specfact-cli 0.4.0__py3-none-any.whl

specfact_cli/analyzers/code_analyzer.py

@@ -0,0 +1,775 @@
+"""Code analyzer for extracting features from brownfield codebases."""
+
+from __future__ import annotations
+
+import ast
+import re
+from collections import defaultdict
+from pathlib import Path
+
+import networkx as nx
+from beartype import beartype
+from icontract import ensure, require
+
+from specfact_cli.models.plan import Feature, Idea, Metadata, PlanBundle, Product, Story
+from specfact_cli.utils.feature_keys import to_classname_key, to_sequential_key
+
+
+class CodeAnalyzer:
+    """
+    Analyzes Python code to auto-derive plan bundles.
+
+    Extracts features from classes and user stories from method patterns
+    following Scrum/Agile practices.
+    """
+
+    # Fibonacci sequence for story points
+    FIBONACCI = [1, 2, 3, 5, 8, 13, 21, 34, 55, 89]
+
+    @beartype
+    @require(lambda repo_path: repo_path is not None and isinstance(repo_path, Path), "Repo path must be Path")
+    @require(lambda confidence_threshold: 0.0 <= confidence_threshold <= 1.0, "Confidence threshold must be 0.0-1.0")
+    @require(lambda plan_name: plan_name is None or isinstance(plan_name, str), "Plan name must be None or str")
+    def __init__(
+        self,
+        repo_path: Path,
+        confidence_threshold: float = 0.5,
+        key_format: str = "classname",
+        plan_name: str | None = None,
+    ) -> None:
+        """
+        Initialize code analyzer.
+
+        Args:
+            repo_path: Path to repository root
+            confidence_threshold: Minimum confidence score (0.0-1.0)
+            key_format: Feature key format ('classname' or 'sequential', default: 'classname')
+            plan_name: Custom plan name (will be used for idea.title, optional)
+        """
+        self.repo_path = Path(repo_path)
+        self.confidence_threshold = confidence_threshold
+        self.key_format = key_format
+        self.plan_name = plan_name
+        self.features: list[Feature] = []
+        self.themes: set[str] = set()
+        self.dependency_graph: nx.DiGraph = nx.DiGraph()  # Module dependency graph
+        self.type_hints: dict[str, dict[str, str]] = {}  # Module -> {function: type_hint}
+        self.async_patterns: dict[str, list[str]] = {}  # Module -> [async_methods]
+        self.commit_bounds: dict[str, tuple[str, str]] = {}  # Feature -> (first_commit, last_commit)
+
+    @beartype
+    @ensure(lambda result: isinstance(result, PlanBundle), "Must return PlanBundle")
+    @ensure(lambda result: result.version == "1.0", "Plan bundle version must be 1.0")
+    @ensure(lambda result: len(result.features) >= 0, "Features list must be non-negative length")
+    def analyze(self) -> PlanBundle:
+        """
+        Analyze repository and generate plan bundle.
+
+        Returns:
+            Generated PlanBundle from code analysis
+        """
+        # Find all Python files
+        python_files = list(self.repo_path.rglob("*.py"))
+
+        # Build module dependency graph first
+        self._build_dependency_graph(python_files)
+
+        # Analyze each file
+        for file_path in python_files:
+            if self._should_skip_file(file_path):
+                continue
+
+            self._analyze_file(file_path)
+
+        # Analyze commit history for feature boundaries
+        self._analyze_commit_history()
+
+        # Enhance features with dependency information
+        self._enhance_features_with_dependencies()
+
+        # If sequential format, update all keys now that we know the total count
+        if self.key_format == "sequential":
+            for idx, feature in enumerate(self.features, start=1):
+                feature.key = to_sequential_key(feature.key, idx)
+
+        # Generate plan bundle
+        # Use plan_name if provided, otherwise use repo name, otherwise fallback
+        if self.plan_name:
+            # Use the plan name (already sanitized, but humanize for title)
+            title = self.plan_name.replace("_", " ").replace("-", " ").title()
+        else:
+            repo_name = self.repo_path.name or "Unknown Project"
+            title = self._humanize_name(repo_name)
+        
+        idea = Idea(
+            title=title,
+            narrative=f"Auto-derived plan from brownfield analysis of {title}",
+            metrics=None,
+        )
+
+        product = Product(
+            themes=sorted(self.themes) if self.themes else ["Core"],
+            releases=[],
+        )
+
+        return PlanBundle(
+            version="1.0",
+            idea=idea,
+            business=None,
+            product=product,
+            features=self.features,
+            metadata=Metadata(stage="draft", promoted_at=None, promoted_by=None),
+        )
+
+    def _should_skip_file(self, file_path: Path) -> bool:
+        """Check if file should be skipped."""
