Update to griptape 1.0 (#142)

* Update to griptape 1.0

* Remove more engines

* Fix seo context

* Update incorrect defaults

* Update factory

* Remove engine reference

* Revert output_task change

* Revert pyproject change

Makefile

@@ -14,7 +14,7 @@ check: check/types check/spell check/lint ## Run all checks.
 
 .PHONY: check/types
 check/types:
-	@poetry run pyright docs/courses/**/code_reviews/*.py
+	@poetry run pyright $(shell find docs -type f -name "*.py" -path "*/code_reviews/*")
 
 .PHONY: check/spell
 check/spell:

docs/courses/create-image-pipeline/02_concepts.md

@@ -13,7 +13,7 @@ In this course, we're going to be taking a topic to draw, and then execute a con
 
 1. Use an LLM to generate an Image Generation prompt.
     * If the input given is "a butterfly", the result of this task might be a text string "Create a watercolor painting of a butterfly.".
-    * This will allow us to ensure the output from this task will always be in the correct format for an Image Generation Engine.
+    * This will allow us to ensure the output from this task will always be in the correct format for an Image Generation Driver.
 3. Generate an image
     * Using the output from the previous step, we'll give that to an Image Generation Model
 4. View the image
@@ -61,7 +61,7 @@ Before we dive in and start setting up our own Pipeline, it's important to revie
 | **Extraction Tasks** | Various tasks associated with extracting information from text. | See examples in the [documentation](https://griptape.readthedocs.io/griptape-framework/structures/tasks/#extraction-task){target="_blank"}. 
 | **TextSummaryTask** | Summarizes text very efficiently | `TextSummaryTask("Imagine this is a massive amount of text.")` |
 | **TextQueryTask** | Can be used to query large bodies of text, for example a vector database. | See examples in the [documentation](https://griptape.readthedocs.io/griptape-framework/structures/tasks/#text-query-task){target="_blank"} |
-| **PromptImageGenerationTask** | Can be used to generate images based off a text prompt. | `PromptImageGenerationTask("watercolor butterfly"), image_generation_engine=image_engine`|
+| **PromptImageGenerationTask** | Can be used to generate images based off a text prompt. | `PromptImageGenerationTask("watercolor butterfly"), image_generation_driver=image_driver`|
 | **CodeExecutionTask** | Can be used to execute code. | `CodeExecutionTask(on_run=reverse_string)`|
 
 In this course, we will be focusing mostly on **Prompt Tasks**, **Image Generation Tasks**, and **Code Execution Tasks**. 
@@ -82,10 +82,8 @@ image_prompt_task = PromptTask(
 #
 image_generation_task = PromptImageGenerationTask(
     "{{ parent_output }}", # The output of the parent task
-    image_generation_engine=PromptImageGenerationEngine(
-        image_generation_driver=OpenAiImageGenerationDriver(
-            model="dall-e-3", api_type="open_ai", image_size="1024x1024"
-        ),
+    image_generation_driver=OpenAiImageGenerationDriver(
+        model="dall-e-3", api_type="open_ai", image_size="1024x1024"
     ),
     output_dir="./images",
     id="image_generation_task"

docs/courses/create-image-pipeline/04_creating_images.md

@@ -6,7 +6,6 @@
 Image Generation via Griptape is handled via a few components.
 
 * **Image Generation Driver** - Determines the model to be used. For example, [OpenAI DALL·E 3](https://openai.com/dall-e-3){target="_blank"} or [Leonardo.AI](https://leonardo.ai/){target="_blank"}.
-* **Image Generation Engines** - The engine that facilitates the use of the Driver.
 * **Image Generation Task** or **Image Generation Tool**. The Task or Tool is what will be provided to the Griptape Structure. Pipelines and Workflows can use `PromptImageGenerationTask` directly. You can provide an `PromptImageGenerationTool` to an Agent, to a ToolTask, or to a ToolkitTask.
 
 For example, to create an image with OpenAI DALL·E 3 as a **task** you could do something like:
@@ -18,33 +17,28 @@ driver = OpenAiImageGenerationDriver(
     model="dall-e-3", api_type="open_ai", image_size="1024x1024"
 )
 
-# Create an Image Generation Engine
-engine = PromptImageGenerationEngine( image_generation_driver=driver )
-
 # Create an Image Generation Task
 task = PromptImageGenerationTask(
     "Create a drawing of a pineapple",
-    image_generation_engine=engine,
+    image_generation_driver=driver,
     output_dir="images"
 )
 ```
 
 Once you generate the task, you will add it to the pipeline or workflow.
 
-You can also use the `PromptImageGenerationTool` tool and assign it to an Agent. It takes many of the same arguments. If you had previously created the `driver` and `engine` as specified above, you would do something like:
+You can also use the `PromptImageGenerationTool` tool and assign it to an Agent. It takes many of the same arguments. If you had previously created the `driver` as specified above, you would do something like:
 
