cicada-mcp 0.1.5__py3-none-any.whl → 0.2.0__py3-none-any.whl

cicada/mcp_entry.py

@@ -0,0 +1,683 @@
+#!/usr/bin/env python
+"""
+Entry point for cicada-mcp command.
+
+Behavior:
+- With no args: Start MCP server
+- With path arg: Start MCP server for that path
+- cicada-mcp install: Interactive setup with editor and model selection
+- With subcommands: Route to appropriate handler (same as cicada CLI)
+
+This provides unified command interface for both cicada and cicada-mcp.
+"""
+
+import argparse
+import sys
+
+
+def main():
+    """Main entry point for cicada-mcp command."""
+    # Known subcommands
+    known_subcommands = [
+        "install",
+        "server",
+        "claude",
+        "cursor",
+        "vs",
+        "index",
+        "index-pr",
+        "find-dead-code",
+        "clean",
+    ]
+
+    # Handle path argument for backward compatibility (cicada-mcp <path>)
+    # If first arg is not a known subcommand and not a flag, treat it as a path
+    server_path = None
+    if (
+        len(sys.argv) > 1
+        and sys.argv[1] not in known_subcommands
+        and not sys.argv[1].startswith("-")
+    ):
+        # Extract the path and remove it from sys.argv so argparse doesn't see it
+        server_path = sys.argv[1]
+        sys.argv = [sys.argv[0]] + sys.argv[2:]
+
+    parser = argparse.ArgumentParser(
+        prog="cicada-mcp",
+        description="Cicada MCP Server - AI-powered Elixir code analysis",
+        epilog="Run 'cicada-mcp <command> --help' for more information on a command.",
+    )
+
+    # Create subparsers for commands (optional to support default server mode)
+    subparsers = parser.add_subparsers(dest="command", help="Available commands", required=False)
+
+    # ========================================================================
+    # INSTALL subcommand - Interactive setup
+    # ========================================================================
+    install_parser = subparsers.add_parser(
+        "install",
+        help="Interactive setup for Cicada",
+        description="Interactive setup with editor and model selection",
+    )
+    install_parser.add_argument(
+        "repo",
+        nargs="?",
+        default=None,
+        help="Path to Elixir repository (default: current directory)",
+    )
+    install_parser.add_argument(
+        "--claude",
+        action="store_true",
+        help="Skip editor selection, use Claude Code",
+    )
+    install_parser.add_argument(
+        "--cursor",
+        action="store_true",
+        help="Skip editor selection, use Cursor",
+    )
+    install_parser.add_argument(
+        "--vs",
+        action="store_true",
+        help="Skip editor selection, use VS Code",
+    )
+    install_parser.add_argument(
+        "--nlp",
+        action="store_true",
+        help="Skip model selection, use Lemminflect",
+    )
+    install_parser.add_argument(
+        "--rag",
+        action="store_true",
+        help="Skip model selection, use BERT (default tier)",
+    )
+    install_parser.add_argument(
+        "--fast",
+        action="store_true",
+        help="Use BERT fast tier (requires --rag)",
+    )
+    install_parser.add_argument(
+        "--max",
+        action="store_true",
+        help="Use BERT max tier (requires --rag)",
+    )
+
+    # ========================================================================
+    # SERVER subcommand - Silent MCP server
+    # ========================================================================
+    server_parser = subparsers.add_parser(
+        "server",
+        help="Start MCP server (silent mode with defaults)",
+        description="Start MCP server with auto-setup using defaults",
+    )
+    server_parser.add_argument(
+        "repo",
+        nargs="?",
+        default=None,
+        help="Path to Elixir repository (default: current directory)",
+    )
+    server_parser.add_argument(
+        "--claude",
+        action="store_true",
+        help="Create Claude Code config before starting server",
+    )
+    server_parser.add_argument(
+        "--cursor",
+        action="store_true",
+        help="Create Cursor config before starting server",
+    )
+    server_parser.add_argument(
+        "--vs",
+        action="store_true",
+        help="Create VS Code config before starting server",
+    )
+    server_parser.add_argument(
+        "--nlp",
+        action="store_true",
+        help="Force Lemminflect (if reindexing needed)",
+    )
+    server_parser.add_argument(
+        "--rag",
+        action="store_true",
+        help="Force BERT (if reindexing needed)",
+    )
+    server_parser.add_argument(
+        "--fast",
+        action="store_true",
+        help="Force BERT fast tier (requires --rag)",
+    )
+    server_parser.add_argument(
+        "--max",