+        skip_patterns = [
+            "__pycache__",
+            ".git",
+            "venv",
+            ".venv",
+            "env",
+            ".pytest_cache",
+            "htmlcov",
+            "dist",
+            "build",
+            ".eggs",
+            "tests",  # Skip test files
+        ]
+
+        return any(pattern in str(file_path) for pattern in skip_patterns)
+
+    def _analyze_file(self, file_path: Path) -> None:
+        """Analyze a single Python file."""
+        try:
+            content = file_path.read_text(encoding="utf-8")
+            tree = ast.parse(content)
+
+            # Extract module-level info
+            self._extract_themes_from_imports(tree)
+
+            # Extract type hints
+            self._extract_type_hints(tree, file_path)
+
+            # Detect async patterns
+            self._detect_async_patterns(tree, file_path)
+
+            # Extract classes as features
+            for node in ast.walk(tree):
+                if isinstance(node, ast.ClassDef):
+                    feature = self._extract_feature_from_class(node, file_path)
+                    if feature:
+                        self.features.append(feature)
+
+        except (SyntaxError, UnicodeDecodeError):
+            # Skip files that can't be parsed
+            pass
+
+    def _extract_themes_from_imports(self, tree: ast.AST) -> None:
+        """Extract themes from import statements."""
+        theme_keywords = {
+            "fastapi": "API",
+            "flask": "API",
+            "django": "Web",
+            "redis": "Caching",
+            "postgres": "Database",
+            "mysql": "Database",
+            "asyncio": "Async",
+            "typer": "CLI",
+            "click": "CLI",
+            "pydantic": "Validation",
+            "pytest": "Testing",
+            "sqlalchemy": "ORM",
+            "requests": "HTTP Client",
+            "aiohttp": "Async HTTP",
+        }
+
+        for node in ast.walk(tree):
+            if isinstance(node, (ast.Import, ast.ImportFrom)):
+                if isinstance(node, ast.Import):
+                    for alias in node.names:
+                        for keyword, theme in theme_keywords.items():
+                            if keyword in alias.name.lower():
+                                self.themes.add(theme)
+                elif isinstance(node, ast.ImportFrom) and node.module:
+                    for keyword, theme in theme_keywords.items():
+                        if keyword in node.module.lower():
+                            self.themes.add(theme)
+
+    def _extract_feature_from_class(self, node: ast.ClassDef, file_path: Path) -> Feature | None:
+        """Extract feature from class definition."""
