Griptape

Fixed

- Restore human-friendly default `ImageArtifact` and `AudioArtifact` names with file type extension.

Added

- `WebScraperTool.text_chunker` default value for `max_tokens`.

Fixed

- `WebScraperTool` not using `text_chunker` override.
- Breaking change in `Chat.handle_output` behavior.

Added

- `griptape.configs.logging.JsonFormatter` for formatting logs as JSON.
- Request/response debug logging to all Prompt Drivers.
- `griptape.schemas.UnionField` for serializing union fields.
- `BaseEventListener.flush_events()` to flush events from an Event Listener.
- Exponential backoff to `BaseEventListenerDriver` for retrying failed event publishing.
- `BaseTask.task_outputs` to get a dictionary of all task outputs. This has been added to `Workflow.context` and `Pipeline.context`.
- `Chat.input_fn` for customizing the input to the Chat utility.
- `GriptapeCloudFileManagerDriver` for managing files on Griptape Cloud.
- `BaseFileManagerDriver.load_artifact()` & `BaseFileManagerDriver.save_artifact()` for loading & saving artifacts as files.
- Events `BaseChunkEvent`, `TextChunkEvent`, `ActionChunkEvent`.
- `wrapt` dependency for more robust decorators.
- `BasePromptDriver.extra_params` for passing extra parameters not explicitly declared by the Driver.
- `RunnableMixin` which adds `on_before_run` and `on_after_run` hooks.
- `griptape.utils.with_contextvars` utility for running functions with the current `contextvars` context.