 ```python
 agent = Agent(
     tools=[PromptImageGenerationTool(
-        image_generation_engine=engine,
+        image_generation_driver=driver,
         output_dir="images",
         off_prompt=False,
     )]
 )
 ```
 
-The main thing to be aware of is that you must use _both_ components - Driver and Engine. You choose the model with the **Driver**, use the **Engine** to facilitate the use of the model, and then access the engine with either a **Task** or a **Tool**.
-
 In this course, because we're focusing on image generation as part of a _Pipeline_, we'll generate images using a _Task_.
 
 ## The Image Task
@@ -53,7 +47,7 @@ To get started, we'll begin by replacing the Fake Image Generation task with a r
 
 ### Imports
 
-To use the Driver, Engine, and Task we'll need to add them to our `imports` section in `app.py`. You'll modify `griptape.tasks` to include `PromptImageGenerationTask`, and add imports for the Driver and Engine.
+To use the Driver and Task we'll need to add them to our `imports` section in `app.py`. You'll modify `griptape.tasks` to include `PromptImageGenerationTask`, and add imports for the Driver and Engine.
 
 ```python hl_lines="5-7"
 # ...
@@ -62,7 +56,6 @@ To use the Driver, Engine, and Task we'll need to add them to our `imports` sect
 from griptape.structures import Pipeline
 from griptape.tasks import PromptTask, PromptImageGenerationTask
 from griptape.drivers import OpenAiImageGenerationDriver
-from griptape.engines import PromptImageGenerationEngine
 
 # ...
 ```
@@ -87,25 +80,6 @@ image_driver = OpenAiImageGenerationDriver(
 # ...
 ```
 
-### Create the Engine
-
-The engine facilitates the use of the particular model. It will be what we pass to the task or tool. After the creation of the driver, create the engine:
-
-```python hl_lines="8-9"
-# ...
-
-# Create the driver
-image_driver = OpenAiImageGenerationDriver(
-    model="dall-e-3", api_type="open_ai", image_size="1024x1024"
-)
-
-# Create the engine
-image_engine = PromptImageGenerationEngine(image_generation_driver=image_driver)
-
-# Create the pipeline object
-# ...
-```
-
 ### Replace the ImageTask
 
 Next, we'll replace our fake image generation task with a _real_ image generation task. Find the section of the code where we're creating the image task with `generate_image_task` and replace it with `PromptImageGenerationTask`.
@@ -115,15 +89,15 @@ Next, we'll replace our fake image generation task with a _real_ image generatio
 
 generate_image_task = PromptImageGenerationTask(
     "{{ parent_output }}",
-    image_generation_engine=image_engine,
+    image_generation_driver=image_driver,
     output_dir="images",
     id="Generate Image Task",
 )
 
 # ...
 ```
 
-Notice we're giving it the `image_generation_engine` we defined earlier as `image_engine`. We're also specifying an `output_dir` of `images`. This will ensure the image is generated in that directory. 
+Notice we're giving it the `image_generation_driver` we defined earlier as `image_driver`. We're also specifying an `output_dir` of `images`. This will ensure the image is generated in that directory. 
 
 !!! tip
     With the `PromptImageGenerationTask`, if you want to save the file to disk you must specify _either_ the output file name (`output_file`) or the directory you want the images to appear in (`output_dir`). If you don't, the image generated will only exist in the `ImageArtifact`. 
@@ -250,7 +224,7 @@ If you scroll down in the code until you find the section where the `run` method
 # ...
 
     def run(self) -> ImageArtifact:
-        image_artifact = self.image_generation_engine.generate_image(
+        image_artifact = self.image_generation_driver.generate_image(
             prompts=[self.input.to_text()], rulesets=self.all_rulesets, negative_rulesets=self.negative_rulesets
         )
 
@@ -358,7 +332,7 @@ Now go down to the `generate_image_task` and use the `output_dir` variable in th
 
 generate_image_task = PromptImageGenerationTask(
     "{{ parent_output }}",
-    image_generation_engine=image_engine,
+    image_generation_driver=image_driver,
     output_dir=output_dir,
     id="Generate Image Task",
 )

docs/courses/create-image-pipeline/07_display_image_task.md

@@ -168,7 +168,6 @@ from griptape.tasks import (
     CodeExecutionTask
 )
 from griptape.drivers import OpenAiImageGenerationDriver
-from griptape.engines import PromptImageGenerationEngine
 from griptape.artifacts import TextArtifact
 