+        # Skip private classes and test classes
+        if node.name.startswith("_") or node.name.startswith("Test"):
+            return None
+
+        # Generate feature key based on configured format
+        if self.key_format == "sequential":
+            # Use sequential numbering (will be updated after all features are collected)
+            feature_key = f"FEATURE-{len(self.features) + 1:03d}"
+        else:
+            # Default: classname format
+            feature_key = to_classname_key(node.name)
+
+        # Extract docstring as outcome
+        docstring = ast.get_docstring(node)
+        outcomes = []
+        if docstring:
+            # Take first paragraph as primary outcome
+            first_para = docstring.split("\n\n")[0].strip()
+            outcomes.append(first_para)
+        else:
+            outcomes.append(f"Provides {self._humanize_name(node.name)} functionality")
+
+        # Collect all methods
+        methods = [item for item in node.body if isinstance(item, ast.FunctionDef)]
+
+        # Group methods into user stories
+        stories = self._extract_stories_from_methods(methods, node.name)
+
+        # Calculate confidence based on documentation and story quality
+        confidence = self._calculate_feature_confidence(node, stories)
+
+        if confidence < self.confidence_threshold:
+            return None
+
+        # Skip if no meaningful stories
+        if not stories:
+            return None
+
+        return Feature(
+            key=feature_key,
+            title=self._humanize_name(node.name),
+            outcomes=outcomes,
+            acceptance=[f"{node.name} class provides documented functionality"],
+            stories=stories,
+            confidence=round(confidence, 2),
+        )
+
+    def _extract_stories_from_methods(self, methods: list[ast.FunctionDef], class_name: str) -> list[Story]:
+        """
+        Extract user stories from methods by grouping related functionality.
+
+        Groups methods by:
+        - CRUD operations (create, read, update, delete)
+        - Common prefixes (get_, set_, validate_, process_)
+        - Functionality patterns
+        """
+        # Group methods by pattern
+        method_groups = self._group_methods_by_functionality(methods)
+
+        stories = []
+        story_counter = 1
+
+        for group_name, group_methods in method_groups.items():
+            if not group_methods:
+                continue
+
+            # Create a user story for this group
+            story = self._create_story_from_method_group(group_name, group_methods, class_name, story_counter)
+
+            if story:
+                stories.append(story)
+                story_counter += 1
+
+        return stories
+
+    def _group_methods_by_functionality(self, methods: list[ast.FunctionDef]) -> dict[str, list[ast.FunctionDef]]:
+        """Group methods by their functionality patterns."""
+        groups = defaultdict(list)
+
+        # Filter out private methods (except __init__)
+        public_methods = [m for m in methods if not m.name.startswith("_") or m.name == "__init__"]
+
+        for method in public_methods:
+            # CRUD operations
+            if any(crud in method.name.lower() for crud in ["create", "add", "insert", "new"]):
+                groups["Create Operations"].append(method)
+            elif any(read in method.name.lower() for read in ["get", "read", "fetch", "find", "list", "retrieve"]):
+                groups["Read Operations"].append(method)
+            elif any(update in method.name.lower() for update in ["update", "modify", "edit", "change", "set"]):
+                groups["Update Operations"].append(method)
+            elif any(delete in method.name.lower() for delete in ["delete", "remove", "destroy"]):
+                groups["Delete Operations"].append(method)
+
+            # Validation
+            elif any(val in method.name.lower() for val in ["validate", "check", "verify", "is_valid"]):
+                groups["Validation"].append(method)
+
+            # Processing/Computation
+            elif any(
+                proc in method.name.lower() for proc in ["process", "compute", "calculate", "transform", "convert"]
+            ):
+                groups["Processing"].append(method)
+
+            # Analysis
+            elif any(an in method.name.lower() for an in ["analyze", "parse", "extract", "detect"]):
+                groups["Analysis"].append(method)
+
+            # Generation
+            elif any(gen in method.name.lower() for gen in ["generate", "build", "create", "make"]):
+                groups["Generation"].append(method)
+
+            # Comparison
+            elif any(cmp in method.name.lower() for cmp in ["compare", "diff", "match"]):
+                groups["Comparison"].append(method)
+
+            # Setup/Configuration
+            elif method.name == "__init__" or any(
+                setup in method.name.lower() for setup in ["setup", "configure", "initialize"]
+            ):
+                groups["Configuration"].append(method)
+
+            # Catch-all for other public methods
+            else:
+                groups["Core Functionality"].append(method)
+
+        return dict(groups)
+
+    def _create_story_from_method_group(
+        self, group_name: str, methods: list[ast.FunctionDef], class_name: str, story_number: int
+    ) -> Story | None:
+        """Create a user story from a group of related methods."""
