mcp-vector-search 0.7.3__py3-none-any.whl → 0.7.4__py3-none-any.whl

mcp_vector_search/__init__.py

@@ -1,7 +1,7 @@
 """MCP Vector Search - CLI-first semantic code search with MCP integration."""
 
-__version__ = "0.7.3"
-__build__ = "25"
+__version__ = "0.7.4"
+__build__ = "26"
 __author__ = "Robert Matsuoka"
 __email__ = "bobmatnyc@gmail.com"
 

mcp_vector_search/cli/commands/demo.py

@@ -0,0 +1,360 @@
+"""Demo command for mcp-vector-search."""
+
+import subprocess
+import sys
+import tempfile
+from pathlib import Path
+
+import typer
+from loguru import logger
+from rich.console import Console
+
+from ..output import print_error, print_info, print_success, print_warning
+
+console = Console()
+
+demo_app = typer.Typer(
+    name="demo",
+    help="🎬 Run interactive demo with sample project",
+    add_completion=False,
+    rich_markup_mode="rich",
+)
+
+
+@demo_app.callback(invoke_without_command=True)
+def demo(
+    ctx: typer.Context,
+    quick: bool = typer.Option(
+        False,
+        "--quick",
+        "-q",
+        help="Skip search demo, only show installation",
+    ),
+    keep_files: bool = typer.Option(
+        False,
+        "--keep",
+        "-k",
+        help="Keep demo files (don't auto-cleanup)",
+    ),
+) -> None:
+    """
+    Run installation demo with sample project.
+
+    [bold cyan]What this does:[/bold cyan]
+      1. Creates a temporary project with sample Python files
+      2. Initializes mcp-vector-search in the demo project
+      3. Indexes the sample code
+      4. Runs a sample semantic search
+      5. Shows you the results
+
+    [bold cyan]Examples:[/bold cyan]
+      [green]mcp-vector-search demo[/green]
+        Run full interactive demo
+
+      [green]mcp-vector-search demo --quick[/green]
+        Skip search demo, only show installation
+
+      [green]mcp-vector-search demo --keep[/green]
+        Keep demo files for inspection
+
+    [dim]Perfect for first-time users to see how semantic search works![/dim]
+    """
+    # If subcommand was invoked, don't run the main demo
+    if ctx.invoked_subcommand is not None:
+        return
+
+    try:
+        print_info("🎬 Running mcp-vector-search installation demo...")
+        console.print(
+            "\n[bold]This demo will:[/bold]\n"
+            "  1. Create a sample Python project\n"
+            "  2. Initialize and index it\n"
+            "  3. Run a semantic search\n"
+            "  4. Show you the results\n"
+        )
+
+        # Create temporary demo directory
+        temp_context = tempfile.TemporaryDirectory(prefix="mcp-demo-")
+        temp_dir = temp_context.__enter__()
+
+        try:
+            demo_dir = Path(temp_dir) / "demo-project"
+            demo_dir.mkdir()
+
+            # Create sample files
+            console.print("[bold blue]📝 Creating sample files...[/bold blue]")
+
+            (demo_dir / "main.py").write_text(
+                '''"""Main application module."""
+
+def main():
+    """Main entry point for the application."""
+    print("Hello, World!")
+    user_service = UserService()
+    user_service.create_user("Alice", "alice@example.com")
+
+
+class UserService:
+    """Service for managing users."""
+
+    def create_user(self, name: str, email: str):
+        """Create a new user with the given name and email.
+
+        Args:
+            name: The user's full name
+            email: The user's email address
+
+        Returns:
+            dict: User object with name and email
+        """
+        print(f"Creating user: {name} ({email})")
+        return {"name": name, "email": email}
+
+    def authenticate_user(self, email: str, password: str):
+        """Authenticate user with email and password.
+
+        Args:
+            email: User's email address
+            password: User's password
+
+        Returns:
+            bool: True if authentication successful
+        """
+        # Simple authentication logic
+        return email.endswith("@example.com")
+
+    def reset_password(self, email: str):
+        """Send password reset email to user.
+
+        Args:
+            email: User's email address
+        """
+        print(f"Sending password reset to {email}")
+
+
+if __name__ == "__main__":
+    main()
+'''
+            )
+
+            (demo_dir / "utils.py").write_text(
+                '''"""Utility functions for the application."""
+
+import hashlib
+import json
+from typing import Any, Dict
+
+
+def load_config(config_path: str) -> Dict[str, Any]:
+    """Load configuration from JSON file.
+
+    Args:
+        config_path: Path to JSON configuration file
+
+    Returns:
+        dict: Configuration dictionary
+    """
+    with open(config_path, "r") as f:
+        return json.load(f)
+
+
+def validate_email(email: str) -> bool:
+    """Validate email address format.
+
+    Args:
+        email: Email address to validate
+
+    Returns:
+        bool: True if email format is valid
+    """
+    return "@" in email and "." in email.split("@")[1]
+
+
+def hash_password(password: str) -> str:
+    """Hash password for secure storage.
+
+    Args:
+        password: Plain text password
+
+    Returns:
+        str: Hexadecimal hash of password
+    """
+    return hashlib.sha256(password.encode()).hexdigest()
+
+
+def validate_password_strength(password: str) -> bool:
+    """Check if password meets security requirements.
+
+    Args:
+        password: Password to validate
+
+    Returns:
+        bool: True if password is strong enough
+    """
+    return len(password) >= 8 and any(c.isdigit() for c in password)
+'''
+            )
+
+            (demo_dir / "api.py").write_text(
+                '''"""API endpoints for user management."""
