synth-ai 0.2.14__py3-none-any.whl → 0.4.4__py3-none-any.whl

synth_ai/cli/commands/scan/core.py

@@ -0,0 +1,344 @@
+"""Main scan command implementation."""
+
+from __future__ import annotations
+
+import asyncio
+import json
+import os
+from pathlib import Path
+
+import click
+
+from synth_ai.cli.commands.scan.cloudflare_scanner import scan_cloudflare_apps
+from synth_ai.cli.commands.scan.local_scanner import scan_local_ports, scan_registry
+from synth_ai.cli.commands.scan.models import ScannedApp
+from synth_ai.cli.lib.env import resolve_env_var
+
+
+def format_app_table(apps: list[ScannedApp]) -> str:
+    """Format apps as a human-readable table."""
+    if not apps:
+        return "No active task apps found."
+
+    # Calculate column widths
+    name_width = max(len(app.name) for app in apps if app.name) if apps else 4
+    port_width = 5  # "Port" header
+    status_width = 10  # Status with icon
+    type_width = max(len(app.type) for app in apps) if apps else 4
+    via_width = max(len(app.discovered_via) for app in apps) if apps else 13
+    
+    # Additional metadata columns
+    app_id_width = max(len(app.app_id or "") for app in apps) if apps else 0
+    app_id_width = max(app_id_width, 7)  # "App ID" header
+    
+    version_width = max(len(app.version or "") for app in apps) if apps else 0
+    version_width = max(version_width, 7)  # "Version" header
+
+    # Ensure minimum widths
+    name_width = max(name_width, 4)
+    type_width = max(type_width, 4)
+    via_width = max(via_width, 13)
+
+    lines = [f"Found {len(apps)} active task app{'s' if len(apps) != 1 else ''}:\n"]
+    
+    # Header row
+    header_parts = [
+        f"{'Name':<{name_width}}",
+        f"{'Port':<{port_width}}",
+        f"{'Status':<{status_width}}",
+        f"{'Type':<{type_width}}",
+    ]
+    if app_id_width > 0:
+        header_parts.append(f"{'App ID':<{app_id_width}}")
+    if version_width > 0:
+        header_parts.append(f"{'Version':<{version_width}}")
+    header_parts.append(f"{'Discovered Via':<{via_width}}")
+    
+    lines.append(" ".join(header_parts))
+    
+    # Separator
+    total_width = sum(len(p) for p in header_parts) + len(header_parts) - 1
+    lines.append("─" * total_width)
+
+    for app in apps:
+        status_icon = "✅" if app.health_status == "healthy" else "⚠️ " if app.health_status == "unhealthy" else "❓"
+        status_display = f"{status_icon} {app.health_status}"
+        
+        port_str = str(app.port) if app.port else "-"
+        
+        row_parts = [
+            f"{app.name:<{name_width}}",
+            f"{port_str:<{port_width}}",
+            f"{status_display:<{status_width}}",
+            f"{app.type:<{type_width}}",
+        ]
+        if app_id_width > 0:
+            row_parts.append(f"{(app.app_id or '-'):<{app_id_width}}")
+        if version_width > 0:
+            row_parts.append(f"{(app.version or '-'):<{version_width}}")
+        row_parts.append(f"{app.discovered_via:<{via_width}}")
+        
+        lines.append(" ".join(row_parts))
+
+    return "\n".join(lines)
+
+
+def format_app_json(apps: list[ScannedApp]) -> str:
+    """Format apps as JSON."""
+    apps_data = []
+    for app in apps:
+        apps_data.append(
+            {
+                "name": app.name,
+                "url": app.url,
+                "type": app.type,
+                "health_status": app.health_status,
+                "port": app.port,
+                "tunnel_mode": app.tunnel_mode,
+                "tunnel_hostname": app.tunnel_hostname,
+                "app_id": app.app_id,
+                "task_name": app.task_name,
+                "dataset_id": app.dataset_id,
+                "version": app.version,
+                "metadata": app.metadata,
+                "discovered_via": app.discovered_via,
+            }
+        )
+
+    summary = {
+        "total_found": len(apps),
+        "healthy": sum(1 for app in apps if app.health_status == "healthy"),
+        "unhealthy": sum(1 for app in apps if app.health_status == "unhealthy"),
+        "local_count": sum(1 for app in apps if app.type == "local"),
+        "cloudflare_count": sum(1 for app in apps if app.type == "cloudflare"),
+    }
+
+    return json.dumps({"apps": apps_data, "scan_summary": summary}, indent=2)
+
+
+async def run_scan(
+    port_range: tuple[int, int],
+    timeout: float,
+    api_key: str | None,
+    env_file: Path | None,
+    verbose: bool,
+) -> list[ScannedApp]:
+    """Run the scan operation.