+        action="store_true",
+        help="Force BERT max tier (requires --rag)",
+    )
+
+    # ========================================================================
+    # CLAUDE subcommand (editor setup)
+    # ========================================================================
+    claude_parser = subparsers.add_parser(
+        "claude",
+        help="Setup Cicada for Claude Code editor",
+        description="One-command setup for Claude Code with keyword extraction",
+    )
+    claude_parser.add_argument(
+        "--nlp",
+        action="store_true",
+        help="Use NLP keyword extraction (lemminflect-based)",
+    )
+    claude_parser.add_argument(
+        "--rag",
+        action="store_true",
+        help="Use RAG-optimized keyword extraction (BERT-based embeddings)",
+    )
+    claude_parser.add_argument(
+        "--fast",
+        action="store_true",
+        help="Use fast tier model (requires --nlp or --rag)",
+    )
+    claude_parser.add_argument(
+        "--max",
+        action="store_true",
+        help="Use maximum quality tier model (requires --nlp or --rag)",
+    )
+
+    # ========================================================================
+    # CURSOR subcommand (editor setup)
+    # ========================================================================
+    cursor_parser = subparsers.add_parser(
+        "cursor",
+        help="Setup Cicada for Cursor editor",
+        description="One-command setup for Cursor with keyword extraction",
+    )
+    cursor_parser.add_argument(
+        "--nlp",
+        action="store_true",
+        help="Use NLP keyword extraction (lemminflect-based)",
+    )
+    cursor_parser.add_argument(
+        "--rag",
+        action="store_true",
+        help="Use RAG-optimized keyword extraction (BERT-based embeddings)",
+    )
+    cursor_parser.add_argument(
+        "--fast",
+        action="store_true",
+        help="Use fast tier model (requires --nlp or --rag)",
+    )
+    cursor_parser.add_argument(
+        "--max",
+        action="store_true",
+        help="Use maximum quality tier model (requires --nlp or --rag)",
+    )
+
+    # ========================================================================
+    # VS subcommand (editor setup)
+    # ========================================================================
+    vs_parser = subparsers.add_parser(
+        "vs",
+        help="Setup Cicada for VS Code editor",
+        description="One-command setup for VS Code with keyword extraction",
+    )
+    vs_parser.add_argument(
+        "--nlp",
+        action="store_true",
+        help="Use NLP keyword extraction (lemminflect-based)",
+    )
+    vs_parser.add_argument(
+        "--rag",
+        action="store_true",
+        help="Use RAG-optimized keyword extraction (BERT-based embeddings)",
+    )
+    vs_parser.add_argument(
+        "--fast",
+        action="store_true",
+        help="Use fast tier model (requires --nlp or --rag)",
+    )
+    vs_parser.add_argument(
+        "--max",
+        action="store_true",
+        help="Use maximum quality tier model (requires --nlp or --rag)",
+    )
+
+    # ========================================================================
+    # INDEX subcommand
+    # ========================================================================
+    index_parser = subparsers.add_parser(
+        "index",
+        help="Index an Elixir repository to extract modules and functions",
+        description="Index current Elixir repository to extract modules and functions",
+    )
+    index_parser.add_argument(
+        "repo",
+        nargs="?",
+        default=".",
+        help="Path to the Elixir repository to index (default: current directory)",
+    )
+    index_parser.add_argument(
+        "--nlp",
+        action="store_true",
+        help="Use NLP keyword extraction (lemminflect-based)",
+    )
+    index_parser.add_argument(
+        "--rag",
+        action="store_true",
+        help="Use RAG-optimized keyword extraction (BERT-based embeddings)",
+    )
+    index_parser.add_argument(
+        "--fast",
+        action="store_true",
+        help="Use fast tier model (requires --nlp or --rag)",
+    )
+    index_parser.add_argument(
+        "--max",
+        action="store_true",
+        help="Use maximum quality tier model (requires --nlp or --rag)",
+    )
+
+    # ========================================================================
+    # INDEX-PR subcommand
+    # ========================================================================
+    index_pr_parser = subparsers.add_parser(
+        "index-pr",
+        help="Index GitHub pull requests for fast offline lookup",
+        description="Index GitHub pull requests for fast offline lookup",
+    )
+    index_pr_parser.add_argument(
+        "repo",
