code-finder 0.1.0__py3-none-any.whl

claude_context/config.py

@@ -0,0 +1,239 @@
+"""
+Balanced Claude Context Configuration - Practical defaults with fail-fast principles
+
+This version keeps useful file extensions while removing over-engineering
+and silent fallbacks.
+"""
+
+import logging
+from dataclasses import dataclass, field
+from pathlib import Path
+from typing import List, Optional
+from pymilvus import MilvusClient
+
+# Clear, simple logging
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+)
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class ClaudeContextConfig:
+    """Configuration for Claude Context - practical defaults, no speculation"""
+    
+    # Database
+    db_path: str = "./data/milvus_vibe2doc.db"
+    collection_name: str = "code_chunks"
+    
+    # Embeddings - start with fast, local model
+    # This is for DEVELOPMENT - easily upgradeable to better models:
+    # - "all-mpnet-base-v2" (better quality, 420MB)
+    # - "BAAI/bge-large-en-v1.5" (state-of-art, 1.3GB)
+    # - OpenAI API "text-embedding-3-large" (best, requires API key)
+    # - Voyage AI "voyage-code-2" (optimized for code, requires API key)
+    embedding_model: str = "all-MiniLM-L6-v2"
+    
+    # Chunking - reasonable defaults based on experience
+    chunk_size: int = 1500
+    chunk_overlap: int = 200
+    
+    # Search - simple for now
+    max_results: int = 10
+    
+    # File patterns - comprehensive but practical
+    file_extensions: List[str] = field(default_factory=list)
+    ignore_patterns: List[str] = field(default_factory=list)
+    
+    def __post_init__(self):
+        """Initialize with practical defaults"""
+        
+        # Include common programming languages and config files
+        # These are languages we're likely to encounter in real projects
+        if not self.file_extensions:
+            self.file_extensions = [
+                # Core languages (most common)
+                ".py",    # Python
+                ".js",    # JavaScript
+                ".ts",    # TypeScript
+                ".jsx",   # React
+                ".tsx",   # React TypeScript
+                
+                # Other common languages
+                ".java",  # Java
+                ".cpp",   # C++
+                ".c",     # C
+                ".h",     # Headers
+                ".hpp",   # C++ headers
+                ".go",    # Go
+                ".rs",    # Rust
+                ".rb",    # Ruby
+                ".php",   # PHP
+                ".cs",    # C#
+                
+                # Documentation
+                ".md",    # Markdown
+                ".rst",   # reStructuredText
+                ".txt",   # Plain text
+                
+                # Configuration
+                ".json",  # JSON
+                ".yaml",  # YAML
+                ".yml",   # YAML
+                ".toml",  # TOML
+                ".ini",   # INI
+                ".env",   # Environment
+                ".xml",   # XML
+                
+                # Web
+                ".html",  # HTML
+                ".css",   # CSS
+                
+                # Shell/Scripts
+                ".sh",    # Shell
+                ".bash",  # Bash
+                ".ps1",   # PowerShell
+                ".bat",   # Batch
+                
+                # Data/Query
+                ".sql",   # SQL
+                ".graphql", # GraphQL
+            ]
+            # Note: We can easily add more as needed, but these cover 99% of projects
+        
+        # Only ignore things that definitely shouldn't be indexed
+        if not self.ignore_patterns:
+            self.ignore_patterns = [
+                # Version control
+                "**/.git/**",
+                
+                # Dependencies (these are huge and not your code)
+                "**/node_modules/**",
+                "**/venv/**",
+                "**/venv_*/**",
+                "**/__pycache__/**",
+                
+                # Build outputs
+                "**/dist/**",
+                "**/build/**",
+
+                # Generated/archived content
+                "**/archive/**",
+                "**/docs/generated/**",
+                "**/docs/planning/**",
+                "**/evaluation/results/**",
+                "**/.vibe2doc/**",
+                "**/vendor/**",
+
+                # OS files
+                "**/.DS_Store",
