griptape-nodes 0.44.0__py3-none-any.whl → 0.45.0__py3-none-any.whl

griptape_nodes/retained_mode/managers/sync_manager.py

@@ -0,0 +1,498 @@
+from __future__ import annotations
+
+import hashlib
+import logging
+import threading
+import uuid
+from collections import defaultdict
+from concurrent.futures import ThreadPoolExecutor
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+from watchfiles import Change, PythonFilter, watch
+
+from griptape_nodes.drivers.storage.griptape_cloud_storage_driver import GriptapeCloudStorageDriver
+from griptape_nodes.retained_mode.events.app_events import AppInitializationComplete
+from griptape_nodes.retained_mode.events.base_events import AppEvent
+from griptape_nodes.retained_mode.events.sync_events import (
+    StartSyncAllCloudWorkflowsRequest,
+    StartSyncAllCloudWorkflowsResultFailure,
+    StartSyncAllCloudWorkflowsResultSuccess,
+    SyncComplete,
+)
+from griptape_nodes.retained_mode.events.workflow_events import (
+    RegisterWorkflowsFromConfigRequest,
+    RegisterWorkflowsFromConfigResultSuccess,
+)
+
+if TYPE_CHECKING:
+    from griptape_nodes.retained_mode.events.base_events import ResultPayload
+    from griptape_nodes.retained_mode.managers.config_manager import ConfigManager
+    from griptape_nodes.retained_mode.managers.event_manager import EventManager
+
+
+logger = logging.getLogger("griptape_nodes")
+
+
+class SyncManager:
+    """Manager for syncing workflows with cloud storage."""
+
+    def __init__(self, event_manager: EventManager, config_manager: ConfigManager) -> None:
+        self._active_sync_tasks: dict[str, threading.Thread] = {}
+        self._watch_task: threading.Thread | None = None
+        self._watching_stopped = threading.Event()
+        self._config_manager = config_manager
+        self._event_manager = event_manager
+
+        # Hash tracking to prevent sync loops
+        self._file_hashes = defaultdict(str)
+        self._hash_lock = threading.Lock()
+
+        # Initialize sync directory
+        self._sync_dir = self._config_manager.workspace_path / self._config_manager.get_config_value(
+            "synced_workflows_directory"
+        )
+        self._sync_dir.mkdir(parents=True, exist_ok=True)
+
+        event_manager.assign_manager_to_request_type(
+            StartSyncAllCloudWorkflowsRequest,
+            self.on_start_sync_all_cloud_workflows_request,
+        )
+
+        event_manager.add_listener_to_app_event(
+            AppInitializationComplete,
+            self.on_app_initialization_complete,
+        )
+
+    def _set_expected_hash(self, path: str | Path, content: bytes) -> None:
+        """Set the expected SHA256 hash for a file we're about to write.
+
+        This should be called before writing a file to establish what content
+        we expect to see, allowing us to distinguish our own writes from
+        external modifications.
+
+        Args:
+            path: Path to the file that will be written
+            content: The exact bytes that will be written to the file
+        """
+        path_str = str(Path(path).resolve())
+        file_hash = hashlib.sha256(content).hexdigest()
+        with self._hash_lock:
+            self._file_hashes[path_str] = file_hash
+            logger.debug("Set expected hash for %s: %s", path_str, file_hash[:8])
+
+    def _is_expected_content(self, path: str) -> bool:
+        """Check if file content matches what we wrote (indicating a self-triggered event).
+
+        This prevents sync loops by identifying when a file system event was caused
+        by our own write operation rather than an external modification.
+
+        Args:
+            path: Path to the file to check
+
+        Returns:
+            True if the file content matches our expected hash (self-triggered event),
+            False if it doesn't match or no expected hash exists (external change)
+        """
+        path_str = str(Path(path).resolve())
+        with self._hash_lock:
+            expected_hash = self._file_hashes.get(path_str)
+
+        if not expected_hash:
+            # No expected hash means this wasn't a self-triggered event
+            return False
+
+        try:
+            actual_hash = hashlib.sha256(Path(path).read_bytes()).hexdigest()
+
+            if actual_hash == expected_hash:
+                # Content matches - this was our own write, clean up and ignore
+                with self._hash_lock:
+                    self._file_hashes.pop(path_str, None)
+                logger.debug("File content matches expected hash for %s", path_str)
+                return True
+            # Content doesn't match - this is an external modification
+            logger.debug("File content does not match expected hash for %s", path_str)
+        except Exception as e:
+            logger.debug("Error checking file hash for %s: %s", path_str, str(e))
+
+        return False
+
+    def on_start_sync_all_cloud_workflows_request(self, _request: StartSyncAllCloudWorkflowsRequest) -> ResultPayload:
+        """Start syncing all cloud workflows to local synced_workflows directory."""