Changed

- **BREAKING**: Removed `BaseEventListenerDriver.publish_event` `flush` argument. Use `BaseEventListenerDriver.flush_events()` instead.
- **BREAKING**: Renamed parameter `driver` on `EventListener` to `event_listener_driver`.
- **BREAKING**: Updated `EventListener.handler` return value behavior.
- If `EventListener.handler` returns `None`, the event will not be published to the `event_listener_driver`.
- If `EventListener.handler` is None, the event will be published to the `event_listener_driver` as-is.
- **BREAKING**: Removed `CompletionChunkEvent`.
- **BREAKING**: Moved `griptape.common.observable.observable` to `griptape.common.decorators.observable`.
- **BREAKING**: `AnthropicDriversConfig` no longer bundles `VoyageAiEmbeddingDriver`.
- **BREAKING**: Removed `HuggingFaceHubPromptDriver.params`, use `HuggingFaceHubPromptDriver.extra_params` instead.
- **BREAKING**: Removed `HuggingFacePipelinePromptDriver.params`, use `HuggingFacePipelinePromptDriver.extra_params` instead.
- **BREAKING**: Renamed `BaseTask.run` to `BaseTask.try_run`.
- **BREAKING**: Renamed `BaseTask.execute` to `BaseTask.run`.
- **BREAKING**: Renamed `BaseTask.can_execute` to `BaseTool.can_run`.
- **BREAKING**: Renamed `BaseTool.run` to `BaseTool.try_run`.
- **BREAKING**: Renamed `BaseTool.execute` to `BaseTool.run`.
- **BREAKING**: Renamed callables throughout the framework for consistency:
- Renamed `LocalStructureRunDriver.structure_factory_fn` to `LocalStructureRunDriver.create_structure`.
- Renamed `SnowflakeSqlDriver.connection_func` to `SnowflakeSqlDriver.get_connection`.
- Renamed `CsvLoader.formatter_fn` to `CsvLoader.format_row`.
- Renamed `SqlLoader.formatter_fn` to `SqlLoader.format_row`.
- Renamed `CsvExtractionEngine.system_template_generator` to `CsvExtractionEngine.generate_system_template`.
- Renamed `CsvExtractionEngine.user_template_generator` to `CsvExtractionEngine.generate_user_template`.
- Renamed `JsonExtractionEngine.system_template_generator` to `JsonExtractionEngine.generate_system_template`.
- Renamed `JsonExtractionEngine.user_template_generator` to `JsonExtractionEngine.generate_user_template`.
- Renamed `PromptResponseRagModule.generate_system_template` to `PromptResponseRagModule.generate_system_template`.
- Renamed `PromptTask.generate_system_template` to `PromptTask.generate_system_template`.
- Renamed `ToolkitTask.generate_assistant_subtask_template` to `ToolkitTask.generate_assistant_subtask_template`.
- Renamed `JsonSchemaRule.template_generator` to `JsonSchemaRule.generate_template`.
- Renamed `ToolkitTask.generate_user_subtask_template` to `ToolkitTask.generate_user_subtask_template`.
- Renamed `TextLoaderRetrievalRagModule.process_query_output_fn` to `TextLoaderRetrievalRagModule.process_query_output`.
- Renamed `FuturesExecutorMixin.futures_executor_fn` to `FuturesExecutorMixin.create_futures_executor`.
- Renamed `VectorStoreTool.process_query_output_fn` to `VectorStoreTool.process_query_output`.
- Renamed `CodeExecutionTask.run_fn` to `CodeExecutionTask.on_run`.
- Renamed `Chat.input_fn` to `Chat.handle_input`.
- Renamed `Chat.output_fn` to `Chat.handle_output`.
- Renamed `EventListener.handler` to `EventListener.on_event`.
- Updated `EventListener.handler` return type to `Optional[BaseEvent | dict]`.
- `BaseTask.parent_outputs` type has changed from `dict[str, str | None]` to `dict[str, BaseArtifact]`.
- `Workflow.context["parent_outputs"]` type has changed from `dict[str, str | None]` to `dict[str, BaseArtifact]`.
- `Pipeline.context["parent_output"]` has changed type from `str | None` to `BaseArtifact | None`.
- `_DefaultsConfig.logging_config` and `Defaults.drivers_config` are now lazily instantiated.
- `griptape.schemas.BaseSchema` now uses `griptape.schemas.UnionField` for `Union` fields.
- `BaseTask.add_parent`/`BaseTask.add_child` now only add the parent/child task to the structure if it is not already present.
- `BaseEventListener.flush_events()` to flush events from an Event Listener.
- `BaseEventListener` no longer requires a thread lock for batching events.
- Updated `ToolkitTask` system prompt to retry/fix actions when using native tool calling.
- `Chat` input now uses a slightly customized version of `Rich.prompt.Prompt` by default.
- `Chat` output now uses `Rich.print` by default.
- `Chat.output_fn`'s now takes an optional kwarg parameter, `stream`.
- Implemented `SerializableMixin` in `Structure`, `BaseTask`, `BaseTool`, and `TaskMemory`
- `activity` decorated functions can now accept kwargs that are defined in the activity schema.
- Updated `ToolkitTask` system prompt to no longer mention `memory_name` and `artifact_namespace`.
- Models in `ToolkitTask` with native tool calling no longer need to provide their final answer as `Answer:`.
- `EventListener.event_types` will now listen on child types of any provided type.
- Only install Tool dependencies if the Tool provides a `requirements.txt` and the dependencies are not already met.
- Implemented `RunnableMixin` in `Structure`, `BaseTask`, and `BaseTool`.
- `EventBus`'s Event Listeners are now thread/coroutine-local. Event Listeners from the spawning thread will be automatically copied when using concurrent griptape features like Workflows.

Fixed

- Structures not flushing events when not listening for `FinishStructureRunEvent`.
- `EventListener.event_types` and the argument to `BaseEventListenerDriver.handler` being out of sync.
- Models occasionally hallucinating `memory_name` and `artifact_namespace` into Tool schemas when using `ToolkitTask`.
- Models occasionally providing overly succinct final answers when using `ToolkitTask`.
- Exception getting raised in `FuturesExecutorMixin.__del__`.
- Issues when using `EventListener` as a context manager in a multi-threaded environment.

Fixed

- Pinned `cohere` at `~5.11.0` to resolve slow dependency resolution.
- Missing `exa-py` from `all` extra.

