discourse-ai/lib/ai_bot/tools/create_artifact.rb

170 lines
5.1 KiB
Ruby
Raw Normal View History

FEATURE: AI artifacts (#898) This is a significant PR that introduces AI Artifacts functionality to the discourse-ai plugin along with several other improvements. Here are the key changes: 1. AI Artifacts System: - Adds a new `AiArtifact` model and database migration - Allows creation of web artifacts with HTML, CSS, and JavaScript content - Introduces security settings (`strict`, `lax`, `disabled`) for controlling artifact execution - Implements artifact rendering in iframes with sandbox protection - New `CreateArtifact` tool for AI to generate interactive content 2. Tool System Improvements: - Adds support for partial tool calls, allowing incremental updates during generation - Better handling of tool call states and progress tracking - Improved XML tool processing with CDATA support - Fixes for tool parameter handling and duplicate invocations 3. LLM Provider Updates: - Updates for Anthropic Claude models with correct token limits - Adds support for native/XML tool modes in Gemini integration - Adds new model configurations including Llama 3.1 models - Improvements to streaming response handling 4. UI Enhancements: - New artifact viewer component with expand/collapse functionality - Security controls for artifact execution (click-to-run in strict mode) - Improved dialog and response handling - Better error management for tool execution 5. Security Improvements: - Sandbox controls for artifact execution - Public/private artifact sharing controls - Security settings to control artifact behavior - CSP and frame-options handling for artifacts 6. Technical Improvements: - Better post streaming implementation - Improved error handling in completions - Better memory management for partial tool calls - Enhanced testing coverage 7. Configuration: - New site settings for artifact security - Extended LLM model configurations - Additional tool configuration options This PR significantly enhances the plugin's capabilities for generating and displaying interactive content while maintaining security and providing flexible configuration options for administrators.
2024-11-18 17:22:39 -05:00
# frozen_string_literal: true
module DiscourseAi
module AiBot
module Tools
class CreateArtifact < Tool
def self.name
"create_artifact"
end
def self.js_dependency_tip
<<~TIP
If you need to include a JavaScript library, you may include assets from:
- unpkg.com
- cdnjs.com
- jsdelivr.com
- ajax.googleapis.com
To include them ensure they are the last tag in your HTML body.
Example: <script crossorigin src="https://cdn.jsdelivr.net/npm/vue@2.6.14/dist/vue.min.js"></script>
TIP
end
def self.js_script_tag_tip
<<~TIP
if you need a custom script tag, you can use the following format:
<script type="module">
// your script here
</script>
If you only need a regular script tag, you can use the following format:
// your script here
TIP
end
FEATURE: AI artifacts (#898) This is a significant PR that introduces AI Artifacts functionality to the discourse-ai plugin along with several other improvements. Here are the key changes: 1. AI Artifacts System: - Adds a new `AiArtifact` model and database migration - Allows creation of web artifacts with HTML, CSS, and JavaScript content - Introduces security settings (`strict`, `lax`, `disabled`) for controlling artifact execution - Implements artifact rendering in iframes with sandbox protection - New `CreateArtifact` tool for AI to generate interactive content 2. Tool System Improvements: - Adds support for partial tool calls, allowing incremental updates during generation - Better handling of tool call states and progress tracking - Improved XML tool processing with CDATA support - Fixes for tool parameter handling and duplicate invocations 3. LLM Provider Updates: - Updates for Anthropic Claude models with correct token limits - Adds support for native/XML tool modes in Gemini integration - Adds new model configurations including Llama 3.1 models - Improvements to streaming response handling 4. UI Enhancements: - New artifact viewer component with expand/collapse functionality - Security controls for artifact execution (click-to-run in strict mode) - Improved dialog and response handling - Better error management for tool execution 5. Security Improvements: - Sandbox controls for artifact execution - Public/private artifact sharing controls - Security settings to control artifact behavior - CSP and frame-options handling for artifacts 6. Technical Improvements: - Better post streaming implementation - Improved error handling in completions - Better memory management for partial tool calls - Enhanced testing coverage 7. Configuration: - New site settings for artifact security - Extended LLM model configurations - Additional tool configuration options This PR significantly enhances the plugin's capabilities for generating and displaying interactive content while maintaining security and providing flexible configuration options for administrators.