+        nargs="?",
+        default=".",
+        help="Path to git repository (default: current directory)",
+    )
+    index_pr_parser.add_argument(
+        "--clean",
+        action="store_true",
+        help="Clean and rebuild the entire index from scratch (default: incremental update)",
+    )
+
+    # ========================================================================
+    # FIND-DEAD-CODE subcommand
+    # ========================================================================
+    dead_code_parser = subparsers.add_parser(
+        "find-dead-code",
+        help="Find potentially unused public functions in Elixir codebase",
+        description="Find potentially unused public functions in Elixir codebase",
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="""
+Confidence Levels:
+  high   - Zero usage, no dynamic call indicators, no behaviors/uses
+  medium - Zero usage, but module has behaviors or uses (possible callbacks)
+  low    - Zero usage, but module passed as value (possible dynamic calls)
+
+Examples:
+  cicada-mcp find-dead-code                      # Show high confidence candidates
+  cicada-mcp find-dead-code --min-confidence low # Show all candidates
+  cicada-mcp find-dead-code --format json        # Output as JSON
+        """,
+    )
+    dead_code_parser.add_argument(
+        "--index",
+        default=None,
+        help="Path to index file (default: uses current directory's centralized index)",
+    )
+    dead_code_parser.add_argument(
+        "--format",
+        choices=["markdown", "json"],
+        default="markdown",
+        help="Output format (default: markdown)",
+    )
+    dead_code_parser.add_argument(
+        "--min-confidence",
+        choices=["high", "medium", "low"],
+        default="high",
+        help="Minimum confidence level to show (default: high)",
+    )
+
+    # ========================================================================
+    # CLEAN subcommand
+    # ========================================================================
+    clean_parser = subparsers.add_parser(
+        "clean",
+        help="Remove Cicada configuration and indexes",
+        description="Remove Cicada configuration and indexes for current repository",
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="""
+Examples:
+  cicada-mcp clean                   # Remove everything (interactive with confirmation)
+  cicada-mcp clean -f                # Remove everything (skip confirmation)
+  cicada-mcp clean --index           # Remove main index (index.json, hashes.json)
+  cicada-mcp clean --pr-index        # Remove PR index (pr_index.json)
+  cicada-mcp clean --all             # Remove ALL project storage
+  cicada-mcp clean --all -f          # Remove ALL project storage (skip confirmation)
+        """,
+    )
+    clean_parser.add_argument(
+        "-f",
+        "--force",
+        action="store_true",
+        help="Skip confirmation prompt (for full clean or --all)",
+    )
+    clean_parser.add_argument(
+        "--index",
+        action="store_true",
+        help="Remove only main index files (index.json, hashes.json)",
+    )
+    clean_parser.add_argument(
+        "--pr-index",
+        action="store_true",
+        help="Remove only PR index file (pr_index.json)",
+    )
+    clean_parser.add_argument(
+        "--all",
+        action="store_true",
+        help="Remove ALL Cicada storage for all projects (~/.cicada/projects/)",
+    )
+
+    # Parse arguments
+    args = parser.parse_args()
+
+    # Store the server path for default handler
+    args._server_path = server_path
+
+    # Route to appropriate handler
+    if args.command == "install":
+        handle_install(args)
+    elif args.command == "server":
+        handle_server(args)
+    elif args.command == "claude":
+        from cicada.cli import handle_editor_setup
+
+        handle_editor_setup(args, "claude")
+    elif args.command == "cursor":
+        from cicada.cli import handle_editor_setup
+
+        handle_editor_setup(args, "cursor")
+    elif args.command == "vs":
+        from cicada.cli import handle_editor_setup
+
+        handle_editor_setup(args, "vs")
+    elif args.command == "index":
+        from cicada.cli import handle_index
+
+        handle_index(args)
+    elif args.command == "index-pr":
+        from cicada.cli import handle_index_pr
+
+        handle_index_pr(args)
+    elif args.command == "find-dead-code":
+        from cicada.cli import handle_find_dead_code
+
+        handle_find_dead_code(args)
+    elif args.command == "clean":
+        from cicada.cli import handle_clean
+
+        handle_clean(args)
+    else:
+        # No subcommand - start server
+        handle_default_server(args)
+
+
+def handle_default_server(args):
+    """
+    Handle default behavior when called with no subcommand.