+        try:
+            storage_driver = self._get_cloud_storage_driver()
+            sync_dir = self._sync_dir
+
+            # List all assets in the bucket to get count
+            files = storage_driver.list_files()
+            workflow_files = [file for file in files if file.endswith(".py")]
+
+            if not workflow_files:
+                logger.info("No workflow files found in cloud storage")
+                return StartSyncAllCloudWorkflowsResultSuccess(sync_directory=str(sync_dir), total_workflows=0)
+
+            # Start background sync with unique ID
+            sync_task_id = str(uuid.uuid4())
+            sync_thread = threading.Thread(
+                target=self._sync_workflows_background,
+                args=(sync_task_id, workflow_files, storage_driver, sync_dir),
+                name=f"SyncWorkflows-{sync_task_id}",
+                daemon=True,
+            )
+
+            self._active_sync_tasks[sync_task_id] = sync_thread
+            sync_thread.start()
+        except Exception as e:
+            logger.error("Failed to start cloud workflow sync: %s", str(e))
+            return StartSyncAllCloudWorkflowsResultFailure()
+        else:
+            logger.info("Started background sync for %d workflow files", len(workflow_files))
+            return StartSyncAllCloudWorkflowsResultSuccess(
+                sync_directory=str(sync_dir), total_workflows=len(workflow_files)
+            )
+
+    def on_app_initialization_complete(self, _payload: AppInitializationComplete) -> None:
+        """Automatically start syncing cloud workflows when the app initializes."""
+        try:
+            # Check if cloud storage is configured before attempting sync
+            self._get_cloud_storage_driver()
+            # Start file watching after successful sync
+            self._start_file_watching()
+
+            logger.info("App initialization complete - starting automatic cloud workflow sync")
+
+            # Create and handle the sync request
+            sync_request = StartSyncAllCloudWorkflowsRequest()
+
+            # Use handle_request to process through normal event system
+            from griptape_nodes.retained_mode.griptape_nodes import GriptapeNodes
+
+            result = GriptapeNodes.handle_request(sync_request)
+
+            if isinstance(result, StartSyncAllCloudWorkflowsResultSuccess):
+                logger.info(
+                    "Automatic cloud workflow sync started successfully - %d workflows will be synced to %s",
+                    result.total_workflows,
+                    result.sync_directory,
+                )
+
+            else:
+                logger.debug("Automatic cloud workflow sync failed to start (likely cloud not configured)")
+
+        except Exception as e:
+            logger.debug("Automatic cloud workflow sync skipped: %s", str(e))
+
+    def _get_cloud_storage_driver(self) -> GriptapeCloudStorageDriver:
+        """Get configured cloud storage driver.
+
+        Returns:
+            Configured GriptapeCloudStorageDriver instance.
+
+        Raises:
+            RuntimeError: If required cloud configuration is missing.
+        """
+        from griptape_nodes.retained_mode.griptape_nodes import GriptapeNodes
+
+        secrets_manager = GriptapeNodes.SecretsManager()
+
+        # Get cloud storage configuration from secrets
+        bucket_id = secrets_manager.get_secret("GT_CLOUD_BUCKET_ID", should_error_on_not_found=False)
+        base_url = secrets_manager.get_secret("GT_CLOUD_BASE_URL", should_error_on_not_found=False)
+        api_key = secrets_manager.get_secret("GT_CLOUD_API_KEY")
+
+        if not bucket_id:
+            msg = "Cloud storage bucket_id not configured. Set GT_CLOUD_BUCKET_ID secret."
+            raise RuntimeError(msg)
+        if not api_key:
+            msg = "Cloud storage api_key not configured. Set GT_CLOUD_API_KEY secret."
+            raise RuntimeError(msg)
+
+        return GriptapeCloudStorageDriver(
+            bucket_id=bucket_id,
+            base_url=base_url,
+            api_key=api_key,
+            static_files_directory=self._config_manager.get_config_value(
+                "synced_workflows_directory", default="synced_workflows"
+            ),
+        )
+
+    def _download_cloud_workflow_to_sync_dir(self, filename: str) -> bool:
+        """Download a workflow file from cloud storage to the sync directory.