+        if not methods:
+            return None
+
+        # Generate story key
+        story_key = f"STORY-{class_name.upper()}-{story_number:03d}"
+
+        # Create user-centric title based on group
+        title = self._generate_story_title(group_name, class_name)
+
+        # Extract acceptance criteria from docstrings
+        acceptance = []
+        tasks = []
+
+        for method in methods:
+            # Add method as task
+            tasks.append(f"{method.name}()")
+
+            # Extract acceptance from docstring
+            docstring = ast.get_docstring(method)
+            if docstring:
+                # Take first line as acceptance criterion
+                first_line = docstring.split("\n")[0].strip()
+                if first_line and first_line not in acceptance:
+                    acceptance.append(first_line)
+
+        # Add default acceptance if none found
+        if not acceptance:
+            acceptance.append(f"{group_name} functionality works as expected")
+
+        # Calculate story points (complexity) based on number of methods and their size
+        story_points = self._calculate_story_points(methods)
+
+        # Calculate value points based on public API exposure
+        value_points = self._calculate_value_points(methods, group_name)
+
+        return Story(
+            key=story_key,
+            title=title,
+            acceptance=acceptance,
+            story_points=story_points,
+            value_points=value_points,
+            tasks=tasks,
+            confidence=0.8 if len(methods) > 1 else 0.6,
+        )
+
+    def _generate_story_title(self, group_name: str, class_name: str) -> str:
+        """Generate user-centric story title."""
+        # Map group names to user-centric titles
+        title_templates = {
+            "Create Operations": f"As a user, I can create new {self._humanize_name(class_name)} records",
+            "Read Operations": f"As a user, I can view {self._humanize_name(class_name)} data",
+            "Update Operations": f"As a user, I can update {self._humanize_name(class_name)} records",
+            "Delete Operations": f"As a user, I can delete {self._humanize_name(class_name)} records",
+            "Validation": f"As a developer, I can validate {self._humanize_name(class_name)} data",
+            "Processing": f"As a user, I can process data using {self._humanize_name(class_name)}",
+            "Analysis": f"As a user, I can analyze data with {self._humanize_name(class_name)}",
+            "Generation": f"As a user, I can generate outputs from {self._humanize_name(class_name)}",
+            "Comparison": f"As a user, I can compare {self._humanize_name(class_name)} data",
+            "Configuration": f"As a developer, I can configure {self._humanize_name(class_name)}",
+            "Core Functionality": f"As a user, I can use {self._humanize_name(class_name)} features",
+        }
+
+        return title_templates.get(group_name, f"As a user, I can work with {self._humanize_name(class_name)}")
+
+    def _calculate_story_points(self, methods: list[ast.FunctionDef]) -> int:
+        """
+        Calculate story points (complexity) using Fibonacci sequence.
+
+        Based on:
+        - Number of methods
+        - Average method size
+        - Complexity indicators (loops, conditionals)
+        """
+        # Base complexity on number of methods
+        method_count = len(methods)
+
+        # Count total lines across all methods
+        total_lines = sum(len(ast.unparse(m).split("\n")) for m in methods)
+        avg_lines = total_lines / method_count if method_count > 0 else 0
+
+        # Simple heuristic: 1-2 methods = small, 3-5 = medium, 6+ = large
+        if method_count <= 2 and avg_lines < 20:
+            base_points = 2  # Small
+        elif method_count <= 5 and avg_lines < 40:
+            base_points = 5  # Medium
+        elif method_count <= 8:
+            base_points = 8  # Large
+        else:
+            base_points = 13  # Extra Large
+
+        # Return nearest Fibonacci number
+        return min(self.FIBONACCI, key=lambda x: abs(x - base_points))
+
+    def _calculate_value_points(self, methods: list[ast.FunctionDef], group_name: str) -> int:
+        """
+        Calculate value points (business value) using Fibonacci sequence.