 ```
@@ -180,9 +179,6 @@ We won't display the image yet, we'll just create the function and make sure we
 ```python title="app.py" hl_lines="6-17"
 # ...
 
-# Create the engine
-image_engine = PromptImageGenerationEngine(image_generation_driver=image_driver)
-
 # Create a function to display an image
 def display_image(task: CodeExecutionTask) --> TextArtifact:
     # Get the filename

docs/courses/create-image-pipeline/assets/code_reviews/04/app.py

@@ -4,7 +4,6 @@
 from griptape.structures import Pipeline
 from griptape.tasks import PromptTask, PromptImageGenerationTask
 from griptape.drivers import OpenAiImageGenerationDriver
-from griptape.engines import PromptImageGenerationEngine
 
 load_dotenv()  # Load your environment
 
@@ -14,8 +13,6 @@
 # Create the driver
 image_driver = OpenAiImageGenerationDriver(model="dall-e-3", api_type="open_ai", image_size="1024x1024")
 
-# Create the engine
-image_engine = PromptImageGenerationEngine(image_generation_driver=image_driver)
 
 # Create the pipeline object
 pipeline = Pipeline()
@@ -33,7 +30,7 @@
 
 generate_image_task = PromptImageGenerationTask(
     "{{ parent_output }}",
-    image_generation_engine=image_engine,
+    image_generation_driver=image_driver,
     output_dir=output_dir,
     id="Generate Image Task",
 )

docs/courses/create-image-pipeline/assets/code_reviews/07/app.py

@@ -1,7 +1,6 @@
 from dotenv import load_dotenv
 from griptape.artifacts import TextArtifact
 from griptape.drivers import OpenAiImageGenerationDriver
-from griptape.engines import PromptImageGenerationEngine
 
 # Griptape
 from griptape.structures import Pipeline
@@ -19,9 +18,6 @@
 # Create the driver
 image_driver = OpenAiImageGenerationDriver(model="dall-e-3", api_type="open_ai", image_size="1024x1024")
 
-# Create the engine
-image_engine = PromptImageGenerationEngine(image_generation_driver=image_driver)
-
 
 # Create a function to display an image
 def display_image(task: CodeExecutionTask) -> TextArtifact:
@@ -61,7 +57,7 @@ def display_image(task: CodeExecutionTask) -> TextArtifact:
 
 generate_image_task = PromptImageGenerationTask(
     "{{ parent_output }}",
-    image_generation_engine=image_engine,
+    image_generation_driver=image_driver,
     output_dir=output_dir,
     id="Generate Image Task",
 )

docs/courses/create-image-pipeline/index.md

@@ -49,8 +49,8 @@ If you don't have those items available, it's highly recommended you go through
 
 It's also recommended to view the [Compare Movies Workflow](../compare-movies-workflow/index.md) course if you haven't viewed it before, as it contains some similar concepts to the Pipeline course.
 
-## Image Generation Engines
-The course will cover some of the Image Generation Engines available for Griptape, including [OpenAI DALL·E 3](https://openai.com/dall-e-3){target="_blank"}, [Leonardo.AI](https://leonardo.ai/){target="_blank"}, and [Image Generation Engines running on Amazon Bedrock](https://aws.amazon.com/bedrock/stable-diffusion/){target="_blank"} 
+## Image Generation Drivers
+The course will cover some of the Image Generation Drivers available for Griptape, including [OpenAI DALL·E 3](https://openai.com/dall-e-3){target="_blank"}, [Leonardo.AI](https://leonardo.ai/){target="_blank"}, and [Image Generation Drivers running on Amazon Bedrock](https://aws.amazon.com/bedrock/stable-diffusion/){target="_blank"} 
 
 ### DALL·E 3
 * DALL·E 3 is available with an OpenAI API key.
@@ -73,7 +73,7 @@ The course will cover:
 
 * Creating Griptape Pipelines
 * Creating Griptape Tasks
-* Investigate Image Generation Engines
+* Investigate Image Generation Drivers
 * Using a `CodeExecutionTask` to display the resulting image
 
 ## Useful Resources and Links

docs/courses/image-query/02_chatbot.md

@@ -64,7 +64,7 @@ from rich import print as print # Modifies print to use the Rich library
 # ...
 ```
 
-Then, create a function to format and print responses from the agent in dark cyan. Add this code before you start the Chat function, passing it the function with the `output_fn` parameter.
+Then, create a function to format and print responses from the agent in dark cyan. Add this code before you start the Chat function, passing it the function with the `handle_output` parameter.
 
 ```python title="app.py" hl_lines="3-5 9"
 
@@ -77,7 +77,7 @@ def formatted_response(response: str) -> None:
 # Begin Chatting
 Chat(
     agent,
-    output_fn=formatted_response, # Uses the formatted_response function
+    handle_output=formatted_response, # Uses the formatted_response function
 ).start()
 
