mcp-code-indexer 1.8.0__tar.gz → 1.9.0__tar.gz

{mcp_code_indexer-1.8.0/src/mcp_code_indexer.egg-info → mcp_code_indexer-1.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mcp-code-indexer
-Version: 1.8.0
+Version: 1.9.0
 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

{mcp_code_indexer-1.8.0 → mcp_code_indexer-1.9.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "mcp-code-indexer"
-version = "1.8.0"
+version = "1.9.0"
 description = "MCP server that tracks file descriptions across codebases, enabling AI agents to efficiently navigate and understand code through searchable summaries and token-aware overviews."
 readme = "README.md"
 license = {text = "MIT"}

{mcp_code_indexer-1.8.0 → mcp_code_indexer-1.9.0}/src/mcp_code_indexer/git_hook_handler.py

@@ -119,9 +119,8 @@ class GitHookHandler:
                 self.logger.info(f"Skipping git hook update - no git diff")
                 return
             
-            diff_chars = len(git_diff)
             diff_tokens = self.token_counter.count_tokens(git_diff)
-            self.logger.info(f"Git diff: {diff_chars} characters, {diff_tokens} tokens")
+            self.logger.info(f"Git diff: {diff_tokens} tokens")
             
             # Fetch current state
             self.logger.info("Fetching current project state...")
@@ -134,11 +133,69 @@ class GitHookHandler:
                 return
             
             self.logger.info(f"Found {len(changed_files)} changed files: {', '.join(changed_files)}")
-            self.logger.info(f"Current overview length: {len(current_overview) if current_overview else 0} characters")
+            overview_tokens = self.token_counter.count_tokens(current_overview) if current_overview else 0
+            self.logger.info(f"Current overview: {overview_tokens} tokens")
             self.logger.info(f"Current descriptions count: {len(current_descriptions)}")
             
-            # Use two-stage approach for large codebases
-            self.logger.info("Starting two-stage analysis approach...")
+            # Try single-stage first, fall back to two-stage if needed
+            updates = await self._analyze_with_smart_staging(
+                git_diff, commit_message, current_overview, current_descriptions, changed_files
+            )
+            
+            # Apply updates to database
+            await self._apply_updates(project_info, updates)
+            
+            self.logger.info(f"Git hook update completed successfully for {len(changed_files)} files")
+            
+        except Exception as e:
+            self.logger.error(f"Git hook mode failed: {e}")
+            self.logger.error(f"Exception details: {type(e).__name__}: {str(e)}")
+            import traceback
+            self.logger.error(f"Full traceback:\n{traceback.format_exc()}")
+            # Don't fail the git operation - just log the error
+            raise GitHookError(f"Git hook processing failed: {e}")
+    
+    async def _analyze_with_smart_staging(
+        self,
+        git_diff: str,
+        commit_message: str,
+        current_overview: str,
+        current_descriptions: Dict[str, str],
+        changed_files: List[str]
+    ) -> Dict[str, Any]:
+        """
+        Smart staging: Try single-stage first, fall back to two-stage if token limit exceeded.
+        
+        Args:
+            git_diff: Git diff content
+            commit_message: Commit message explaining the changes
+            current_overview: Current project overview
+            current_descriptions: Current file descriptions
+            changed_files: List of changed file paths
+            
+        Returns:
+            Dict containing file_updates and overview_update
+        """
+        # Build single-stage prompt and check token count
+        single_stage_prompt = self._build_single_stage_prompt(
+            git_diff, commit_message, current_overview, current_descriptions, changed_files
+        )
+        
+        prompt_tokens = self.token_counter.count_tokens(single_stage_prompt)
+        token_limit = self.config.get("max_diff_tokens", 130000)  # Conservative limit under 136k
+        
+        self.logger.info(f"Single-stage prompt: {prompt_tokens} tokens")
+        self.logger.info(f"Token limit: {token_limit}")
+        
+        if prompt_tokens <= token_limit:
+            # Use single-stage approach
+            self.logger.info("Using single-stage analysis (within token limit)")
+            result = await self._call_openrouter(single_stage_prompt)
+            self.logger.info("Single-stage analysis completed")
+            return result
+        else:
+            # Fall back to two-stage approach
+            self.logger.info(f"Single-stage prompt too large ({prompt_tokens} tokens), falling back to two-stage analysis")
             
             # Stage 1: Check if overview needs updating
             overview_updates = await self._analyze_overview_updates(
@@ -156,20 +213,76 @@ class GitHookHandler:
                 "overview_update": overview_updates.get("overview_update")
             }
             