+
+    Args:
+        port_range: Tuple of (start_port, end_port)
+        timeout: Health check timeout
+        api_key: API key for health checks
+        env_file: Specific .env file to check
+        verbose: Show detailed progress
+
+    Returns:
+        List of discovered apps
+    """
+    start_port, end_port = port_range
+    all_apps: list[ScannedApp] = []
+
+    # Resolve API key
+    env_api_key = api_key
+    if not env_api_key:
+        try:
+            env_api_key = resolve_env_var("ENVIRONMENT_API_KEY")
+        except Exception:
+            env_api_key = os.getenv("ENVIRONMENT_API_KEY")
+
+    if verbose:
+        click.echo(f"Scanning ports {start_port}-{end_port}...", err=True)
+
+    # Scan local ports
+    local_apps = await scan_local_ports(start_port, end_port, env_api_key, timeout)
+    all_apps.extend(local_apps)
+
+    if verbose:
+        click.echo(f"Found {len(local_apps)} local app(s)", err=True)
+
+    # Scan Cloudflare apps
+    synth_api_key = os.getenv("SYNTH_API_KEY")
+    cloudflare_apps = await scan_cloudflare_apps(synth_api_key, env_api_key, env_file, timeout)
+    all_apps.extend(cloudflare_apps)
+
+    if verbose:
+        click.echo(f"Found {len(cloudflare_apps)} Cloudflare app(s)", err=True)
+
+    # Scan service records (local services deployed via synth-ai)
+    from synth_ai.cli.commands.scan.local_scanner import scan_service_records
+    service_record_apps = await scan_service_records(env_api_key, timeout)
+    all_apps.extend(service_record_apps)
+
+    if verbose:
+        click.echo(f"Found {len(service_record_apps)} service record(s)", err=True)
+
+    # Scan registry (for reference, but these don't have URLs)
+    registry_apps = scan_registry()
+    # Only add registry apps that weren't already discovered
+    discovered_app_ids = {app.app_id for app in all_apps if app.app_id}
+    new_registry_apps = [app for app in registry_apps if app.app_id and app.app_id not in discovered_app_ids]
+    all_apps.extend(new_registry_apps)
+
+    if verbose:
+        click.echo(f"Found {len(new_registry_apps)} registry app(s) not yet running", err=True)
+
+    # Deduplicate apps by URL (prefer better sources and metadata)
+    seen_urls: dict[str, ScannedApp] = {}
+    for app in all_apps:
+        if app.url and app.url in seen_urls:
+            # Prefer apps with better discovery method and metadata
+            existing = seen_urls[app.url]
+            
+            # Priority order for discovery methods:
+            # 1. service_records / tunnel_records (have metadata like app_id, task_app_path)
+            # 2. cloudflared_process / backend_api (tunnels)
+            # 3. port_scan (basic discovery)
+            # 4. registry (no URL)
+            
+            priority_order = {
+                "service_records": 1,
+                "tunnel_records": 1,
+                "backend_api": 2,
+                "cloudflared_process": 2,
+                "port_scan": 3,
+                "registry": 4,
+            }
+            
+            app_priority = priority_order.get(app.discovered_via, 99)
+            existing_priority = priority_order.get(existing.discovered_via, 99)
+            
+            # Prefer higher priority discovery method
+            if app_priority < existing_priority:
+                seen_urls[app.url] = app
+            elif app_priority == existing_priority:
+                # Same priority - prefer better health status or better metadata
+                if app.health_status != "unknown" and existing.health_status == "unknown":
+                    seen_urls[app.url] = app
+                elif app.app_id and not existing.app_id:
+                    # Prefer app with app_id
+                    seen_urls[app.url] = app
+                elif app.discovered_via != "registry" and existing.discovered_via == "registry":
+                    seen_urls[app.url] = app
+        elif app.url:
+            seen_urls[app.url] = app
+        elif app.discovered_via == "registry":
+            # Registry apps without URLs - add them separately
+            seen_urls[f"registry:{app.app_id}"] = app
+
+    return list(seen_urls.values())
+
+
+@click.command("scan")
+@click.option(
+    "--port-range",
+    type=str,
+    default="8000:8100",
+    help="Port range to scan for local apps (format: START:END)",
+)
+@click.option(
+    "--timeout",
+    type=float,
+    default=2.0,
+    show_default=True,
+    help="Health check timeout in seconds",
+)
+@click.option(
+    "--api-key",
+    type=str,
+    default=None,
+    envvar="ENVIRONMENT_API_KEY",
+    help="API key for health checks (default: from ENVIRONMENT_API_KEY env var)",
+)
+@click.option("--json", "output_json", is_flag=True, help="Output results as JSON")
+@click.option("--verbose", is_flag=True, help="Show detailed scanning progress")
+@click.option(
+    "--env-file",
+    type=click.Path(exists=True, dir_okay=False, path_type=Path),
+    default=None,
+    help="(Deprecated) Not used - tunnels are discovered from running processes",
+    hidden=True,
+)
+def scan_command(
+    port_range: str,
+    timeout: float,
+    api_key: str | None,
+    output_json: bool,
+    verbose: bool,
+    env_file: Path | None,
+) -> None:
+    """Scan for active Cloudflare and local task apps.