+    Starts MCP server silently.
+    """
+    import asyncio
+    import os
+    from pathlib import Path
+
+    # Check if a path was provided (backward compatibility: cicada-mcp <path>)
+    if hasattr(args, "_server_path") and args._server_path:
+        repo_path = Path(args._server_path).resolve()
+        os.environ["CICADA_REPO_PATH"] = str(repo_path)
+
+    # Import and run MCP server
+    from cicada.mcp_server import async_main
+
+    asyncio.run(async_main())
+
+
+def handle_install(args):
+    """
+    Handle the install subcommand (interactive setup).
+
+    Behavior:
+    - INTERACTIVE: shows prompts and menus
+    - Can skip prompts with flags (--claude, --cursor, --vs, --nlp, --rag)
+    - Creates editor config and indexes repository
+    """
+    from pathlib import Path
+
+    from cicada.interactive_setup import show_first_time_setup
+    from cicada.setup import EditorType, setup
+    from cicada.utils import get_config_path, get_index_path
+
+    # Determine repository path
+    repo_path = Path(args.repo).resolve() if args.repo else Path.cwd().resolve()
+
+    # Validate it's an Elixir project
+    if not (repo_path / "mix.exs").exists():
+        print(f"Error: {repo_path} does not appear to be an Elixir project", file=sys.stderr)
+        print("(mix.exs not found)", file=sys.stderr)
+        sys.exit(1)
+
+    # Validate flag combinations
+    if (args.fast or args.max) and not args.rag:
+        print("Error: --fast or --max requires --rag", file=sys.stderr)
+        sys.exit(1)
+
+    if args.nlp and args.rag:
+        print("Error: Cannot specify both --nlp and --rag", file=sys.stderr)
+        sys.exit(1)
+
+    # Count editor flags
+    editor_flags = [args.claude, args.cursor, args.vs]
+    editor_count = sum(editor_flags)
+
+    if editor_count > 1:
+        print("Error: Can only specify one editor flag for install command", file=sys.stderr)
+        sys.exit(1)
+
+    # Determine editor from flags
+    editor: EditorType | None = None
+    if args.claude:
+        editor = "claude"
+    elif args.cursor:
+        editor = "cursor"
+    elif args.vs:
+        editor = "vs"
+
+    # Determine keyword method and tier from flags
+    keyword_method = None
+    keyword_tier = None
+
+    if args.nlp:
+        keyword_method = "lemminflect"
+        keyword_tier = "regular"
+    elif args.rag:
+        keyword_method = "bert"
+        if args.fast:
+            keyword_tier = "fast"
+        elif args.max:
+            keyword_tier = "max"
+        else:
+            keyword_tier = "regular"
+
+    # Check if index already exists
+    config_path = get_config_path(repo_path)
+    index_path = get_index_path(repo_path)
+    index_exists = config_path.exists() and index_path.exists()
+
+    # If no flags provided, use full interactive setup
+    if editor is None and keyword_method is None:
+        from cicada.interactive_setup import show_full_interactive_setup
+
+        show_full_interactive_setup(repo_path)
+        return
+
+    # If only model flags provided (no editor), prompt for editor
+    if editor is None:
+        # Show editor selection menu
+        from simple_term_menu import TerminalMenu
+
+        print("Select editor to configure:")
+        print()
+        editor_options = [
+            "Claude Code (Claude AI assistant)",
+            "Cursor (AI-powered code editor)",
+            "VS Code (Visual Studio Code)",
+        ]
+        editor_menu = TerminalMenu(editor_options, title="Choose your editor:")
+        menu_idx = editor_menu.show()
+
+        if menu_idx is None:
+            print("\nSetup cancelled.")