+
+        Args:
+            filename: Name of the workflow file to download from cloud
+
+        Returns:
+            True if download was successful, False otherwise
+        """
+        try:
+            storage_driver = self._get_cloud_storage_driver()
+            sync_dir = self._sync_dir
+
+            # Download file content from cloud
+            file_content = storage_driver.download_file(filename)
+
+            # Write to local sync directory
+            local_file_path = sync_dir / filename
+
+            # Check if file exists and has same content hash
+            if local_file_path.exists():
+                try:
+                    existing_content = local_file_path.read_bytes()
+
+                    existing_hash = hashlib.sha256(existing_content).hexdigest()
+                    new_hash = hashlib.sha256(file_content).hexdigest()
+
+                    if existing_hash == new_hash:
+                        logger.debug("Skipping write - file already has same content hash: %s", filename)
+                        return True
+                except Exception as e:
+                    logger.debug("Error checking existing file hash for %s: %s", filename, str(e))
+
+            # Set expected hash before writing to prevent sync loops
+            self._set_expected_hash(local_file_path, file_content)
+
+            local_file_path.write_bytes(file_content)
+
+            logger.info("Successfully downloaded cloud workflow to sync directory: %s", filename)
+        except Exception as e:
+            logger.error("Failed to download cloud workflow '%s': %s", filename, str(e))
+            return False
+        else:
+            return True
+
+    def _upload_workflow_file(self, file_path: Path) -> None:
+        """Upload a single workflow file to cloud storage.
+
+        Args:
+            file_path: Path to the workflow file to upload.
+        """
+        try:
+            # Check if valid workflow file
+            if not file_path.name.endswith(".py"):
+                logger.error("Invalid workflow file path: %s", file_path)
+                return
+
+            # Proceed with upload
+            storage_driver = self._get_cloud_storage_driver()
+
+            # Read file content
+            file_content = file_path.read_bytes()
+
+            # Upload to cloud storage using the upload_file method
+            filename = file_path.name
+            storage_driver.upload_file(filename, file_content)
+
+            logger.info("Successfully uploaded workflow file to cloud: %s", filename)
+
+        except Exception as e:
+            logger.error("Failed to upload workflow file '%s': %s", file_path.name, str(e))
+
+    def _delete_workflow_file(self, file_path: Path) -> None:
+        """Delete a workflow file from cloud storage.
+
+        Args:
+            file_path: Path to the workflow file that was deleted locally.
+        """
+        try:
+            storage_driver = self._get_cloud_storage_driver()
+            filename = file_path.name
+
+            # Use the storage driver's delete method
+            storage_driver.delete_file(filename)
+            logger.info("Successfully deleted workflow file from cloud: %s", filename)
+
+        except Exception as e:
+            logger.error("Failed to delete workflow file '%s' from cloud: %s", file_path.name, str(e))
+
+    def _start_file_watching(self) -> None:
+        """Start watching the synced_workflows directory for changes."""
+        try:
+            sync_dir = self._sync_dir
+
+            # Stop any existing watching
+            if self._watch_task and self._watch_task.is_alive():
+                self._watching_stopped.set()
+                self._watch_task.join(timeout=2.0)
+
+            # Reset the stop event for new watching
+            self._watching_stopped.clear()
+
+            # Start new watching thread
+            self._watch_task = threading.Thread(
+                target=self._watch_files_thread,
+                args=(str(sync_dir),),
+                name="WatchFiles-SyncManager",
+                daemon=True,
+            )
+            self._watch_task.start()
+
+            logger.info("Started watching synced workflows directory: %s", sync_dir)
+
+        except Exception as e:
+            logger.error("Failed to start file watching: %s", str(e))
+
+    def _watch_files_thread(self, sync_dir: str) -> None:
+        """Background thread that watches for file changes using watchfiles."""
+        try:
+            logger.debug("File watching thread started for directory: %s", sync_dir)
+
+            # Watch for changes in the sync directory using PythonFilter
+            for changes in watch(sync_dir, watch_filter=PythonFilter(), stop_event=self._watching_stopped):
+                if self._watching_stopped.is_set():
+                    break
+
+                for change, path_str in changes:
+                    path = Path(path_str)
+
+                    # Check if this was a self-triggered event for add/modify
+                    if change in (Change.added, Change.modified) and self._is_expected_content(path_str):
+                        logger.debug("Ignoring self-triggered %s event for: %s", change.name, path_str)
+                        continue
+
+                    # Handle the file change
+                    self._handle_file_change(change, path, path_str)
+
+        except Exception as e:
+            if not self._watching_stopped.is_set():
+                logger.error("Error in file watching thread: %s", str(e))
+        finally:
+            logger.debug("File watching thread stopped")
+
+    def _handle_file_change(self, change: Change, path: Path, path_str: str) -> None:
+        """Handle a file system change event."""