+
+    Discovers and performs health checks on running task applications deployed
+    locally or via Cloudflare tunnels. Returns structured information in table
+    or JSON format, suitable for terminal use or programmatic consumption by
+    CLI agents and automation tools.
+
+    Discovery Methods:
+        - Port scanning: Scans specified port range for local HTTP servers
+        - Service records: Reads deployed local services from persistent records
+        - Tunnel records: Reads deployed Cloudflare tunnels from persistent records
+        - Process scanning: Inspects running cloudflared processes for tunnel URLs
+        - Backend API: Queries backend for managed tunnel information
+        - Registry: Checks task app registry for registered apps
+
+    Health Checks:
+        - Performs HTTP GET requests to /health endpoints
+        - Extracts metadata from /info endpoints (app_id, version, task_name, etc.)
+        - Supports API key authentication via X-API-Key header
+
+    Output Formats:
+        - Table (default): Human-readable table with columns for name, port, status,
+          type, app ID, version, and discovery method
+        - JSON (--json): Machine-readable JSON with full metadata and scan summary
+
+    Examples:
+        # Scan default port range (8000-8100) and show table
+        $ synth-ai scan
+
+        # Scan specific port range with verbose output
+        $ synth-ai scan --port-range 8000:9000 --verbose
+
+        # Get JSON output for programmatic use
+        $ synth-ai scan --json
+
+        # Use custom API key and timeout
+        $ synth-ai scan --api-key YOUR_KEY --timeout 5.0
+
+    Args:
+        port_range: Port range to scan (format: START:END, e.g., "8000:8100")
+        timeout: Health check timeout in seconds (default: 2.0)
+        api_key: API key for health checks (default: from ENVIRONMENT_API_KEY env var)
+        output_json: Output results as JSON instead of table
+        verbose: Show detailed scanning progress
+        env_file: (Deprecated) Not used - tunnels are discovered from processes
+
+    Returns:
+        None (outputs to stdout)
+
+    Raises:
+        click.BadParameter: If port range format is invalid or out of valid range
+    """
+    # Parse port range
+    try:
+        if ":" in port_range:
+            start_str, end_str = port_range.split(":", 1)
+            start_port = int(start_str.strip())
+            end_port = int(end_str.strip())
+        else:
+            start_port = int(port_range.strip())
+            end_port = start_port
+    except ValueError as e:
+        raise click.BadParameter(f"Invalid port range format: {port_range}. Use START:END (e.g., 8000:8100)") from e
+
+    if start_port < 1 or end_port > 65535 or start_port > end_port:
+        raise click.BadParameter(f"Invalid port range: {start_port}-{end_port}")
+
+    # Run scan
+    apps = asyncio.run(run_scan((start_port, end_port), timeout, api_key, env_file, verbose))
+
+    # Output results
+    if output_json:
+        click.echo(format_app_json(apps))
+    else:
+        click.echo(format_app_table(apps))
+

synth_ai/cli/commands/scan/health_checker.py

@@ -0,0 +1,242 @@
+"""Health check utilities for scan command."""
+
+from __future__ import annotations
+
+import asyncio
+from typing import Any, Literal
+
+import httpx
+
+
+async def check_app_health(
+    url: str,
+    api_key: str | None,
+    timeout: float = 2.0,
+) -> tuple[Literal["healthy", "unhealthy", "unknown"], dict[str, Any]]:
+    """Check health and fetch metadata from a task app.
