shotgun-sh 0.2.29.dev2__py3-none-any.whl → 0.6.1.dev1__py3-none-any.whl

shotgun/codebase/core/kuzu_compat.py

@@ -0,0 +1,119 @@
+"""Lazy import wrapper for kuzu (real_ladybug) with Windows compatibility handling."""
+
+from __future__ import annotations
+
+import sys
+import webbrowser
+from typing import Any
+
+_kuzu_module: Any = None
+_import_error: ImportError | None = None
+
+# Windows VC++ download URL
+_VC_REDIST_URL = "https://aka.ms/vs/17/release/vc_redist.x64.exe"
+
+# PowerShell installation script for VC++ Redistributable
+_VC_INSTALL_SCRIPT = f"""\
+# Download and install Visual C++ Redistributable (run as Administrator)
+Import-Module BitsTransfer
+Start-BitsTransfer -Source "{_VC_REDIST_URL}" -Destination "$env:TEMP\\vc_redist.x64.exe"
+Start-Process -FilePath "$env:TEMP\\vc_redist.x64.exe" -ArgumentList "/install", "/quiet", "/norestart" -Wait\
+"""
+
+# Windows VC++ installation instructions
+_WINDOWS_INSTALL_INSTRUCTIONS = f"""
+To fix this, install the Visual C++ Redistributable.
+
+Option 1: Run this PowerShell script (as Administrator):
+
+{_VC_INSTALL_SCRIPT}
+
+Option 2: Download manually from:
+  {_VC_REDIST_URL}
+"""
+
+
+def copy_vcpp_script_to_clipboard() -> bool:
+    """Copy the VC++ installation PowerShell script to clipboard.
+
+    Returns:
+        True if successful, False otherwise
+    """
+    try:
+        import pyperclip  # type: ignore[import-untyped]
+
+        pyperclip.copy(_VC_INSTALL_SCRIPT)
+        return True
+    except Exception:
+        return False
+
+
+def open_vcpp_download_page() -> bool:
+    """Open the VC++ Redistributable download page in the default browser.
+
+    Returns:
+        True if successful, False otherwise
+    """
+    try:
+        webbrowser.open(_VC_REDIST_URL)
+        return True
+    except Exception:
+        return False
+
+
+class KuzuImportError(ImportError):
+    """Raised when kuzu cannot be imported, typically on Windows due to DLL issues."""
+
+    def __init__(self, original_error: ImportError) -> None:
+        self.original_error = original_error
+        if sys.platform == "win32":
+            message = (
+                "Failed to load the graph database library (real_ladybug).\n\n"
+                "This error typically occurs on Windows when the Visual C++ "
+                "Redistributable is not installed.\n"
+                f"{_WINDOWS_INSTALL_INSTRUCTIONS}\n"
+                f"Original error: {original_error}"
+            )
+        else:
+            message = f"Failed to import real_ladybug: {original_error}"
+        super().__init__(message)
+
+
+def get_kuzu() -> Any:
+    """Get the kuzu module, importing it lazily on first use.
+
+    Raises:
+        KuzuImportError: If kuzu cannot be imported (e.g., Windows DLL issues)
+
+    Returns:
+        The real_ladybug module
+    """
+    global _kuzu_module, _import_error
+
+    if _kuzu_module is not None:
+        return _kuzu_module
+
+    if _import_error is not None:
+        raise KuzuImportError(_import_error)
+
+    try:
+        import real_ladybug
+
+        _kuzu_module = real_ladybug
+        return _kuzu_module
+    except ImportError as e:
+        _import_error = e
+        raise KuzuImportError(e) from e
+
+
+def is_kuzu_available() -> bool:
+    """Check if kuzu is available without raising an error.
+
+    Returns:
+        True if kuzu can be imported, False otherwise
+    """
+    try:
+        get_kuzu()
+        return True
+    except KuzuImportError:
+        return False

shotgun/codebase/core/language_config.py

@@ -1,6 +1,152 @@
 """Language configuration for Tree-sitter parsing."""
 
