViJitaxGyREyfOXDPiklRh17jXeV0U,9523
+mcp_code_indexer/main.py,sha256=U-f3AJYdycWhjh-vLryj7aH8DGCs4d3x1yjA852HTxM,31546
 mcp_code_indexer/merge_handler.py,sha256=lJR8eVq2qSrF6MW9mR3Fy8UzrNAaQ7RsI2FMNXne3vQ,14692
 mcp_code_indexer/token_counter.py,sha256=WrifOkbF99nWWHlRlhCHAB2KN7qr83GOHl7apE-hJcE,8460
 mcp_code_indexer/data/stop_words_english.txt,sha256=7Zdd9ameVgA6tN_zuXROvHXD4hkWeELVywPhb7FJEkw,6343
@@ -17,9 +17,9 @@ mcp_code_indexer/server/__init__.py,sha256=16xMcuriUOBlawRqWNBk6niwrvtv_JD5xvI36
 mcp_code_indexer/server/mcp_server.py,sha256=4goDZmRmhPgipImgfhTVa6nYJM7L1p56h34ITO6JhSw,64431
 mcp_code_indexer/tiktoken_cache/9b5ad71b2ce5302211f9c61530b329a4922fc6a4,sha256=Ijkht27pm96ZW3_3OFE-7xAPtR0YyTWXoRO8_-hlsqc,1681126
 mcp_code_indexer/tools/__init__.py,sha256=m01mxML2UdD7y5rih_XNhNSCMzQTz7WQ_T1TeOcYlnE,49
-mcp_code_indexer-1.6.5.dist-info/licenses/LICENSE,sha256=JN9dyPPgYwH9C-UjYM7FLNZjQ6BF7kAzpF3_4PwY4rY,1086
-mcp_code_indexer-1.6.5.dist-info/METADATA,sha256=1vw3p7ngL5uKvvRWF8yfZXGWduc-Z7auUzuWHfQDsVk,17571
-mcp_code_indexer-1.6.5.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-mcp_code_indexer-1.6.5.dist-info/entry_points.txt,sha256=8HqWOw1Is7jOP1bvIgaSwouvT9z_Boe-9hd4NzyJOhY,68
-mcp_code_indexer-1.6.5.dist-info/top_level.txt,sha256=yKYCM-gMGt-cnupGfAhnZaoEsROLB6DQ1KFUuyKx4rw,17
-mcp_code_indexer-1.6.5.dist-info/RECORD,,
+mcp_code_indexer-1.8.0.dist-info/licenses/LICENSE,sha256=JN9dyPPgYwH9C-UjYM7FLNZjQ6BF7kAzpF3_4PwY4rY,1086
+mcp_code_indexer-1.8.0.dist-info/METADATA,sha256=GHDNwAAJ9Q-MBd3t9AMKkPDoHsz9ncyKzEkxf6Y_YIg,17573
+mcp_code_indexer-1.8.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+mcp_code_indexer-1.8.0.dist-info/entry_points.txt,sha256=8HqWOw1Is7jOP1bvIgaSwouvT9z_Boe-9hd4NzyJOhY,68
+mcp_code_indexer-1.8.0.dist-info/top_level.txt,sha256=yKYCM-gMGt-cnupGfAhnZaoEsROLB6DQ1KFUuyKx4rw,17
+mcp_code_indexer-1.8.0.dist-info/RECORD,,