+
+        Based on:
+        - Public API exposure
+        - CRUD operations have high value
+        - Validation has medium value
+        """
+        # CRUD operations are high value
+        crud_groups = ["Create Operations", "Read Operations", "Update Operations", "Delete Operations"]
+        if group_name in crud_groups:
+            base_value = 8  # High business value
+
+        # User-facing operations
+        elif group_name in ["Processing", "Analysis", "Generation", "Comparison"]:
+            base_value = 5  # Medium-high value
+
+        # Developer/internal operations
+        elif group_name in ["Validation", "Configuration"]:
+            base_value = 3  # Medium value
+
+        # Core functionality
+        else:
+            base_value = 3  # Default medium value
+
+        # Adjust based on number of public methods (more = higher value)
+        public_count = sum(1 for m in methods if not m.name.startswith("_"))
+        if public_count >= 3:
+            base_value = min(base_value + 2, 13)
+
+        # Return nearest Fibonacci number
+        return min(self.FIBONACCI, key=lambda x: abs(x - base_value))
+
+    def _calculate_feature_confidence(self, node: ast.ClassDef, stories: list[Story]) -> float:
+        """Calculate confidence score for a feature."""
+        score = 0.3  # Base score
+
+        # Has docstring
+        if ast.get_docstring(node):
+            score += 0.2
+
+        # Has stories
+        if stories:
+            score += 0.2
+
+        # Has multiple stories (better coverage)
+        if len(stories) > 2:
+            score += 0.2
+
+        # Stories are well-documented
+        documented_stories = sum(1 for s in stories if s.acceptance and len(s.acceptance) > 1)
+        if stories and documented_stories > len(stories) / 2:
+            score += 0.1
+
+        return min(score, 1.0)
+
+    def _humanize_name(self, name: str) -> str:
+        """Convert snake_case or PascalCase to human-readable title."""
+        # Handle PascalCase
+        name = re.sub(r"([A-Z])", r" \1", name).strip()
+        # Handle snake_case
+        name = name.replace("_", " ").replace("-", " ")
+        return name.title()
+
+    def _build_dependency_graph(self, python_files: list[Path]) -> None:
+        """
+        Build module dependency graph using AST imports.
+
+        Creates a directed graph where nodes are modules and edges represent imports.
+        """
+        # First pass: collect all modules as nodes
+        modules: dict[str, Path] = {}
+        for file_path in python_files:
+            if self._should_skip_file(file_path):
+                continue
+
+            # Convert file path to module name
+            module_name = self._path_to_module_name(file_path)
+            modules[module_name] = file_path
+            self.dependency_graph.add_node(module_name, path=file_path)
+
+        # Second pass: add edges based on imports
+        for module_name, file_path in modules.items():
+            try:
+                content = file_path.read_text(encoding="utf-8")
+                tree = ast.parse(content)
+
+                # Extract imports
+                imports = self._extract_imports_from_ast(tree, file_path)
+                for imported_module in imports:
+                    # Only add edges for modules we know about (within repo)
+                    # Try exact match first, then partial match
+                    if imported_module in modules:
+                        self.dependency_graph.add_edge(module_name, imported_module)
+                    else:
+                        # Try to find matching module (e.g., "module_a" matches "src.module_a")
+                        matching_module = None
+                        for known_module in modules:
+                            # Check if imported name matches the module name (last part)
+                            if imported_module == known_module.split(".")[-1]:
+                                matching_module = known_module
+                                break
+                        if matching_module:
+                            self.dependency_graph.add_edge(module_name, matching_module)
+            except (SyntaxError, UnicodeDecodeError):
+                # Skip files that can't be parsed
+                continue
+
+    def _path_to_module_name(self, file_path: Path) -> str:
+        """Convert file path to module name (e.g., src/foo/bar.py -> src.foo.bar)."""