Added

- `Workflow.input_tasks` and `Workflow.output_tasks` to access the input and output tasks of a Workflow.
- Ability to pass nested list of `Tasks` to `Structure.tasks` allowing for more complex declarative Structure definitions.
- `TavilyWebSearchDriver` to integrate Tavily's web search capabilities.
- `ExaWebSearchDriver` to integrate Exa's web search capabilities.
- `Workflow.outputs` to access the outputs of a Workflow.
- `BaseFileLoader` for Loaders that load from a path.
- `BaseLoader.fetch()` method for fetching data from a source.
- `BaseLoader.parse()` method for parsing fetched data.
- `BaseFileManager.encoding` to specify the encoding when loading and saving files.
- `BaseWebScraperDriver.extract_page()` method for extracting data from an already scraped web page.
- `TextLoaderRetrievalRagModule.chunker` for specifying the chunking strategy.
- `file_utils.get_mime_type` utility for getting the MIME type of a file.
- `BaseRulesetDriver` for loading a `Ruleset` from an external source.
- `LocalRulesetDriver` for loading a `Ruleset` from a local `.json` file.
- `GriptapeCloudRulesetDriver` for loading a `Ruleset` resource from Griptape Cloud.
- Parameter `alias` on `GriptapeCloudConversationMemoryDriver` for fetching a Thread by alias.
- Basic support for OpenAi Structured Output via `OpenAiChatPromptDriver.response_format` parameter.
- Ability to pass callable to `activity.schema` for dynamic schema generation.

Changed

- **BREAKING**: Renamed parameters on several classes to `client`:
- `bedrock_client` on `AmazonBedrockCohereEmbeddingDriver`.
- `bedrock_client` on `AmazonBedrockCohereEmbeddingDriver`.
- `bedrock_client` on `AmazonBedrockTitanEmbeddingDriver`.
- `bedrock_client` on `AmazonBedrockImageGenerationDriver`.
- `bedrock_client` on `AmazonBedrockImageQueryDriver`.
- `bedrock_client` on `AmazonBedrockPromptDriver`.
- `sagemaker_client` on `AmazonSageMakerJumpstartEmbeddingDriver`.
- `sagemaker_client` on `AmazonSageMakerJumpstartPromptDriver`.
- `sqs_client` on `AmazonSqsEventListenerDriver`.
- `iotdata_client` on `AwsIotCoreEventListenerDriver`.
- `s3_client` on `AmazonS3FileManagerDriver`.
- `s3_client` on `AwsS3Tool`.
- `iam_client` on `AwsIamTool`.
- `pusher_client` on `PusherEventListenerDriver`.
- `mq` on `MarqoVectorStoreDriver`.
- `model_client` on `GooglePromptDriver`.
- `model_client` on `GoogleTokenizer`.
- **BREAKING**: Renamed parameter `pipe` on `HuggingFacePipelinePromptDriver` to `pipeline`.
- **BREAKING**: Removed `BaseFileManager.default_loader` and `BaseFileManager.loaders`.
- **BREAKING**: Loaders no longer chunk data, use a Chunker to chunk the data.
- **BREAKING**: Removed `fileutils.load_file` and `fileutils.load_files`.
- **BREAKING**: Removed `loaders-dataframe` and `loaders-audio` extras as they are no longer needed.
- **BREKING**: `TextLoader`, `PdfLoader`, `ImageLoader`, and `AudioLoader` now take a `str | PathLike` instead of `bytes`. Passing `bytes` is still supported but deprecated.
- **BREAKING**: Removed `DataframeLoader`.
- **BREAKING**: Update `pypdf` dependency to `^5.0.1`.
- **BREAKING**: Update `redis` dependency to `^5.1.0`.
- **BREAKING**: Remove `torch` extra from `transformers` dependency. This must be installed separately.
- **BREAKING**: Split `BaseExtractionEngine.extract` into `extract_text` and `extract_artifacts` for consistency with `BaseSummaryEngine`.
- **BREAKING**: `BaseExtractionEngine` no longer catches exceptions and returns `ErrorArtifact`s.
- **BREAKING**: `JsonExtractionEngine.template_schema` is now required.
- **BREAKING**: `CsvExtractionEngine.column_names` is now required.
- **BREAKING**: Renamed`RuleMixin.all_rulesets` to `RuleMixin.rulesets`.
- **BREAKING**: Renamed `GriptapeCloudKnowledgeBaseVectorStoreDriver` to `GriptapeCloudVectorStoreDriver`.
- **BREAKING**: `OpenAiChatPromptDriver.response_format` is now a `dict` instead of a `str`.
- `MarkdownifyWebScraperDriver.DEFAULT_EXCLUDE_TAGS` now includes media/blob-like HTML tags
- `StructureRunTask` now inherits from `PromptTask`.
- Several places where API clients are initialized are now lazy loaded.
- `BaseVectorStoreDriver.upsert_text_artifacts` now returns a list or dictionary of upserted vector ids.
- `LocalFileManagerDriver.workdir` is now optional.
- `filetype` is now a core dependency.
- `FileManagerTool` now uses `filetype` for more accurate file type detection.
- `BaseFileLoader.load_file()` will now either return a `TextArtifact` or a `BlobArtifact` depending on whether `BaseFileManager.encoding` is set.
- `Structure.output`'s type is now `BaseArtifact` and raises an exception if the output is `None`.
- `JsonExtractionEngine.extract_artifacts` now returns a `ListArtifact[JsonArtifact]`.
- `CsvExtractionEngine.extract_artifacts` now returns a `ListArtifact[CsvRowArtifact]`.
- Remove `manifest.yml` requirements for custom tool creation.