2024-11-18 17:22:39 -05:00
def self.signature
{
name: "create_artifact",
description:
"Creates a web artifact with HTML, CSS, and JavaScript that can be displayed in an iframe",
parameters: [
{
name: "name",
description: "A name for the artifact (max 255 chars)",
type: "string",
required: true,
},
{
name: "html_body",
description:
"The HTML content for the BODY tag (do not include the BODY tag). #{js_dependency_tip}",
FEATURE: AI artifacts (#898) This is a significant PR that introduces AI Artifacts functionality to the discourse-ai plugin along with several other improvements. Here are the key changes: 1. AI Artifacts System: - Adds a new `AiArtifact` model and database migration - Allows creation of web artifacts with HTML, CSS, and JavaScript content - Introduces security settings (`strict`, `lax`, `disabled`) for controlling artifact execution - Implements artifact rendering in iframes with sandbox protection - New `CreateArtifact` tool for AI to generate interactive content 2. Tool System Improvements: - Adds support for partial tool calls, allowing incremental updates during generation - Better handling of tool call states and progress tracking - Improved XML tool processing with CDATA support - Fixes for tool parameter handling and duplicate invocations 3. LLM Provider Updates: - Updates for Anthropic Claude models with correct token limits - Adds support for native/XML tool modes in Gemini integration - Adds new model configurations including Llama 3.1 models - Improvements to streaming response handling 4. UI Enhancements: - New artifact viewer component with expand/collapse functionality - Security controls for artifact execution (click-to-run in strict mode) - Improved dialog and response handling - Better error management for tool execution 5. Security Improvements: - Sandbox controls for artifact execution - Public/private artifact sharing controls - Security settings to control artifact behavior - CSP and frame-options handling for artifacts 6. Technical Improvements: - Better post streaming implementation - Improved error handling in completions - Better memory management for partial tool calls - Enhanced testing coverage 7. Configuration: - New site settings for artifact security - Extended LLM model configurations - Additional tool configuration options This PR significantly enhances the plugin's capabilities for generating and displaying interactive content while maintaining security and providing flexible configuration options for administrators.
2024-11-18 17:22:39 -05:00
type: "string",
required: true,
},
{ name: "css", description: "Optional CSS styles for the artifact", type: "string" },
{
name: "js",
description:
"Optional
JavaScript code for the artifact, this will be the last <script> tag in the BODY. #{js_script_tag_tip}",
FEATURE: AI artifacts (#898) This is a significant PR that introduces AI Artifacts functionality to the discourse-ai plugin along with several other improvements. Here are the key changes: 1. AI Artifacts System: - Adds a new `AiArtifact` model and database migration - Allows creation of web artifacts with HTML, CSS, and JavaScript content - Introduces security settings (`strict`, `lax`, `disabled`) for controlling artifact execution - Implements artifact rendering in iframes with sandbox protection - New `CreateArtifact` tool for AI to generate interactive content 2. Tool System Improvements: - Adds support for partial tool calls, allowing incremental updates during generation - Better handling of tool call states and progress tracking - Improved XML tool processing with CDATA support - Fixes for tool parameter handling and duplicate invocations 3. LLM Provider Updates: - Updates for Anthropic Claude models with correct token limits - Adds support for native/XML tool modes in Gemini integration - Adds new model configurations including Llama 3.1 models - Improvements to streaming response handling 4. UI Enhancements: - New artifact viewer component with expand/collapse functionality - Security controls for artifact execution (click-to-run in strict mode) - Improved dialog and response handling - Better error management for tool execution 5. Security Improvements: - Sandbox controls for artifact execution - Public/private artifact sharing controls - Security settings to control artifact behavior - CSP and frame-options handling for artifacts 6. Technical Improvements: - Better post streaming implementation - Improved error handling in completions - Better memory management for partial tool calls - Enhanced testing coverage 7. Configuration: - New site settings for artifact security - Extended LLM model configurations - Additional tool configuration options This PR significantly enhances the plugin's capabilities for generating and displaying interactive content while maintaining security and providing flexible configuration options for administrators.
2024-11-18 17:22:39 -05:00
type: "string",
},
],
}
end
def self.allow_partial_tool_calls?
true
end
def partial_invoke
@selected_tab = :html_body
if @prev_parameters
@selected_tab = parameters.keys.find { |k| @prev_parameters[k] != parameters[k] }
end
update_custom_html
@prev_parameters = parameters.dup
end
def invoke
yield parameters[:name] || "Web Artifact"
# Get the current post from context
post = Post.find_by(id: context[:post_id])
return error_response("No post context found") unless post
html = parameters[:html_body].to_s
css = parameters[:css].to_s
js = parameters[:js].to_s
# Create the artifact
artifact =
AiArtifact.new(
user_id: bot_user.id,
post_id: post.id,
name: parameters[:name].to_s[0...255],
html: html,
css: css,
js: js,
metadata: parameters[:metadata],
)
if artifact.save
update_custom_html(artifact)
success_response(artifact)
else
error_response(artifact.errors.full_messages.join(", "))
end
end
def chain_next_response?