+
+from typing import Dict, List
+
+
+class UserAPI:
+    """REST API endpoints for user operations."""
+
+    def get_user(self, user_id: int) -> Dict:
+        """Get user by ID.
+
+        Args:
+            user_id: Unique user identifier
+
+        Returns:
+            dict: User object
+        """
+        return {"id": user_id, "name": "Example User"}
+
+    def list_users(self, limit: int = 10) -> List[Dict]:
+        """List all users with pagination.
+
+        Args:
+            limit: Maximum number of users to return
+
+        Returns:
+            list: List of user objects
+        """
+        return [{"id": i, "name": f"User {i}"} for i in range(limit)]
+
+    def create_user_endpoint(self, name: str, email: str) -> Dict:
+        """Create a new user via API.
+
+        Args:
+            name: User's full name
+            email: User's email address
+
+        Returns:
+            dict: Created user object with ID
+        """
+        return {"id": 1, "name": name, "email": email}
+'''
+            )
+
+            console.print(f"[green]✅ Created 3 sample files in:[/green] {demo_dir}\n")
+
+            # Run initialization
+            print_info("🔧 Initializing mcp-vector-search in demo project...")
+
+            result = subprocess.run(
+                [
+                    sys.executable,
+                    "-m",
+                    "mcp_vector_search.cli.main",
+                    "--project-root",
+                    str(demo_dir),
+                    "init",
+                    "--extensions",
+                    ".py",
+                    "--no-mcp",  # Skip MCP for demo
+                    "--auto-index",  # Auto-index after init
+                ],
+                input="y\n",  # Auto-confirm the setup
+                capture_output=True,
+                text=True,
+            )
+
+            if result.returncode != 0:
+                print_error(f"Demo initialization failed: {result.stderr}")
+                raise typer.Exit(1)
+
+            print_success("✅ Demo project initialized and indexed!")
+
+            if not quick:
+                # Give the index a moment to settle
+                import time
+
+                time.sleep(1)
+
+                # Run sample searches to demonstrate different features
+                searches = [
+                    ("user authentication", "Finding authentication-related code"),
+                    ("password", "Finding password-related functions"),
+                    ("email validation", "Finding email validation logic"),
+                ]
+
+                for query, description in searches:
+                    console.print(f"\n[bold blue]🔍 {description}:[/bold blue]")
+                    console.print(f"[dim]Query: '{query}'[/dim]\n")
+
+                    search_result = subprocess.run(
+                        [
+                            sys.executable,
+                            "-m",
+                            "mcp_vector_search.cli.main",
+                            "--project-root",
+                            str(demo_dir),
+                            "search",
+                            "--limit",
+                            "2",
+                            query,
+                        ],
+                        capture_output=True,
+                        text=True,
+                    )
+
+                    if search_result.returncode == 0 and search_result.stdout.strip():
+                        console.print(search_result.stdout)
+                    else:
+                        # Show a friendly message if search didn't return results
+                        console.print(
+                            "[dim]Note: Search completed but may need more time "
+                            "for the index to fully settle.[/dim]"
+                        )
+
+            # Show summary
+            console.print("\n" + "=" * 70)
+            console.print("[bold green]🎉 Demo completed successfully![/bold green]")
+            console.print("=" * 70 + "\n")
+
+            console.print("[bold]What you just saw:[/bold]")
+            console.print("  ✅ Sample project creation")
+            console.print("  ✅ Automatic code indexing")
+            if not quick:
+                console.print("  ✅ Semantic code search in action")
+                console.print(
+                    "  ✅ Finding code by meaning (not just keywords)\n"
+                )
+
+            console.print("[bold cyan]Next steps to use in your project:[/bold cyan]")
+            console.print("  1. [green]cd /your/project[/green]")
+            console.print("  2. [green]mcp-vector-search init[/green]")
+            console.print("  3. [green]mcp-vector-search search 'your query'[/green]\n")
+
+            if keep_files:
+                console.print(
+                    f"[bold]Demo files saved at:[/bold] [cyan]{demo_dir}[/cyan]"
+                )
+                console.print(
+                    "[yellow]Note: This is still in a temp directory, "
+                    "copy files if you want to keep them![/yellow]\n"
+                )
+            else:
+                console.print(
+                    "[dim]Demo files will be cleaned up automatically.[/dim]\n"
+                )
+
+        finally:
+            if not keep_files:
+                temp_context.__exit__(None, None, None)
+            else:
+                # Keep the context manager open
+                pass
+
+    except Exception as e:
+        logger.error(f"Demo failed: {e}")
+        print_error(f"Demo failed: {e}")
+        raise typer.Exit(1)
+
+
+if __name__ == "__main__":
+    demo_app()

