gnosisllm-knowledge 0.2.0__py3-none-any.whl

gnosisllm_knowledge/cli/app.py

@@ -0,0 +1,509 @@
+"""GnosisLLM Knowledge CLI Application.
+
+Main entry point assembling all CLI commands with enterprise-grade UX.
+"""
+
+from __future__ import annotations
+
+import asyncio
+from typing import Annotated, Optional
+
+import typer
+from rich.console import Console
+
+from gnosisllm_knowledge.cli.display import RichDisplayService
+from gnosisllm_knowledge.cli.utils import CliConfig
+
+# Main application
+app = typer.Typer(
+    name="gnosisllm-knowledge",
+    help="Enterprise-grade knowledge loading, indexing, and semantic search.",
+    no_args_is_help=True,
+    rich_markup_mode="rich",
+    pretty_exceptions_enable=True,
+    pretty_exceptions_show_locals=False,
+)
+
+# Shared console and display service
+console = Console()
+display = RichDisplayService(console)
+
+
+def version_callback(value: bool) -> None:
+    """Display version and exit."""
+    if value:
+        from gnosisllm_knowledge import __version__
+
+        console.print(f"gnosisllm-knowledge [cyan]{__version__}[/cyan]")
+        raise typer.Exit()
+
+
+@app.callback()
+def main_callback(
+    version: Annotated[
+        Optional[bool],
+        typer.Option(
+            "--version",
+            "-v",
+            help="Show version and exit.",
+            callback=version_callback,
+            is_eager=True,
+        ),
+    ] = None,
+) -> None:
+    """GnosisLLM Knowledge - Enterprise knowledge management CLI."""
+    pass
+
+
+# ============================================================================
+# SETUP COMMAND
+# ============================================================================
+
+
+@app.command()
+def setup(
+    host: Annotated[
+        str,
+        typer.Option("--host", "-h", help="OpenSearch host."),
+    ] = "localhost",
+    port: Annotated[
+        int,
+        typer.Option("--port", "-p", help="OpenSearch port."),
+    ] = 9200,
+    username: Annotated[
+        Optional[str],
+        typer.Option("--username", "-u", help="OpenSearch username."),
+    ] = None,
+    password: Annotated[
+        Optional[str],
+        typer.Option("--password", help="OpenSearch password."),
+    ] = None,
+    use_ssl: Annotated[
+        bool,
+        typer.Option("--use-ssl", help="Enable SSL."),
+    ] = False,
+    verify_certs: Annotated[
+        bool,
+        typer.Option("--verify-certs", help="Verify SSL certificates."),
+    ] = False,
+    force: Annotated[
+        bool,
+        typer.Option("--force", "-f", help="Clean up existing resources first."),
+    ] = False,
+    no_sample_data: Annotated[
+        bool,
+        typer.Option("--no-sample-data", help="Skip sample data ingestion."),
+    ] = False,
+    no_hybrid: Annotated[
+        bool,
+        typer.Option("--no-hybrid", help="Skip hybrid search pipeline."),
+    ] = False,
+) -> None:
+    """Configure OpenSearch with ML model for neural search.