+
+    Performs HTTP health checks on a task app by:
+    1. Checking the /health endpoint for health status
+    2. Fetching the /info endpoint for metadata (app_id, version, etc.)
+
+    Health Status Determination:
+        - "healthy": HTTP 200 with valid JSON response containing "status": "healthy"
+            or "healthy": true, or HTTP 200 with any valid JSON
+        - "unhealthy": HTTP error status (4xx, 5xx), Cloudflare tunnel errors (530, 502),
+            or HTML responses (Cloudflare error pages)
+        - "unknown": Request timeout, connection errors, or other exceptions
+
+    Args:
+        url: Base URL of the task app (e.g., "http://localhost:8000" or
+            "https://abc123.trycloudflare.com"). Trailing slashes are handled.
+        api_key: API key for authentication via X-API-Key header. If None, requests
+            are made without authentication (may fail for apps requiring auth).
+        timeout: Request timeout in seconds for both /health and /info endpoints.
+            Default is 2.0 seconds.
+
+    Returns:
+        Tuple of (health_status, metadata_dict) where:
+            - health_status: One of "healthy", "unhealthy", or "unknown"
+            - metadata_dict: Dictionary containing:
+                - Full /info endpoint response (if available)
+                - Error information (if health check failed)
+                - HTTP status codes and error messages
+
+    Examples:
+        >>> import asyncio
+        >>> status, metadata = asyncio.run(check_app_health("http://localhost:8000", "key123"))
+        >>> print(status)  # "healthy", "unhealthy", or "unknown"
+        >>> print(metadata.get("service", {}).get("task", {}).get("id"))  # app_id
+
+    Raises:
+        No exceptions are raised - all errors are captured and returned as "unknown"
+        status with error details in metadata.
+    """
+    headers: dict[str, str] = {}
+    if api_key:
+        headers["X-API-Key"] = api_key
+
+    metadata: dict[str, Any] = {}
+
+    # Check /health endpoint
+    health_status: Literal["healthy", "unhealthy", "unknown"] = "unknown"
+    try:
+        async with httpx.AsyncClient(timeout=timeout, follow_redirects=True) as client:
+            health_url = f"{url.rstrip('/')}/health"
+            health_resp = await client.get(health_url, headers=headers)
+            
+            # Check for Cloudflare tunnel errors (530, 502, etc.) or HTML responses
+            if health_resp.status_code in (530, 502, 503, 504):
+                # Cloudflare error - tunnel pointing to dead server
+                health_status = "unhealthy"
+                metadata["tunnel_error"] = f"HTTP {health_resp.status_code}"
+            elif "text/html" in health_resp.headers.get("content-type", "").lower():
+                # HTML response (Cloudflare error page) - tunnel not working
+                health_status = "unhealthy"
+                metadata["tunnel_error"] = "HTML response (tunnel error)"
+            elif health_resp.status_code == 200:
+                try:
+                    health_data = health_resp.json()
+                    if isinstance(health_data, dict):
+                        # Check for "status": "healthy" or "healthy": true
+                        status = health_data.get("status", "").lower()
+                        healthy_flag = health_data.get("healthy")
+                        if status == "healthy" or healthy_flag is True:
+                            health_status = "healthy"
+                        elif status == "unhealthy" or healthy_flag is False:
+                            health_status = "unhealthy"
+                        else:
+                            # If status is 200 but no clear health indicator, consider healthy
+                            health_status = "healthy"
+                    else:
+                        # If JSON parsing fails but status is 200, consider it healthy
+                        health_status = "healthy"
+                except Exception:
+                    # If JSON parsing fails but status is 200, consider it healthy
+                    health_status = "healthy"
+            else:
+                health_status = "unhealthy"
+                metadata["http_status"] = health_resp.status_code
+    except httpx.TimeoutException:
+        health_status = "unknown"
+        metadata["error"] = "timeout"
+    except Exception as exc:
+        health_status = "unknown"
+        metadata["error"] = str(exc)
+
+    # Fetch /info endpoint for metadata
+    try:
+        async with httpx.AsyncClient(timeout=timeout) as client:
+            info_url = f"{url.rstrip('/')}/info"
+            info_resp = await client.get(info_url, headers=headers)
+            if info_resp.status_code == 200:
+                try:
+                    info_data = info_resp.json()
+                    if isinstance(info_data, dict):
+                        metadata.update(info_data)
+                except Exception:
+                    pass
+    except Exception:
+        pass
+
+    return health_status, metadata
+
+
+def extract_app_info(metadata: dict[str, Any]) -> tuple[str | None, str | None, str | None, str | None]:
+    """Extract app information from /info endpoint metadata.