mcp_vector_search/cli/commands/index.py

@@ -217,10 +217,15 @@ async def _run_batch_indexing(
             show_progress=show_progress,
         )
 
-    print_success(f"Indexed {indexed_count} files")
-
     # Show statistics
     stats = await indexer.get_indexing_stats()
+
+    # Display success message with chunk count for clarity
+    total_chunks = stats.get("total_chunks", 0)
+    print_success(
+        f"Processed {indexed_count} files ({total_chunks} searchable chunks created)"
+    )
+
     print_index_stats(stats)
 
     # Add next-step hints
@@ -381,10 +386,15 @@ async def _reindex_entire_project(project_root: Path) -> None:
 
                 progress.update(task, completed=indexed_count, total=indexed_count)
 
-            print_success(f"Successfully reindexed {indexed_count} files")
-
             # Show statistics
             stats = await indexer.get_indexing_stats()
+
+            # Display success message with chunk count for clarity
+            total_chunks = stats.get("total_chunks", 0)
+            print_success(
+                f"Processed {indexed_count} files ({total_chunks} searchable chunks created)"
+            )
+
             print_index_stats(stats)
 
     except Exception as e:

mcp_vector_search/cli/commands/init.py

@@ -6,7 +6,6 @@ import typer
 from loguru import logger
 
 from ...config.constants import (
-    SUBPROCESS_INSTALL_TIMEOUT,
     SUBPROCESS_MCP_TIMEOUT,
     SUBPROCESS_TEST_TIMEOUT,
 )

mcp_vector_search/cli/main.py

@@ -34,6 +34,7 @@ unfamiliar codebases, finding similar patterns, and integrating with AI tools.
 
 [bold cyan]Main Commands:[/bold cyan]
   init      🔧 Initialize project
+  demo      🎬 Run interactive demo
   doctor    🩺 Check system health
   status    📊 Show project status
   search    🔍 Search code semantically
@@ -51,6 +52,7 @@ unfamiliar codebases, finding similar patterns, and integrating with AI tools.
 