+
+    Sets up the complete neural search infrastructure:
+    - OpenAI connector for embeddings
+    - Model group and deployed ML model
+    - Ingest pipeline for automatic embedding generation
+    - Search pipeline for hybrid scoring
+    - Knowledge index with k-NN vector mapping
+
+    [bold]Example:[/bold]
+        $ gnosisllm-knowledge setup
+        $ gnosisllm-knowledge setup --host opensearch.example.com --port 443 --use-ssl
+        $ gnosisllm-knowledge setup --force  # Clean and recreate
+    """
+    from gnosisllm_knowledge.cli.commands.setup import setup_command
+
+    asyncio.run(
+        setup_command(
+            display=display,
+            host=host,
+            port=port,
+            username=username,
+            password=password,
+            use_ssl=use_ssl,
+            verify_certs=verify_certs,
+            force=force,
+            no_sample_data=no_sample_data,
+            no_hybrid=no_hybrid,
+        )
+    )
+
+
+# ============================================================================
+# LOAD COMMAND
+# ============================================================================
+
+
+@app.command()
+def load(
+    source: Annotated[
+        str,
+        typer.Argument(help="URL or sitemap to load content from."),
+    ],
+    source_type: Annotated[
+        Optional[str],
+        typer.Option(
+            "--type",
+            "-t",
+            help="Source type: website, sitemap (auto-detects if not specified).",
+        ),
+    ] = None,
+    index: Annotated[
+        str,
+        typer.Option("--index", "-i", help="Target index name."),
+    ] = "knowledge",
+    account_id: Annotated[
+        Optional[str],
+        typer.Option("--account-id", "-a", help="Multi-tenant account ID."),
+    ] = None,
+    collection_id: Annotated[
+        Optional[str],
+        typer.Option("--collection-id", "-c", help="Collection grouping ID."),
+    ] = None,
+    source_id: Annotated[
+        Optional[str],
+        typer.Option("--source-id", "-s", help="Source identifier (defaults to URL)."),
+    ] = None,
+    batch_size: Annotated[
+        int,
+        typer.Option("--batch-size", "-b", help="Documents per indexing batch."),
+    ] = 100,
+    max_urls: Annotated[
+        int,
+        typer.Option("--max-urls", "-m", help="Maximum URLs to process from sitemap."),
+    ] = 1000,
+    force: Annotated[
+        bool,
+        typer.Option("--force", "-f", help="Delete existing source documents first."),
+    ] = False,
+    dry_run: Annotated[
+        bool,
+        typer.Option("--dry-run", help="Preview without indexing."),
+    ] = False,
+    verbose: Annotated[
+        bool,
+        typer.Option("--verbose", "-V", help="Show per-document progress."),
+    ] = False,
+) -> None:
+    """Load and index content from URLs or sitemaps.
+
+    Fetches content, chunks it for optimal embedding, and indexes
+    into OpenSearch with automatic embedding generation.
+
+    [bold]Example:[/bold]
+        $ gnosisllm-knowledge load https://docs.example.com/intro
+        $ gnosisllm-knowledge load https://example.com/sitemap.xml --type sitemap
+        $ gnosisllm-knowledge load https://docs.example.com/sitemap.xml --max-urls 500
+    """
+    from gnosisllm_knowledge.cli.commands.load import load_command
+
+    asyncio.run(
+        load_command(
+            display=display,
+            source=source,
+            source_type=source_type,
+            index_name=index,
+            account_id=account_id,
+            collection_id=collection_id,
+            source_id=source_id,
+            batch_size=batch_size,
+            max_urls=max_urls,
+            force=force,
+            dry_run=dry_run,
+            verbose=verbose,
+        )
+    )
+
+
+# ============================================================================
+# SEARCH COMMAND
+# ============================================================================
+
+
+@app.command()
+def search(
+    query: Annotated[
+        Optional[str],
+        typer.Argument(help="Search query text."),
+    ] = None,
+    mode: Annotated[
+        str,
+        typer.Option(
+            "--mode",
+            "-m",
+            help="Search mode: semantic, keyword, hybrid (default), agentic.",
+        ),
+    ] = "hybrid",
+    index: Annotated[
+        str,
+        typer.Option("--index", "-i", help="Index to search."),
+    ] = "knowledge",
+    limit: Annotated[
+        int,
+        typer.Option("--limit", "-l", help="Maximum results to return."),
+    ] = 5,
+    offset: Annotated[
+        int,
+        typer.Option("--offset", "-o", help="Pagination offset."),
+    ] = 0,
+    account_id: Annotated[
+        Optional[str],
+        typer.Option("--account-id", "-a", help="Filter by account ID."),
+    ] = None,
+    collection_ids: Annotated[
+        Optional[str],
+        typer.Option("--collection-ids", "-c", help="Filter by collection IDs (comma-separated)."),
+    ] = None,
+    source_ids: Annotated[
+        Optional[str],
+        typer.Option("--source-ids", "-s", help="Filter by source IDs (comma-separated)."),
+    ] = None,
+    min_score: Annotated[
+        float,
+        typer.Option("--min-score", help="Minimum score threshold (0.0-1.0)."),
+    ] = 0.0,
+    explain: Annotated[
+        bool,
+        typer.Option("--explain", "-e", help="Show score explanation."),
+    ] = False,
+    json_output: Annotated[
+        bool,
+        typer.Option("--json", "-j", help="Output as JSON for scripting."),
+    ] = False,
+    interactive: Annotated[
+        bool,
+        typer.Option("--interactive", "-I", help="Interactive search session."),
+    ] = False,
+    verbose: Annotated[
+        bool,
+        typer.Option("--verbose", "-V", help="Show full content (not truncated)."),
+    ] = False,
+) -> None:
+    """Search indexed content with semantic, keyword, or hybrid modes.
