Orchestrator/privat/NoteGraph/ingestion/cli.py

"""Click-based CLI for the NoteGraph ingestion pipeline.

Subcommands:
  import  - Import files or directories into NoteGraph
  watch   - Start drop-folder watcher for continuous ingestion
  sync    - One-shot Nextcloud inbox sync
  upload  - Start web upload server
"""

import logging
import sys
from pathlib import Path

import click

from ingestion.config import IngestionConfig


def _setup_logging(verbose: bool) -> None:
    """Configure logging based on verbosity."""
    level = logging.DEBUG if verbose else logging.INFO
    logging.basicConfig(
        level=level,
        format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
        datefmt="%H:%M:%S",
    )


@click.group()
@click.version_option(version="0.1.0", prog_name="notegraph-ingest")
def cli():
    """NoteGraph Ingestion — import documents into your knowledge base."""
    pass


@cli.command("import")
@click.argument("path", type=click.Path(exists=True))
@click.option("--output-dir", "-o", type=click.Path(), default=None, help="Override output directory")
@click.option("--create-tasks", is_flag=True, help="Create OrgMyLife tasks for action items")
@click.option("--provider", type=str, default=None, help="Override LLM provider (openai, anthropic, etc.)")
@click.option("--dry-run", is_flag=True, help="Show what would be created without writing files")
@click.option("--no-commit", is_flag=True, help="Skip git auto-commit")
@click.option("--verbose", "-v", is_flag=True, help="Enable debug logging")
def import_cmd(path, output_dir, create_tasks, provider, dry_run, no_commit, verbose):
    """Import files or directories into NoteGraph.

    PATH can be a single file or a directory to scan recursively.
    """
    _setup_logging(verbose)

    config = IngestionConfig()
    if provider:
        config.llm_provider = provider

    config.validate_api_key()

    from ingestion.discovery import discover_files
    from ingestion.pipeline import process_batch

    source_path = Path(path)
    files = discover_files(source_path)

    if not files:
        click.echo("No supported files found.")
        sys.exit(0)

    click.echo(f"Found {len(files)} file(s) to process.")

    if dry_run:
        click.echo("[DRY RUN] No files will be written.\n")

    def progress(current, total, filename):
        click.echo(f"  [{current}/{total}] {filename}")

    result = process_batch(
        paths=files,
        config=config,
        dry_run=dry_run,
        no_commit=no_commit,
        create_tasks=create_tasks,
        output_dir=Path(output_dir) if output_dir else None,
        progress_callback=progress,
    )

    # Summary
    click.echo("")
    click.echo(f"Done! Processed {result.total} file(s):")
    click.echo(f"  ✓ Success: {result.success}")
    if result.failed:
        click.echo(f"  ✗ Failed:  {result.failed}")


@cli.command("watch")
@click.option("--verbose", "-v", is_flag=True, help="Enable debug logging")
def watch_cmd(verbose):
    """Start drop-folder watcher for continuous ingestion."""
    _setup_logging(verbose)

    from ingestion.watcher import start_watcher

    config = IngestionConfig()
    config.validate_api_key()

    click.echo(f"Watching inbox: {config.inbox_dir}")
    click.echo("Press Ctrl+C to stop.\n")

    try:
        start_watcher(config)
    except KeyboardInterrupt:
        click.echo("\nWatcher stopped.")


@cli.command("sync")
@click.option("--verbose", "-v", is_flag=True, help="Enable debug logging")
def sync_cmd(verbose):
    """One-shot Nextcloud inbox sync."""
    _setup_logging(verbose)

    from ingestion.integrations.nextcloud import NextcloudInboxSync

    config = IngestionConfig()

    if not config.nextcloud_inbox_url:
        click.echo("ERROR: NEXTCLOUD_INBOX_URL not configured.")
        sys.exit(1)

    config.validate_api_key()

    sync = NextcloudInboxSync(config)
    downloaded = sync.poll()

    if not downloaded:
        click.echo("No new files in Nextcloud inbox.")
        return

    click.echo(f"Downloaded {len(downloaded)} file(s) from Nextcloud.")

    # Process downloaded files
    from ingestion.pipeline import process_batch

    result = process_batch(
        paths=downloaded,
        config=config,
    )

    # Mark processed files on Nextcloud
    for r in result.results:
        if r.success:
            sync.mark_processed(r.source_file.name)

    click.echo(f"Processed: {result.success} success, {result.failed} failed.")


@cli.command("upload")
@click.option("--host", default="0.0.0.0", help="Bind host")
@click.option("--port", default=8001, type=int, help="Bind port")
@click.option("--verbose", "-v", is_flag=True, help="Enable debug logging")
def upload_cmd(host, port, verbose):
    """Start web upload server for mobile file ingestion."""
    _setup_logging(verbose)

    import uvicorn

    click.echo(f"Starting upload server on {host}:{port}")
    uvicorn.run("ingestion.web.upload:app", host=host, port=port, log_level="info")


if __name__ == "__main__":
    cli()