@chain_next_response
end
private
def update_custom_html(artifact = nil)
html = parameters[:html_body].to_s
css = parameters[:css].to_s
js = parameters[:js].to_s
artifact_div =
"<div class=\"ai-artifact\" data-ai-artifact-id=\"#{artifact.id}\"></div>" if artifact
FEATURE: AI artifacts (#898) This is a significant PR that introduces AI Artifacts functionality to the discourse-ai plugin along with several other improvements. Here are the key changes: 1. AI Artifacts System: - Adds a new `AiArtifact` model and database migration - Allows creation of web artifacts with HTML, CSS, and JavaScript content - Introduces security settings (`strict`, `lax`, `disabled`) for controlling artifact execution - Implements artifact rendering in iframes with sandbox protection - New `CreateArtifact` tool for AI to generate interactive content 2. Tool System Improvements: - Adds support for partial tool calls, allowing incremental updates during generation - Better handling of tool call states and progress tracking - Improved XML tool processing with CDATA support - Fixes for tool parameter handling and duplicate invocations 3. LLM Provider Updates: - Updates for Anthropic Claude models with correct token limits - Adds support for native/XML tool modes in Gemini integration - Adds new model configurations including Llama 3.1 models - Improvements to streaming response handling 4. UI Enhancements: - New artifact viewer component with expand/collapse functionality - Security controls for artifact execution (click-to-run in strict mode) - Improved dialog and response handling - Better error management for tool execution 5. Security Improvements: - Sandbox controls for artifact execution - Public/private artifact sharing controls - Security settings to control artifact behavior - CSP and frame-options handling for artifacts 6. Technical Improvements: - Better post streaming implementation - Improved error handling in completions - Better memory management for partial tool calls - Enhanced testing coverage 7. Configuration: - New site settings for artifact security - Extended LLM model configurations - Additional tool configuration options This PR significantly enhances the plugin's capabilities for generating and displaying interactive content while maintaining security and providing flexible configuration options for administrators.
2024-11-18 17:22:39 -05:00
content = []
content << [:html_body, "### HTML\n\n```html\n#{html}\n```"] if html.present?
content << [:css, "### CSS\n\n```css\n#{css}\n```"] if css.present?
content << [:js, "### JavaScript\n\n```javascript\n#{js}\n```"] if js.present?
content.sort_by! { |c| c[0] === @selected_tab ? 1 : 0 } if !artifact
if artifact
content.unshift([nil, "[details='#{I18n.t("discourse_ai.ai_artifact.view_source")}']"])
content << [nil, "[/details]"]
end
content << [:preview, "### Preview\n\n#{artifact_div}"] if artifact_div
FEATURE: AI artifacts (#898) This is a significant PR that introduces AI Artifacts functionality to the discourse-ai plugin along with several other improvements. Here are the key changes: 1. AI Artifacts System: - Adds a new `AiArtifact` model and database migration - Allows creation of web artifacts with HTML, CSS, and JavaScript content - Introduces security settings (`strict`, `lax`, `disabled`) for controlling artifact execution - Implements artifact rendering in iframes with sandbox protection - New `CreateArtifact` tool for AI to generate interactive content 2. Tool System Improvements: - Adds support for partial tool calls, allowing incremental updates during generation - Better handling of tool call states and progress tracking - Improved XML tool processing with CDATA support - Fixes for tool parameter handling and duplicate invocations 3. LLM Provider Updates: - Updates for Anthropic Claude models with correct token limits - Adds support for native/XML tool modes in Gemini integration - Adds new model configurations including Llama 3.1 models - Improvements to streaming response handling 4. UI Enhancements: - New artifact viewer component with expand/collapse functionality - Security controls for artifact execution (click-to-run in strict mode) - Improved dialog and response handling - Better error management for tool execution 5. Security Improvements: - Sandbox controls for artifact execution - Public/private artifact sharing controls - Security settings to control artifact behavior - CSP and frame-options handling for artifacts 6. Technical Improvements: - Better post streaming implementation - Improved error handling in completions - Better memory management for partial tool calls - Enhanced testing coverage 7. Configuration: - New site settings for artifact security - Extended LLM model configurations - Additional tool configuration options This PR significantly enhances the plugin's capabilities for generating and displaying interactive content while maintaining security and providing flexible configuration options for administrators.
2024-11-18 17:22:39 -05:00
self.custom_raw = content.map { |c| c[1] }.join("\n\n")
end
def success_response(artifact)
@chain_next_response = false
{
status: "success",
artifact_id: artifact.id,
message: "Artifact created successfully and rendered to user.",
}
end
def error_response(message)
@chain_next_response = false
{ status: "error", error: message }
end
def help
"Creates a web artifact with HTML, CSS, and JavaScript that can be displayed in an iframe. " \
"Requires a name and HTML content. CSS and JavaScript are optional. " \
"The artifact will be associated with the current post and can be displayed using an iframe."
end
end
end
end
end