+
+    Supports multiple search strategies:
+    - [cyan]semantic[/cyan]: Meaning-based vector search using embeddings
+    - [cyan]keyword[/cyan]: Traditional BM25 text matching
+    - [cyan]hybrid[/cyan]: Combined semantic + keyword (default, best results)
+    - [cyan]agentic[/cyan]: AI-powered search with reasoning
+
+    [bold]Example:[/bold]
+        $ gnosisllm-knowledge search "how to configure auth"
+        $ gnosisllm-knowledge search "API reference" --mode semantic --limit 10
+        $ gnosisllm-knowledge search --interactive
+    """
+    from gnosisllm_knowledge.cli.commands.search import search_command
+
+    asyncio.run(
+        search_command(
+            display=display,
+            query=query,
+            mode=mode,
+            index_name=index,
+            limit=limit,
+            offset=offset,
+            account_id=account_id,
+            collection_ids=collection_ids,
+            source_ids=source_ids,
+            min_score=min_score,
+            explain=explain,
+            json_output=json_output,
+            interactive=interactive,
+            verbose=verbose,
+        )
+    )
+
+
+# ============================================================================
+# INFO COMMAND
+# ============================================================================
+
+
+@app.command()
+def info() -> None:
+    """Display configuration and environment information.
+
+    Shows current settings from environment variables and
+    validates connectivity to required services.
+    """
+    config = CliConfig.from_env()
+
+    display.header("GnosisLLM Knowledge", "Configuration and Environment Info")
+
+    display.table(
+        "OpenSearch Configuration",
+        [
+            ("Host", f"{config.opensearch_host}:{config.opensearch_port}"),
+            ("SSL", "Enabled" if config.opensearch_use_ssl else "Disabled"),
+            ("Auth", "Configured" if config.opensearch_username else "None"),
+            ("Model ID", config.opensearch_model_id or "[dim]Not set[/dim]"),
+            ("Index", config.opensearch_index_name),
+        ],
+    )
+
+    display.newline()
+
+    display.table(
+        "Embedding Configuration",
+        [
+            ("OpenAI Key", "✓ Set" if config.openai_api_key else "✗ Not set"),
+            ("Model", config.openai_embedding_model),
+            ("Dimension", str(config.openai_embedding_dimension)),
+        ],
+    )
+
+    display.newline()
+
+    display.table(
+        "Agentic Search Configuration",
+        [
+            ("Flow Agent", config.opensearch_flow_agent_id or "[dim]Not set[/dim]"),
+            ("Conversational Agent", config.opensearch_conversational_agent_id or "[dim]Not set[/dim]"),
+            ("LLM Model", config.agentic_llm_model),
+        ],
+    )
+
+    display.newline()
+
+    display.table(
+        "Content Fetching",
+        [
+            ("Neoreader", config.neoreader_host),
+        ],
+    )
+
+    # Validation
+    setup_errors = config.validate_for_setup()
+    search_errors = config.validate_for_search()
+
+    if setup_errors or search_errors:
+        display.newline()
+        display.warning("Configuration Issues:")
+        for error in setup_errors + search_errors:
+            display.error(f"  {error}")
+
+
+# ============================================================================
+# AGENTIC SUBCOMMAND GROUP
+# ============================================================================
+
+agentic_app = typer.Typer(
+    name="agentic",
+    help="AI-powered agentic search commands.",
+    no_args_is_help=True,
+    rich_markup_mode="rich",
+)
+app.add_typer(agentic_app, name="agentic")
+
+
+@agentic_app.command("setup")
+def agentic_setup(
+    agent_type: Annotated[
+        str,
+        typer.Option(
+            "--type",
+            "-t",
+            help="Agent type to setup: flow, conversational, or all (default).",
+        ),
+    ] = "all",
+    force: Annotated[
+        bool,
+        typer.Option("--force", "-f", help="Force recreate existing agents."),
+    ] = False,
+) -> None:
+    """Setup agentic search agents in OpenSearch.