 # Import command modules
 from .commands.config import config_app  # noqa: E402
+from .commands.demo import demo_app  # noqa: E402
 from .commands.index import index_app  # noqa: E402
 from .commands.init import init_app  # noqa: E402
 from .commands.mcp import mcp_app  # noqa: E402
@@ -65,29 +67,32 @@ from .commands.status import main as status_main  # noqa: E402
 # Use Typer group for init to support both direct call and subcommands
 app.add_typer(init_app, name="init", help="🔧 Initialize project for semantic search")
 
-# 2. DOCTOR - System health check
+# 2. DEMO - Interactive demo
+app.add_typer(demo_app, name="demo", help="🎬 Run interactive demo with sample project")
+
+# 3. DOCTOR - System health check
 # (defined below inline)
 
-# 3. STATUS - Project status
+# 4. STATUS - Project status
 app.command("status", help="📊 Show project status and statistics")(status_main)
 
-# 4. SEARCH - Search code
+# 5. SEARCH - Search code
 # Register search as both a command and a typer group
 app.add_typer(search_app, name="search", help="🔍 Search code semantically")
 
-# 5. INDEX - Index codebase
+# 6. INDEX - Index codebase
 app.add_typer(index_app, name="index", help="📇 Index codebase for semantic search")
 
-# 6. MCP - MCP integration
+# 7. MCP - MCP integration
 app.add_typer(mcp_app, name="mcp", help="🤖 Manage Claude Code MCP integration")
 
-# 7. CONFIG - Configuration
+# 8. CONFIG - Configuration
 app.add_typer(config_app, name="config", help="⚙️  Manage project configuration")
 
-# 8. HELP - Enhanced help
+# 9. HELP - Enhanced help
 # (defined below inline)
 
-# 9. VERSION - Version info
+# 10. VERSION - Version info
 # (defined below inline)
 
 
@@ -117,15 +122,6 @@ def deprecated_install():
     _deprecated_command("install", "init")()
 
 
-# Deprecated: demo -> removed
-@app.command("demo", hidden=True)
-def deprecated_demo():
-    """[DEPRECATED] Command removed."""
-    print_warning(
-        "⚠️  The 'demo' command has been removed.\n"
-        "   Use [cyan]mcp-vector-search init --help[/cyan] for setup instructions."
-    )
-    raise typer.Exit(1)
 
 
 # Deprecated: find -> search

mcp_vector_search/core/database.py

@@ -377,39 +377,36 @@ class ChromaVectorDatabase(VectorDatabase):
             # Get total count
             count = self._collection.count()
 
-            # Get sample for language distribution
-            sample_results = self._collection.get(
-                limit=min(1000, count) if count > 0 else 0,
-                include=["metadatas"],
-            )
+            # Get ALL metadata to analyze (not just a sample)
+            # Only fetch metadata, not embeddings, for performance
+            results = self._collection.get(include=["metadatas"])
 
-            languages = {}
-            file_types = {}
+            # Count unique files from all chunks
+            files = {m.get("file_path", "") for m in results.get("metadatas", [])}
 
-            if sample_results["metadatas"]:
-                for metadata in sample_results["metadatas"]:
-                    # Count languages
-                    lang = metadata.get("language", "unknown")
-                    languages[lang] = languages.get(lang, 0) + 1
+            # Count languages and file types
+            language_counts = {}
+            file_type_counts = {}
 
-                    # Count file types
-                    file_path = metadata.get("file_path", "")
+            for metadata in results.get("metadatas", []):
+                # Count languages
+                lang = metadata.get("language", "unknown")
+                language_counts[lang] = language_counts.get(lang, 0) + 1
+
+                # Count file types
+                file_path = metadata.get("file_path", "")
+                if file_path:
                     ext = Path(file_path).suffix or "no_extension"
-                    file_types[ext] = file_types.get(ext, 0) + 1
+                    file_type_counts[ext] = file_type_counts.get(ext, 0) + 1
 
             # Estimate index size (rough approximation)
             index_size_mb = count * 0.001  # Rough estimate
 
             return IndexStats(
-                total_files=len(
-                    {
-                        m.get("file_path", "")
-                        for m in sample_results.get("metadatas", [])
-                    }
-                ),
+                total_files=len(files),
                 total_chunks=count,
-                languages=languages,
-                file_types=file_types,
+                languages=language_counts,
+                file_types=file_type_counts,
                 index_size_mb=index_size_mb,
                 last_updated="unknown",  # TODO: Track this
                 embedding_model="unknown",  # TODO: Track this