+            sys.exit(0)
+
+        # Map menu index to editor type (menu_idx is guaranteed to be int here)
+        assert isinstance(menu_idx, int), "menu_idx must be an integer"
+        editor_map: tuple[EditorType, EditorType, EditorType] = ("claude", "cursor", "vs")
+        editor = editor_map[menu_idx]
+
+    # If only editor flag provided (no model), prompt for model (unless index exists)
+    if keyword_method is None and not index_exists:
+        keyword_method, keyword_tier = show_first_time_setup()
+
+    # If index exists but no model flags, use existing settings
+    if keyword_method is None and index_exists:
+        import yaml
+
+        try:
+            with open(config_path) as f:
+                existing_config = yaml.safe_load(f)
+                keyword_method = existing_config.get("keyword_extraction", {}).get(
+                    "method", "lemminflect"
+                )
+                keyword_tier = existing_config.get("keyword_extraction", {}).get("tier", "regular")
+        except Exception:
+            # If we can't read config, use defaults
+            keyword_method = "lemminflect"
+            keyword_tier = "regular"
+
+    # Run setup
+    try:
+        setup(
+            editor,
+            repo_path,
+            keyword_method=keyword_method,
+            keyword_tier=keyword_tier,
+            index_exists=index_exists,
+        )
+    except Exception as e:
+        print(f"\nError: Setup failed: {e}", file=sys.stderr)
+        sys.exit(1)
+
+
+def handle_server(args):
+    """
+    Handle the server subcommand (silent MCP server with optional configs).
+
+    Behavior:
+    - SILENT: no prompts, no interactive menus
+    - Auto-setup if needed (uses default model: lemminflect)
+    - Creates editor configs if flags provided (--claude, --cursor, --vs)
+    - Starts MCP server on stdio
+    """
+    import asyncio
+    import os
+    from pathlib import Path
+
+    from cicada.setup import (
+        EditorType,
+        create_config_yaml,
+        index_repository,
+        setup_multiple_editors,
+    )
+    from cicada.utils import create_storage_dir, get_config_path, get_index_path
+
+    # Determine repository path
+    repo_path = Path(args.repo).resolve() if args.repo else Path.cwd().resolve()
+
+    # Validate it's an Elixir project
+    if not (repo_path / "mix.exs").exists():
+        print(
+            f"Error: {repo_path} does not appear to be an Elixir project (mix.exs not found)",
+            file=sys.stderr,
+        )
+        sys.exit(1)
+
+    # Validate flag combinations
+    if (args.fast or args.max) and not args.rag:
+        print("Error: --fast or --max requires --rag", file=sys.stderr)
+        sys.exit(1)
+
+    if args.nlp and args.rag:
+        print("Error: Cannot specify both --nlp and --rag", file=sys.stderr)
+        sys.exit(1)
+
+    # Create storage directory
+    storage_dir = create_storage_dir(repo_path)
+
+    # Determine keyword extraction method and tier
+    keyword_method = None
+    keyword_tier = None
+
+    if args.nlp:
+        keyword_method = "lemminflect"
+        keyword_tier = "regular"
+    elif args.rag:
+        keyword_method = "bert"
+        if args.fast:
+            keyword_tier = "fast"
+        elif args.max:
+            keyword_tier = "max"
+        else:
+            keyword_tier = "regular"
+
+    # Check if setup is needed
+    config_path = get_config_path(repo_path)
+    index_path = get_index_path(repo_path)
+    needs_setup = not (config_path.exists() and index_path.exists())
+
+    if needs_setup:
+        # Silent setup with defaults
+        # If no method specified, default to lemminflect (fastest, no downloads)
+        if keyword_method is None:
+            keyword_method = "lemminflect"
+            keyword_tier = "regular"
+
+        # Create config.yaml (silent)
+        create_config_yaml(repo_path, storage_dir, keyword_method, keyword_tier, verbose=False)
+
+        # Index repository (silent)
+        try:
+            index_repository(repo_path, force_full=False, verbose=False)
+        except Exception as e:
+            print(f"Error during indexing: {e}", file=sys.stderr)
+            sys.exit(1)
+
+    # Create editor configs if flags provided
+    editors_to_configure: list[EditorType] = []
+    if args.claude:
+        editors_to_configure.append("claude")
+    if args.cursor:
+        editors_to_configure.append("cursor")
+    if args.vs:
+        editors_to_configure.append("vs")
+
+    if editors_to_configure:
+        try:
+            setup_multiple_editors(editors_to_configure, repo_path, storage_dir, verbose=False)
+        except Exception as e:
+            print(f"Error creating editor configs: {e}", file=sys.stderr)
+            sys.exit(1)
+
+    # Set environment variable for MCP server
+    os.environ["CICADA_REPO_PATH"] = str(repo_path)
+
+    # Start MCP server (silent)
+    from cicada.mcp_server import async_main
+
+    asyncio.run(async_main())
+
+
+if __name__ == "__main__":
+    main()