 from dataclasses import dataclass, field
+from pathlib import Path
+
+# =============================================================================
+# COMMON EXCLUSIONS (applied to all languages)
+# =============================================================================
+
+# Version control
+VCS_DIRECTORIES = {
+    ".git",
+    ".hg",
+    ".svn",
+}
+
+# Editor and IDE directories
+EDITOR_DIRECTORIES = {
+    ".idea",
+    ".vscode",
+    ".vs",
+    ".eclipse",
+    ".settings",
+}
+
+# AI assistant directories
+AI_ASSISTANT_DIRECTORIES = {
+    ".claude",
+    ".cursor",
+    ".copilot",
+}
+
+# Generic build/output directories (language-agnostic)
+GENERIC_BUILD_DIRECTORIES = {
+    "build",
+    "dist",
+    "out",
+    "tmp",
+    "temp",
+    "coverage",
+}
+
+# All common directories to ignore (union of above)
+COMMON_IGNORE_DIRECTORIES = (
+    VCS_DIRECTORIES
+    | EDITOR_DIRECTORIES
+    | AI_ASSISTANT_DIRECTORIES
+    | GENERIC_BUILD_DIRECTORIES
+)
+
+
+# =============================================================================
+# LANGUAGE-SPECIFIC EXCLUSIONS
+# =============================================================================
+
+PYTHON_IGNORE_DIRECTORIES = {
+    # Virtual environments
+    "venv",
+    ".venv",
+    "env",
+    ".env",
+    "virtualenv",
+    ".virtualenv",
+    # Package/build artifacts
+    "__pycache__",
+    ".eggs",
+    "*.egg-info",
+    ".tox",
+    ".nox",
+    # Tool caches
+    ".pytest_cache",
+    ".mypy_cache",
+    ".ruff_cache",
+    ".coverage",
+    ".hypothesis",
+    # Build outputs
+    "site-packages",
+}
+
+JAVASCRIPT_IGNORE_DIRECTORIES = {
+    # Dependencies
+    "node_modules",
+    "bower_components",
+    # Framework build outputs
+    ".next",
+    ".nuxt",
+    ".vite",
+    ".svelte-kit",
+    ".output",
+    # Package managers
+    ".yarn",
+    ".pnpm-store",
+    # Build tools
+    ".turbo",
+    ".parcel-cache",
+    ".cache",
+    # Deployment
+    ".vercel",
+    ".netlify",
+    ".serverless",
+}
+
+TYPESCRIPT_IGNORE_DIRECTORIES = JAVASCRIPT_IGNORE_DIRECTORIES | {
+    # TypeScript-specific
+    ".tsbuildinfo",
+}
+
+GO_IGNORE_DIRECTORIES = {
+    # Vendored dependencies
+    "vendor",
+    # Build output (common convention)
+    "bin",
+}
+
+RUST_IGNORE_DIRECTORIES = {
+    # Cargo build output
+    "target",
+}
+
+RUBY_IGNORE_DIRECTORIES = {
+    # Bundler
+    ".bundle",
+    "vendor/bundle",
+}
+
+JAVA_IGNORE_DIRECTORIES = {
+    # Gradle
+    ".gradle",
+    "gradle",
+    # Maven
+    ".m2",
+    "target",
+    # IDE
+    ".apt_generated",
+}
+
+CSHARP_IGNORE_DIRECTORIES = {
+    # Build outputs
+    "bin",
+    "obj",
+    # NuGet
+    "packages",
+    ".nuget",
+}
+
+PHP_IGNORE_DIRECTORIES = {
+    # Composer
+    "vendor",
+}
 