{mcp_vector_search-0.7.3.dist-info → mcp_vector_search-0.7.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mcp-vector-search
-Version: 0.7.3
+Version: 0.7.4
 Summary: CLI-first semantic code search with MCP integration
 Project-URL: Homepage, https://github.com/bobmatnyc/mcp-vector-search
 Project-URL: Documentation, https://mcp-vector-search.readthedocs.io

{mcp_vector_search-0.7.3.dist-info → mcp_vector_search-0.7.4.dist-info}/RECORD

@@ -1,18 +1,19 @@
-mcp_vector_search/__init__.py,sha256=oYKNMl_U9KpUkmcl6mSeiFsg3eALZB2BhLLkDp9Edq4,299
+mcp_vector_search/__init__.py,sha256=dwjtm9sE-mB8ai4WJ0qaTi2dWD55TXfdknB0YzqYGzQ,299
 mcp_vector_search/py.typed,sha256=lCKeV9Qcn9sGtbRsgg-LJO2ZwWRuknnnlmomq3bJFH0,43
 mcp_vector_search/cli/__init__.py,sha256=TNB7CaOASz8u3yHWLbNmo8-GtHF0qwUjVKWAuNphKgo,40
 mcp_vector_search/cli/didyoumean.py,sha256=F_ss-EX4F9RgnMsEhdTwLpyNCah9SqnBZc2tBtzASck,15918
 mcp_vector_search/cli/export.py,sha256=iluxuRT2KELdKlQeDAlVkteiel4GGrng153UAw9H0as,10804
 mcp_vector_search/cli/history.py,sha256=6wRrSfxpUe9hJXuaEeVxOVkFlcpqkIiGfwzDgd5N6c8,9323
 mcp_vector_search/cli/interactive.py,sha256=T7P4dAdvbglznzQYgiePv5YNyOx9FeE57Y3OKYnnbYE,12744
-mcp_vector_search/cli/main.py,sha256=ZmTuPFIIqiZzn7Ml04EkzgNveBm4uIXe0kA6t4jysB8,14602
+mcp_vector_search/cli/main.py,sha256=5BJeueirPMGqNNf9NYZ8x6NhKOgu_EUkuRYNaP25I80,14495
 mcp_vector_search/cli/output.py,sha256=7ShIk_UKzhDzRGxI6JluPu0gGkbmKOevqgIAKR4oCa0,12560
 mcp_vector_search/cli/suggestions.py,sha256=h-UaxoLcHmFbhZSm0WG7nKJXAIRIqhv7aGsXijp7vA8,13273
 mcp_vector_search/cli/commands/__init__.py,sha256=vQls-YKZ54YEwmf7g1dL0T2SS9D4pdQljXzsUChG_V4,42
 mcp_vector_search/cli/commands/auto_index.py,sha256=imVVbxWRlA128NPdK9BetNNl3ELrsdq-hqcsLqyAmoM,12712
 mcp_vector_search/cli/commands/config.py,sha256=mKE8gUgAOqCM__4yzEEu9HJPbx9X15lN264zkDJBRxg,12399
-mcp_vector_search/cli/commands/index.py,sha256=AWQk_nc5vOW0nxy2yjRre_2uktA4J32D-302IS3dT3M,17862
-mcp_vector_search/cli/commands/init.py,sha256=TdgFspoWAeXSkeWHa_o4FWJHIFwI4OW8TD4GL2TPOwM,23274
+mcp_vector_search/cli/commands/demo.py,sha256=HOa5g4vDu_zjSq77bMcFaCck7RN9YsYgsIknAfeYIC8,10683
+mcp_vector_search/cli/commands/index.py,sha256=tqn6KjDygAHam5mINthYFBm-hA6I8QYDjrSVRmUtXLE,18213
+mcp_vector_search/cli/commands/init.py,sha256=2kdjtIPPeutKUXs65-6W1VQPF_BQrbV6_U3TCE7U5mw,23242