+                
+                # Binary files
+                "**/*.pyc",
+                "**/*.pyo",
+                "**/*.so",
+                "**/*.dylib",
+                "**/*.dll",
+                "**/*.exe",
+            ]
+        
+        # Create data directory - FAIL if we can't
+        data_dir = Path(self.db_path).parent
+        try:
+            data_dir.mkdir(parents=True, exist_ok=True)
+            logger.info(f"Data directory ready: {data_dir.absolute()}")
+        except Exception as e:
+            # FAIL FAST - if we can't create the data dir, something is wrong
+            raise RuntimeError(f"Cannot create data directory {data_dir}: {e}")
+        
+        # Validate configuration - FAIL FAST on bad config
+        self._validate()
+    
+    def _validate(self):
+        """Validate configuration - fail fast on problems"""
+        if self.chunk_size <= 0:
+            raise ValueError(f"chunk_size must be positive, got {self.chunk_size}")
+        
+        if self.chunk_overlap >= self.chunk_size:
+            raise ValueError(f"chunk_overlap ({self.chunk_overlap}) must be less than chunk_size ({self.chunk_size})")
+        
+        if self.max_results <= 0:
+            raise ValueError(f"max_results must be positive, got {self.max_results}")
+        
+        if not self.file_extensions:
+            raise ValueError("file_extensions cannot be empty")
+        
+        logger.debug("Configuration validated")
+    
+    def add_extension(self, ext: str):
+        """Add a file extension at runtime if needed"""
+        if not ext.startswith('.'):
+            ext = f'.{ext}'
+        if ext not in self.file_extensions:
+            self.file_extensions.append(ext)
+            logger.info(f"Added file extension: {ext}")
+    
+    def add_ignore_pattern(self, pattern: str):
+        """Add an ignore pattern at runtime if needed"""
+        if pattern not in self.ignore_patterns:
+            self.ignore_patterns.append(pattern)
+            logger.info(f"Added ignore pattern: {pattern}")
+
+
+class MilvusManager:
+    """Simple Milvus client manager - fail fast, no silent errors"""
+    
+    def __init__(self, config: ClaudeContextConfig):
+        """Initialize manager - will fail fast if config is invalid"""
+        self.config = config
+        self.client: Optional[MilvusClient] = None
+        logger.info(f"MilvusManager initialized with: {config.db_path}")
+    
+    def get_client(self) -> MilvusClient:
+        """Get or create client - fails fast if connection fails"""
+        if not self.client:
+            logger.info(f"Creating Milvus client: {self.config.db_path}")
+            
+            # No try/except - if this fails, we want to know
+            self.client = MilvusClient(self.config.db_path)
+            
+            # Verify it works - fail fast if not
+            collections = self.client.list_collections()
+            logger.info(f"✅ Connected. Collections: {collections or 'none'}")
+        
+        return self.client
+    
+    def __enter__(self) -> MilvusClient:
+        """Context manager entry"""
+        return self.get_client()
+    
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        """Context manager exit - don't hide exceptions"""
+        if exc_type:
+            logger.error(f"ERROR in MilvusManager: {exc_type.__name__}: {exc_val}")
+            # Don't suppress the exception
+            return False
+        logger.debug("MilvusManager context closed cleanly")
+
+
+# Example usage showing the balance
+if __name__ == "__main__":
+    print("Testing balanced configuration...")
+    
+    # Create config with good defaults
+    config = ClaudeContextConfig()
+    print(f"✅ Config created with {len(config.file_extensions)} useful extensions")
+    print(f"   Covers: Python, JS/TS, Java, Go, Rust, Ruby, configs, docs, etc.")
+    
+    # Show we can easily extend at runtime
+    config.add_extension(".vue")  # Add Vue files when needed
+    config.add_extension("swift")  # Add Swift when we encounter iOS code
+    
+    # Test connection - will fail fast if problem
+    with MilvusManager(config) as client:
+        print("✅ Connected to Milvus")
+        collections = client.list_collections()
+        print(f"   Collections: {collections or 'none'}")
+    
+    print("\n✨ Balanced config: practical defaults + expandability + fail fast!")