+        # Get relative path from repo root
+        try:
+            relative_path = file_path.relative_to(self.repo_path)
+        except ValueError:
+            # File is outside repo, use full path
+            relative_path = file_path
+
+        # Convert to module name
+        parts = list(relative_path.parts[:-1]) + [relative_path.stem]  # Remove .py extension
+        return ".".join(parts)
+
+    def _extract_imports_from_ast(self, tree: ast.AST, file_path: Path) -> list[str]:
+        """
+        Extract imported module names from AST.
+
+        Returns:
+            List of module names (relative to repo root if possible)
+        """
+        imports: set[str] = set()
+
+        for node in ast.walk(tree):
+            if isinstance(node, ast.Import):
+                for alias in node.names:
+                    # Import aliases (e.g., import foo as bar)
+                    if "." in alias.name:
+                        # Extract root module (e.g., foo.bar.baz -> foo)
+                        root_module = alias.name.split(".")[0]
+                        imports.add(root_module)
+                    else:
+                        imports.add(alias.name)
+
+            elif isinstance(node, ast.ImportFrom) and node.module:
+                # From imports (e.g., from foo.bar import baz)
+                if "." in node.module:
+                    # Extract root module
+                    root_module = node.module.split(".")[0]
+                    imports.add(root_module)
+                else:
+                    imports.add(node.module)
+
+        # Try to resolve local imports (relative to current file)
+        resolved_imports = []
+        current_module = self._path_to_module_name(file_path)
+
+        for imported in imports:
+            # Skip stdlib imports (common patterns)
+            stdlib_modules = {
+                "sys",
+                "os",
+                "json",
+                "yaml",
+                "pathlib",
+                "typing",
+                "collections",
+                "dataclasses",
+                "enum",
+                "abc",
+                "asyncio",
+                "functools",
+                "itertools",
+                "re",
+                "datetime",
+                "time",
+                "logging",
+                "hashlib",
+                "base64",
+                "urllib",
+                "http",
+                "socket",
+                "threading",
+                "multiprocessing",
+            }
+
+            if imported in stdlib_modules:
+                continue
+
+            # Try to resolve relative imports
+            # If imported module matches a pattern from our repo, resolve it
+            potential_module = self._resolve_local_import(imported, current_module)
+            if potential_module:
+                resolved_imports.append(potential_module)
+            else:
+                # Keep as external dependency
+                resolved_imports.append(imported)
+
+        return resolved_imports
+
+    def _resolve_local_import(self, imported: str, current_module: str) -> str | None:
+        """
+        Try to resolve a local import relative to current module.
+
+        Returns:
+            Resolved module name if found in repo, None otherwise
+        """
+        # Check if it's already in our dependency graph
+        if imported in self.dependency_graph:
+            return imported
+
+        # Try relative import resolution (e.g., from .foo import bar)
+        # This is simplified - full resolution would need to handle package structure
+        current_parts = current_module.split(".")
+        if len(current_parts) > 1:
+            # Try parent package
+            parent_module = ".".join(current_parts[:-1])
+            potential = f"{parent_module}.{imported}"
+            if potential in self.dependency_graph:
+                return potential
+
+        return None
+
+    def _extract_type_hints(self, tree: ast.AST, file_path: Path) -> dict[str, str]:
+        """
+        Extract type hints from function/method signatures.