 mcp_vector_search/cli/commands/install.py,sha256=phk7Eb7UOU5IsRfJyaDPdOfdUWli9gyA4cHjhgXcNEI,24609
 mcp_vector_search/cli/commands/mcp.py,sha256=FKZNxYrDc7HfPTFBUEypCv-8atsrHEdbtU6Yfg9QUMA,18569
 mcp_vector_search/cli/commands/reset.py,sha256=bsIT6zjDf6gsvIkVaRaUClYzlTyNe--8t0NWkBY0ldU,13724
@@ -26,7 +27,7 @@ mcp_vector_search/config/settings.py,sha256=m8o8j-tvWcuzrnNL6YWbi2fFbcB3lZY1kMNi
 mcp_vector_search/core/__init__.py,sha256=bWKtKmmaFs7gG5XPCbrx77UYIVeO1FF8wIJxpj1dLNw,48
 mcp_vector_search/core/auto_indexer.py,sha256=0S4lZXaUgqEytMSA2FxQsh5hN7V1mbSLYVzEf_dslYQ,10307
 mcp_vector_search/core/connection_pool.py,sha256=Yo-gUQQbHawtuvh6OcJiAlbbvWQGQBd31QZOvs498fg,11224
-mcp_vector_search/core/database.py,sha256=4CDqtmxtg1EsajZP36Xwc43jD_84szVfq5SFLpJh3eg,35515
+mcp_vector_search/core/database.py,sha256=fMcclsO2dGOfsUSN38rJDRGVNEuPvGKHZlSSuuyIKag,35512
 mcp_vector_search/core/embeddings.py,sha256=wSMUNxZcuGPMxxQ1AbKqA1a3-0c6AiOqmuuI7OqTyaQ,10578
 mcp_vector_search/core/exceptions.py,sha256=3bCjT8wmrLz_0e_Tayr90049zNTKYFWZa19kl0saKz8,1597
 mcp_vector_search/core/factory.py,sha256=tM6Ft-V9buF7nn9xbRMU1ngji-BJOKt6BhtfQhFLmF4,10384
@@ -55,8 +56,8 @@ mcp_vector_search/utils/__init__.py,sha256=Eq6lY-oPMfCt-GpPUbg9QbmTHuQVmTaVDBMU2
 mcp_vector_search/utils/gitignore.py,sha256=bzie3V5gOGIN7j3FNVLLCx8O_hfZJDUqqAy5T3lT3Ek,7685
 mcp_vector_search/utils/timing.py,sha256=THC7mfbTYnUpnnDcblgQacYMzbEkfFoIShx6plmhCgg,11285
 mcp_vector_search/utils/version.py,sha256=d7fS-CLemxb8UzZ9j18zH0Y0Ud097ljKKYYOPulnGPE,1138
-mcp_vector_search-0.7.3.dist-info/METADATA,sha256=eIS0WUMDN9V0Qxw4Z48xiB9EDHxDxwblcyTrIS_sXXQ,19120
-mcp_vector_search-0.7.3.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-mcp_vector_search-0.7.3.dist-info/entry_points.txt,sha256=y3Ygtc_JiBchNEIL-tPABo7EbzBExGAxwGdkkeP5D2I,86
-mcp_vector_search-0.7.3.dist-info/licenses/LICENSE,sha256=FqZUgGJH_tZKZLQsMCpXaLawRyLmyFKRVfMwYyEcyTs,1072
-mcp_vector_search-0.7.3.dist-info/RECORD,,
+mcp_vector_search-0.7.4.dist-info/METADATA,sha256=k-RZl6CDcmhMyiS_w4oNCUzkrg7syOyIeKyCdYoereo,19120
+mcp_vector_search-0.7.4.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+mcp_vector_search-0.7.4.dist-info/entry_points.txt,sha256=y3Ygtc_JiBchNEIL-tPABo7EbzBExGAxwGdkkeP5D2I,86
+mcp_vector_search-0.7.4.dist-info/licenses/LICENSE,sha256=FqZUgGJH_tZKZLQsMCpXaLawRyLmyFKRVfMwYyEcyTs,1072
+mcp_vector_search-0.7.4.dist-info/RECORD,,