+
+    Creates and deploys AI agents for intelligent search:
+    - [cyan]flow[/cyan]: Fast RAG for single-turn queries
+    - [cyan]conversational[/cyan]: Multi-turn with memory support
+
+    [bold]Example:[/bold]
+        $ gnosisllm-knowledge agentic setup
+        $ gnosisllm-knowledge agentic setup --type flow
+        $ gnosisllm-knowledge agentic setup --force
+    """
+    from gnosisllm_knowledge.cli.commands.agentic import agentic_setup_command
+
+    asyncio.run(
+        agentic_setup_command(
+            display=display,
+            agent_type=agent_type,
+            force=force,
+        )
+    )
+
+
+@agentic_app.command("chat")
+def agentic_chat(
+    index: Annotated[
+        str,
+        typer.Option("--index", "-i", help="Index to search."),
+    ] = "knowledge",
+    agent_type: Annotated[
+        str,
+        typer.Option(
+            "--type",
+            "-t",
+            help="Agent type: flow or conversational (default).",
+        ),
+    ] = "conversational",
+    account_id: Annotated[
+        Optional[str],
+        typer.Option("--account-id", "-a", help="Filter by account ID."),
+    ] = None,
+    collection_ids: Annotated[
+        Optional[str],
+        typer.Option("--collection-ids", "-c", help="Filter by collection IDs (comma-separated)."),
+    ] = None,
+    verbose: Annotated[
+        bool,
+        typer.Option("--verbose", "-V", help="Show reasoning steps."),
+    ] = False,
+) -> None:
+    """Interactive agentic chat session.
+
+    Start a conversation with the AI-powered knowledge assistant.
+    The agent remembers context for multi-turn dialogue.
+
+    [bold]Example:[/bold]
+        $ gnosisllm-knowledge agentic chat
+        $ gnosisllm-knowledge agentic chat --type flow
+        $ gnosisllm-knowledge agentic chat --verbose
+    """
+    from gnosisllm_knowledge.cli.commands.agentic import agentic_chat_command
+
+    asyncio.run(
+        agentic_chat_command(
+            display=display,
+            index_name=index,
+            agent_type=agent_type,
+            account_id=account_id,
+            collection_ids=collection_ids,
+            verbose=verbose,
+        )
+    )
+
+
+@agentic_app.command("status")
+def agentic_status() -> None:
+    """Show agentic search configuration status.
+
+    Displays configured agents and their health status.
+
+    [bold]Example:[/bold]
+        $ gnosisllm-knowledge agentic status
+    """
+    from gnosisllm_knowledge.cli.commands.agentic import agentic_status_command
+
+    asyncio.run(agentic_status_command(display=display))
+
+
+def main() -> None:
+    """CLI entry point."""
+    app()
+
+
+if __name__ == "__main__":
+    main()

gnosisllm_knowledge/cli/commands/__init__.py

@@ -0,0 +1,7 @@
+"""CLI commands for gnosisllm-knowledge."""
+
+from gnosisllm_knowledge.cli.commands.load import load_command
+from gnosisllm_knowledge.cli.commands.search import search_command
+from gnosisllm_knowledge.cli.commands.setup import setup_command
+
+__all__ = ["setup_command", "load_command", "search_command"]