+
+        Returns:
+            Dictionary mapping function names to their return type hints
+        """
+        type_hints: dict[str, str] = {}
+        module_name = self._path_to_module_name(file_path)
+
+        for node in ast.walk(tree):
+            if isinstance(node, (ast.FunctionDef, ast.AsyncFunctionDef)):
+                func_name = node.name
+                return_type = "None"
+
+                # Extract return type annotation
+                if node.returns:
+                    # Convert AST node to string representation
+                    if isinstance(node.returns, ast.Name):
+                        return_type = node.returns.id
+                    elif isinstance(node.returns, ast.Subscript):
+                        # Handle generics like List[str], Dict[str, int]
+                        container = node.returns.value.id if isinstance(node.returns.value, ast.Name) else "Any"
+                        return_type = str(container)  # Simplified representation
+
+                type_hints[func_name] = return_type
+
+        # Store per module
+        if module_name not in self.type_hints:
+            self.type_hints[module_name] = {}
+        self.type_hints[module_name].update(type_hints)
+
+        return type_hints
+
+    def _detect_async_patterns(self, tree: ast.AST, file_path: Path) -> list[str]:
+        """
+        Detect async/await patterns in code.
+
+        Returns:
+            List of async method/function names
+        """
+        async_methods: list[str] = []
+        module_name = self._path_to_module_name(file_path)
+
+        for node in ast.walk(tree):
+            # Check for async functions
+            if isinstance(node, ast.AsyncFunctionDef):
+                async_methods.append(node.name)
+
+            # Check for await statements (even in sync functions)
+            if isinstance(node, ast.Await):
+                # Find containing function
+                for parent in ast.walk(tree):
+                    if isinstance(parent, (ast.FunctionDef, ast.AsyncFunctionDef)):
+                        for child in ast.walk(parent):
+                            if child == node:
+                                if parent.name not in async_methods:
+                                    async_methods.append(parent.name)
+                                break
+
+        # Store per module
+        self.async_patterns[module_name] = async_methods
+
+        return async_methods
+
+    def _analyze_commit_history(self) -> None:
+        """
+        Mine commit history to identify feature boundaries.
+
+        Uses GitPython to analyze commit messages and associate them with features.
+        Limits analysis to recent commits to avoid performance issues.
+        """
+        try:
+            from git import Repo
+
+            if not (self.repo_path / ".git").exists():
+                return
+
+            repo = Repo(self.repo_path)
+            # Limit to last 100 commits to avoid performance issues with large repositories
+            max_commits = 100
+            commits = list(repo.iter_commits(max_count=max_commits))
+
+            # Map commits to files to features
+            file_to_feature: dict[str, list[str]] = {}
+            for feature in self.features:
+                # Extract potential file paths from feature key
+                # This is simplified - in reality we'd track which files contributed to which features
+                pass
+
+            # Analyze commit messages for feature references
+            for commit in commits:
+                try:
+                    # Skip commits that can't be accessed (corrupted or too old)
+                    # Use commit.message which is lazy-loaded but faster than full commit object
+                    commit_message = commit.message
+                    if isinstance(commit_message, bytes):
+                        commit_message = commit_message.decode("utf-8", errors="ignore")
+                    message = commit_message.lower()
+                    # Look for feature patterns (e.g., FEATURE-001, feat:, feature:)
+                    if "feat" in message or "feature" in message:
+                        # Try to extract feature keys from commit message
+                        feature_match = re.search(r"feature[-\s]?(\d+)", message, re.IGNORECASE)
+                        if feature_match:
+                            feature_num = feature_match.group(1)
+                            # Associate commit with feature (simplified)
+                except Exception:
+                    # Skip individual commits that fail (corrupted, etc.)
+                    continue
+
+        except ImportError:
+            # GitPython not available, skip
+            pass
+        except Exception:
+            # Git operations failed, skip gracefully
+            pass
+
+    def _enhance_features_with_dependencies(self) -> None:
+        """Enhance features with dependency graph information."""
+        for feature in self.features:
+            # Find dependencies for this feature's module
+            # This is simplified - would need to track which module each feature comes from
+            pass
+
+    def _get_module_dependencies(self, module_name: str) -> list[str]:
+        """Get list of modules that the given module depends on."""
+        if module_name not in self.dependency_graph:
+            return []
+
+        return list(self.dependency_graph.successors(module_name))