-            self.logger.info(f"Two-stage analysis completed, processing updates...")
-            
-            # Apply updates to database
-            await self._apply_updates(project_info, updates)
-            
-            self.logger.info(f"Git hook update completed successfully for {len(changed_files)} files")
+            self.logger.info("Two-stage analysis completed")
+            return updates
+    
+    def _build_single_stage_prompt(
+        self,
+        git_diff: str,
+        commit_message: str,
+        current_overview: str,
+        current_descriptions: Dict[str, str],
+        changed_files: List[str]
+    ) -> str:
+        """
+        Build single-stage prompt that handles both overview and file updates.
+        
+        Args:
+            git_diff: Git diff content
+            commit_message: Commit message explaining the changes
+            current_overview: Current project overview
+            current_descriptions: Current file descriptions
+            changed_files: List of changed file paths
             
-        except Exception as e:
-            self.logger.error(f"Git hook mode failed: {e}")
-            self.logger.error(f"Exception details: {type(e).__name__}: {str(e)}")
-            import traceback
-            self.logger.error(f"Full traceback:\n{traceback.format_exc()}")
-            # Don't fail the git operation - just log the error
-            raise GitHookError(f"Git hook processing failed: {e}")
+        Returns:
+            Complete single-stage prompt
+        """
+        # 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
+        }
+        
+        return f"""Analyze this git commit and update both the project overview (if needed) and file descriptions.
+
+COMMIT MESSAGE:
+{commit_message or "No commit message available"}
+
+CURRENT PROJECT OVERVIEW:
+{current_overview or "No overview available"}
+
+CURRENT FILE DESCRIPTIONS (for changed files only):
+{json.dumps(relevant_descriptions, indent=2)}
+
+CHANGED FILES:
+{', '.join(changed_files)}
+
+GIT DIFF:
+{git_diff}
+
+INSTRUCTIONS:
+
+1. OVERVIEW UPDATE: 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.
+
+2. FILE UPDATES: 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:
+{{
+  "overview_update": "Updated overview text" or null,
+  "file_updates": {{
+    "path/to/file1.py": "Updated description for file1",
+    "path/to/file2.js": "Updated description for file2"
+  }}
+}}"""
     
     async def _identify_project_from_git(self) -> Dict[str, Any]:
         """
@@ -489,9 +602,8 @@ Return ONLY a JSON object:
 }}"""
 
         # 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")
+        self.logger.info(f"Stage 1 prompt: {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")
@@ -559,9 +671,8 @@ Return ONLY a JSON object:
 }}"""
 
         # 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")
+        self.logger.info(f"Stage 2 prompt: {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")
@@ -667,8 +778,49 @@ Return ONLY a JSON object:
         Returns:
             Validated response data
         """
+        def extract_json_from_response(text: str) -> str:
+            """Extract JSON from response that might have extra text before/after."""
+            text = text.strip()
+            
+            # Try to find JSON in the response
+            json_start = -1
+            json_end = -1
+            
+            # Look for opening brace
+            for i, char in enumerate(text):
+                if char == '{':
+                    json_start = i
+                    break
+            
+            if json_start == -1:
+                return text  # No JSON found, return original
+            
+            # Find matching closing brace
+            brace_count = 0
+            for i in range(json_start, len(text)):
+                if text[i] == '{':
+                    brace_count += 1
+                elif text[i] == '}':
+                    brace_count -= 1
+                    if brace_count == 0:
+                        json_end = i + 1
+                        break
+            
+            if json_end == -1:
+                return text  # No matching brace found, return original
+            
+            return text[json_start:json_end]
+        
         try:
-            data = json.loads(response_text.strip())
+            # First try parsing as-is
+            try:
+                data = json.loads(response_text.strip())
+            except json.JSONDecodeError:
+                # Try extracting JSON from response
+                extracted_json = extract_json_from_response(response_text)
+                if extracted_json != response_text.strip():
+                    self.logger.debug(f"Extracted JSON from response: {extracted_json}")
+                data = json.loads(extracted_json)
             
             # Handle both single-stage and two-stage responses
             if "file_updates" in data and "overview_update" in data:
@@ -700,8 +852,10 @@ Return ONLY a JSON object:
             return data
             
         except json.JSONDecodeError as e:
+            self.logger.error(f"Raw response content: {repr(response_text)}")
             raise GitHookError(f"Invalid JSON response from API: {e}")
         except ValueError as e:
+            self.logger.error(f"Raw response content: {repr(response_text)}")
             raise GitHookError(f"Invalid response structure: {e}")
     
     async def _apply_updates(self, project_info: Dict[str, Any], updates: Dict[str, Any]) -> None:

{mcp_code_indexer-1.8.0 → mcp_code_indexer-1.9.0/src/mcp_code_indexer.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mcp-code-indexer
-Version: 1.8.0
+Version: 1.9.0
 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