 ```
@@ -114,7 +114,7 @@ Chat(
     prompt_prefix="\nYou: ",
     processing_text="\nThinking...",
     response_prefix="\nAgent: ",
-    output_fn=formatted_response,  # Uses the formatted_response function
+    handle_output=formatted_response,  # Uses the formatted_response function
 ).start()
 
 ```

docs/courses/image-query/03_image_query_tool.md

@@ -26,19 +26,15 @@ Add the following to your `import` section:
 # ...
 
 from griptape.tools import ImageQueryTool
-from griptape.engines import ImageQueryEngine
-from griptape.drivers import OpenAiImageQueryDriver
+from griptape.drivers import OpenAiChatPromptDriver
 
 # ...
 ```
 
 ### ImageQueryTool
 This component acts as the intermediary that will manage your interactions with images. It's responsible for sending image data to the engine and retrieving results.
 
-### ImageQueryEngine
-This is where the magic happens. The engine processes the image data, analyzes it, and interprets what's in the image based on the capabilities of the underlying model.
-
-### OpenAiImageQueryDriver
+### OpenAiChatPromptDriver
 This driver leverages OpenAI's advanced vision capabilities to understand and analyze images. It's a powerful tool that enables your agent to not just see images but truly understand their content, like I said before.
 
 An example of another ImageQueryDriver is Anthropic. Check out other available drivers in the [Griptape docs](https://docs.griptape.ai/stable/griptape-framework/drivers/image-query-drivers/)
@@ -64,34 +60,19 @@ driver = OpenAiImageQueryDriver(
 # ...
 ```
 
-### Create the Image Query Engine
-
-This section effectively connects our previously created driver to the engine. Think of the engine as the middle manager in your corporate structure—it doesn’t do the grunt work itself (that’s the driver’s job), but it makes sure the instructions and data flow properly to get the job done.
-
-```python title="app.py" hl_lines="3-6"
-# ...
-
-# Create an Image Query Engine
-engine = ImageQueryEngine(
-    image_query_driver=driver,
-)
-
-# ...
-```
-
 With the driver and engine now set up, your agent is almost ready to start describing pictures as if it were an art critic at a gallery opening. But remember, even though it can now "see" images, don't expect it to understand why your dog looks at you like that—some mysteries remain unsolved even for the most advanced AI.
 
 Next, we'll tie all these components together and create a client that can handle all the image querying magic.
 
 ### Create the Tool
 
-First up, let’s create and configure the `ImageQueryTool`. This component uses the engine we set up earlier to process image queries. The `off_prompt` parameter is set to `False` to ensure that the image descriptions are not just processed but also sent back to the large language model for some insightful summarization. Here's how to do it:
+First up, let’s create and configure the `ImageQueryTool`. This component uses the driver we set up earlier to process image queries. The `off_prompt` parameter is set to `False` to ensure that the image descriptions are not just processed but also sent back to the large language model for some insightful summarization. Here's how to do it:
 
 ```python title="app.py" hl_lines="3-6"
 # ...
 
 image_query_tool = ImageQueryTool(
-    image_query_engine=engine, 
+    prompt_driver=driver, 
     off_prompt=False
 )
 

docs/courses/image-query/06_querying_multiple_images.md

@@ -109,7 +109,7 @@ Modify the code as follows:
 ```python title="app.py" hl_lines="5-25"
 # ...
 # Configure the ImageQueryTool
-image_query_tool = ImageQueryTool(image_query_engine=engine, off_prompt=False)
+image_query_tool = ImageQueryTool(prompt_driver=driver, off_prompt=False)
 
 flow = "AGENT"
 if flow == "WORKFLOW":
@@ -131,7 +131,7 @@ else:
       prompt_prefix="\nYou: ",
       processing_text="\nThinking...",
       response_prefix="\nAgent: ",
-      output_fn=formatted_response,  # Uses the formatted_response function
+      handle_output=formatted_response,  # Uses the formatted_response function
   ).start()
 ```
 
@@ -285,7 +285,7 @@ if flow == "WORKFLOW":
           + "display the image. Save this to image_descriptions/{{ filename }}.yml\n"
           + "in YAML format.\n\n{{ parent_outputs }}",
           tools=[FileManagerTool(off_prompt=False)],
-          context={"image": image},
+          context={"filename": filename},
           id=f"seo_{image}",
       )
 

docs/courses/image-query/assets/code_reviews/02/app.py

@@ -25,5 +25,5 @@ def formatted_response(response: str) -> None:
     prompt_prefix="\nYou: ",
     processing_text="\nThinking...",
     response_prefix="\nAgent: ",
-    output_fn=formatted_response,  # Uses the formatted_response function
+    handle_output=formatted_response,  # Uses the formatted_response function
 ).start()