+        if change == Change.added:
+            logger.info("Detected external creation of workflow file: %s", path_str)
+            self._upload_workflow_file(path)
+        elif change == Change.modified:
+            logger.info("Detected external modification of workflow file: %s", path_str)
+            self._upload_workflow_file(path)
+        elif change == Change.deleted:
+            logger.info("Detected deletion of workflow file: %s", path_str)
+            self._delete_workflow_file(path)
+
+    def _download_single_workflow(
+        self, file_name: str, storage_driver: GriptapeCloudStorageDriver, sync_dir: Path
+    ) -> tuple[str, bool, str | None]:
+        """Download a single workflow file.
+
+        Args:
+            file_name: Name of the workflow file to download from cloud
+            storage_driver: Griptape Cloud storage driver instance for downloading
+            sync_dir: Local directory path where the workflow file will be saved
+
+        Returns:
+            (filename, success, error_message)
+        """
+        try:
+            # Download file content
+            file_content = storage_driver.download_file(file_name)
+
+            # Extract just the filename (remove any directory prefixes)
+            local_filename = Path(file_name).name
+            local_file_path = sync_dir / local_filename
+
+            # Check if file exists and has same content hash
+            should_write = True
+            if local_file_path.exists():
+                try:
+                    existing_content = local_file_path.read_bytes()
+
+                    existing_hash = hashlib.sha256(existing_content).hexdigest()
+                    new_hash = hashlib.sha256(file_content).hexdigest()
+
+                    if existing_hash == new_hash:
+                        logger.debug("Skipping write - file already has same content hash: %s", local_filename)
+                        should_write = False
+                except Exception as e:
+                    logger.debug("Error checking existing file hash for %s: %s", local_filename, str(e))
+
+            if should_write:
+                # Set expected hash before writing to prevent sync loops
+                self._set_expected_hash(local_file_path, file_content)
+
+                # Write to local file
+                local_file_path.write_bytes(file_content)
+
+        except Exception as e:
+            error_msg = str(e)
+            logger.warning("Failed to sync workflow '%s': %s", file_name, error_msg)
+            return file_name, False, error_msg
+        else:
+            logger.debug("Successfully synced workflow: %s", local_filename)
+            return local_filename, True, None
+
+    def _sync_workflows_background(
+        self, sync_id: str, workflow_files: list[str], storage_driver: GriptapeCloudStorageDriver, sync_dir: Path
+    ) -> None:
+        """Background thread function to sync workflows."""
+        from griptape_nodes.app.app import event_queue
+
+        synced_workflows = []
+        failed_downloads = []
+        total_workflows = len(workflow_files)
+
+        logger.info("Starting background sync of %d workflows (sync_id: %s)", total_workflows, sync_id)
+
+        # Use thread pool for concurrent downloads
+        with ThreadPoolExecutor() as executor:
+            # Submit all download tasks
+            future_to_filename = {
+                executor.submit(self._download_single_workflow, filename, storage_driver, sync_dir): filename
+                for filename in workflow_files
+            }
+
+            # Collect results as they complete
+            for future in future_to_filename:
+                filename, success, error = future.result()
+                if success:
+                    synced_workflows.append(filename)
+                else:
+                    failed_downloads.append(filename)
+
+        if failed_downloads:
+            logger.warning("Failed to sync %d workflows: %s", len(failed_downloads), failed_downloads)
+
+        logger.info(
+            "Background sync completed: %d of %d workflows synced to %s (sync_id: %s)",
+            len(synced_workflows),
+            len(workflow_files),
+            sync_dir,
+            sync_id,
+        )
+
+        # Emit sync complete event
+        sync_complete_event = SyncComplete(
+            sync_directory=str(sync_dir),
+            synced_workflows=synced_workflows,
+            failed_workflows=failed_downloads,
+            total_workflows=total_workflows,
+        )
+        event_queue.put(AppEvent(payload=sync_complete_event))
+
+        # Register workflows from the synced directory
+        if synced_workflows:
+            logger.info("Registering %d synced workflows from configuration", len(synced_workflows))
+            try:
+                from griptape_nodes.retained_mode.griptape_nodes import GriptapeNodes
+
+                register_request = RegisterWorkflowsFromConfigRequest(
+                    config_section="app_events.on_app_initialization_complete.workflows_to_register"
+                )
+                register_result = GriptapeNodes.handle_request(register_request)
+
+                if isinstance(register_result, RegisterWorkflowsFromConfigResultSuccess):
+                    logger.info(
+                        "Successfully registered %d workflows after sync completion: %s",
+                        len(register_result.succeeded_workflows),
+                        register_result.succeeded_workflows,
+                    )
+                else:
+                    logger.warning("Failed to register workflows after sync completion")
+            except Exception as e:
+                logger.error("Error registering workflows after sync: %s", str(e))
+
+        # Clean up task tracking
+        if sync_id in self._active_sync_tasks:
+            del self._active_sync_tasks[sync_id]