 
 @dataclass
@@ -17,6 +163,7 @@ class LanguageConfig:
     import_node_types: list[str] = field(default_factory=list)
     import_from_node_types: list[str] = field(default_factory=list)
     package_indicators: list[str] = field(default_factory=list)
+    ignore_directories: set[str] = field(default_factory=set)
     function_query: str | None = None
     class_query: str | None = None
     call_query: str | None = None
@@ -36,6 +183,7 @@ LANGUAGE_CONFIGS = {
         import_node_types=["import_statement"],
         import_from_node_types=["import_from_statement"],
         package_indicators=["__init__.py"],
+        ignore_directories=PYTHON_IGNORE_DIRECTORIES,
         function_query="""
     (function_definition
       name: (identifier) @function_name
@@ -96,6 +244,7 @@ LANGUAGE_CONFIGS = {
         import_node_types=["import_statement"],
         import_from_node_types=["import_statement"],
         package_indicators=["package.json"],
+        ignore_directories=JAVASCRIPT_IGNORE_DIRECTORIES,
         function_query="""
     [
       (function_declaration name: (identifier) @function_name) @function
@@ -151,6 +300,7 @@ LANGUAGE_CONFIGS = {
         import_node_types=["import_statement"],
         import_from_node_types=["import_statement"],
         package_indicators=["package.json", "tsconfig.json"],
+        ignore_directories=TYPESCRIPT_IGNORE_DIRECTORIES,
         function_query="""
     [
       (function_declaration name: (identifier) @function_name) @function
@@ -197,6 +347,7 @@ LANGUAGE_CONFIGS = {
         import_node_types=["import_declaration"],
         import_from_node_types=["import_spec"],
         package_indicators=["go.mod", "go.sum"],
+        ignore_directories=GO_IGNORE_DIRECTORIES,
         function_query="""
     [
       (function_declaration
@@ -248,6 +399,7 @@ LANGUAGE_CONFIGS = {
         import_node_types=["use_declaration"],
         import_from_node_types=["use_as_clause", "use_list"],
         package_indicators=["Cargo.toml"],
+        ignore_directories=RUST_IGNORE_DIRECTORIES,
         function_query="""
     [
       (function_item
@@ -295,3 +447,90 @@ def get_language_config(file_extension: str) -> LanguageConfig | None:
         if file_extension in config.file_extensions:
             return config
     return None
+
+
+def get_all_ignore_directories() -> set[str]:
+    """Get all ignore directories (common + all language-specific).
+
+    This is useful when indexing a codebase with multiple languages
+    or when you don't know what languages are present.
+    """
+    all_ignores = COMMON_IGNORE_DIRECTORIES.copy()
+    for config in LANGUAGE_CONFIGS.values():
+        all_ignores |= config.ignore_directories
+    return all_ignores
+
+
+def get_ignore_directories_for_languages(languages: list[str]) -> set[str]:
+    """Get ignore directories for specific languages.
+
+    Args:
+        languages: List of language names (e.g., ["python", "typescript"])
+
+    Returns:
+        Common ignores + language-specific ignores for the given languages
+    """
+    ignores = COMMON_IGNORE_DIRECTORIES.copy()
+    for lang in languages:
+        if lang in LANGUAGE_CONFIGS:
+            ignores |= LANGUAGE_CONFIGS[lang].ignore_directories
+    return ignores
+
+
+def should_ignore_directory(name: str, ignore_patterns: set[str] | None = None) -> bool:
+    """Check if a directory should be ignored.
+
+    Args:
+        name: Directory name (not full path)
+        ignore_patterns: Optional custom ignore patterns. If None, uses all patterns.
+
+    Returns:
+        True if directory should be ignored
+    """
+    patterns = (
+        ignore_patterns if ignore_patterns is not None else get_all_ignore_directories()
+    )
+
+    # Check explicit patterns
+    if name in patterns:
+        return True
+
+    # Check for glob patterns like "*.egg-info"
+    for pattern in patterns:
+        if pattern.startswith("*") and name.endswith(pattern[1:]):
+            return True
+
+    # Hidden directories (starting with .) are always ignored
+    if name.startswith("."):
+        return True
+
+    return False
+
+
+def should_ignore_file(name: str) -> bool:
+    """Check if a file should be ignored.
+
+    Args:
+        name: File name (not full path)
+
+    Returns:
+        True if file should be ignored (hidden files starting with .)
+    """
+    return name.startswith(".")
+
+
+def is_path_ignored(path: str | Path, ignore_patterns: set[str] | None = None) -> bool:
+    """Check if any part of a path should be ignored.
+
+    Args:
+        path: File or directory path (can be relative or absolute)
+        ignore_patterns: Optional custom ignore patterns
+
+    Returns:
+        True if any path component should be ignored
+    """
+    path_obj = Path(path) if isinstance(path, str) else path
+    for part in path_obj.parts:
+        if should_ignore_directory(part, ignore_patterns):
+            return True
+    return False