Fixed

- Anthropic native Tool calling.
- Empty `ActionsSubtask.thought` being logged.
- `RuleMixin` no longer prevents setting `rulesets` _and_ `rules` at the same time.
- `PromptTask` will merge in its Structure's Rulesets and Rules.
- `PromptTask` not checking whether Structure was set before building Prompt Stack.
- `BaseTask.full_context` context being empty when not connected to a Structure.
- Tool calling when using `OpenAiChatPromptDriver` with Groq.

Added

- `BaseArtifact.to_bytes()` method to convert an Artifact's value to bytes.
- `BlobArtifact.base64` property for converting a `BlobArtifact`'s value to a base64 string.
- `CsvLoader`/`SqlLoader`/`DataframeLoader` `formatter_fn` field for customizing how SQL results are formatted into `TextArtifact`s.
- `AzureOpenAiTextToSpeechDriver`.
- `JsonSchemaRule` for instructing the LLM to output a JSON object that conforms to a schema.
- Ability to use Event Listeners as Context Managers for temporarily setting the Event Bus listeners.
- Ability to use Drivers Configs as Context Managers for temporarily setting the default Drivers.
- Generic type support to `ListArtifact`.
- Iteration support to `ListArtifact`.

Changed

- **BREAKING**: Removed `CsvRowArtifact`. Use `TextArtifact` instead.
- **BREAKING**: Removed `DataframeLoader`.
- **BREAKING**: Removed `MediaArtifact`, use `ImageArtifact` or `AudioArtifact` instead.
- **BREAKING**: `CsvLoader` and `SqlLoader` now return `ListArtifact[TextArtifact]`.
- **BREAKING**: Removed `ImageArtifact.media_type`.
- **BREAKING**: Removed `AudioArtifact.media_type`.
- **BREAKING**: Removed `BlobArtifact.dir_name`.
- **BREAKING**: Moved `ImageArtifact.prompt` and `ImageArtifact.model` into `ImageArtifact.meta`.
- **BREAKING**: `ImageArtifact.format` is now required.
- **BREAKING**: Removed the `__all__` declaration from the `griptape.mixins` module.
- Updated `JsonArtifact` value converter to properly handle more types.
- `AudioArtifact` now subclasses `BlobArtifact` instead of `MediaArtifact`.
- `ImageArtifact` now subclasses `BlobArtifact` instead of `MediaArtifact`.
- Removed `__add__` method from `BaseArtifact`, implemented it where necessary.

Fixed

- Crash when passing "empty" Artifacts or no Artifacts to `CohereRerankDriver`.