+
+    Parses the metadata dictionary (typically from /info endpoint response) to
+    extract key application identifiers and version information. Handles missing
+    or malformed data gracefully by returning None for unavailable fields.
+
+    Metadata Structure Expected:
+        {
+            "service": {
+                "task": {
+                    "id": "app_id",
+                    "name": "Task Name",
+                    "version": "1.0.0"
+                }
+            },
+            "dataset": {
+                "id": "dataset_id"
+            }
+        }
+
+    Args:
+        metadata: Dictionary containing /info endpoint response or service records.
+            May contain nested "service" and "dataset" dictionaries.
+
+    Returns:
+        Tuple of (app_id, task_name, dataset_id, version) where:
+            - app_id: Task app identifier from service.task.id
+            - task_name: Human-readable task name from service.task.name
+            - dataset_id: Dataset identifier from dataset.id (if available)
+            - version: App version from service.task.version
+
+    Examples:
+        >>> metadata = {
+        ...     "service": {"task": {"id": "banking77", "name": "Banking77", "version": "1.0.0"}},
+        ...     "dataset": {"id": "banking77_dataset"}
+        ... }
+        >>> app_id, task_name, dataset_id, version = extract_app_info(metadata)
+        >>> print(app_id)  # "banking77"
+        >>> print(version)  # "1.0.0"
+    """
+    app_id: str | None = None
+    task_name: str | None = None
+    dataset_id: str | None = None
+    version: str | None = None
+
+    try:
+        service = metadata.get("service", {})
+        if isinstance(service, dict):
+            task = service.get("task", {})
+            if isinstance(task, dict):
+                app_id = task.get("id")
+                task_name = task.get("name")
+                version = task.get("version")
+
+        dataset = metadata.get("dataset", {})
+        if isinstance(dataset, dict):
+            dataset_id = dataset.get("id")
+    except Exception:
+        pass
+
+    return app_id, task_name, dataset_id, version
+
+
+async def check_multiple_apps_health(
+    urls: list[str],
+    api_key: str | None,
+    timeout: float = 2.0,
+    max_concurrent: int = 10,
+) -> dict[str, tuple[Literal["healthy", "unhealthy", "unknown"], dict[str, Any]]]:
+    """Check health for multiple apps concurrently.
+
+    Performs health checks on multiple task apps in parallel using asyncio
+    semaphores to limit concurrent requests. This is more efficient than
+    checking apps sequentially, especially when scanning many ports or services.
+
+    Concurrency Control:
+        Uses asyncio.Semaphore to limit the number of concurrent HTTP requests,
+        preventing resource exhaustion when checking many apps simultaneously.
+
+    Args:
+        urls: List of app URLs to check (e.g., ["http://localhost:8000", ...]).
+            Each URL will be checked independently.
+        api_key: API key for authentication via X-API-Key header. Applied to
+            all requests. If None, requests are made without authentication.
+        timeout: Request timeout in seconds per app. Default is 2.0 seconds.
+        max_concurrent: Maximum number of concurrent HTTP requests. Default is 10.
+            Increase for faster scanning, decrease to reduce resource usage.
+
+    Returns:
+        Dictionary mapping URL -> (health_status, metadata) where:
+            - Keys are the input URLs
+            - Values are tuples of (health_status, metadata_dict) as returned
+              by check_app_health()
+
+    Examples:
+        >>> import asyncio
+        >>> urls = ["http://localhost:8000", "http://localhost:8001"]
+        >>> results = asyncio.run(check_multiple_apps_health(urls, "key123"))
+        >>> for url, (status, metadata) in results.items():
+        ...     print(f"{url}: {status}")
+
+    Note:
+        Exceptions during individual health checks are caught and result in
+        "unknown" status. The function never raises exceptions.
+    """
+    semaphore = asyncio.Semaphore(max_concurrent)
+    results: dict[str, tuple[Literal["healthy", "unhealthy", "unknown"], dict[str, Any]]] = {}
+
+    async def check_one(url: str) -> None:
+        async with semaphore:
+            status, metadata = await check_app_health(url, api_key, timeout)
+            results[url] = (status, metadata)
+
+    await asyncio.gather(*[check_one(url) for url in urls], return_exceptions=True)
+
+    return results
+