Implement Drupal Recipe Integration for AI Functionalities Configuration

[robertoperuzzo]

Is there an existing issue for this?

• I have searched the existing issues

Is your feature request related to a problem?

Currently, the ddev drupal-ai setup wizard reads the drupal_modules configuration from functionalities.yaml to determine which modules to install for each selected functionality. This approach is reductive because:

1. Limited Configuration Scope: Installing modules alone doesn't provide a fully functional Drupal AI environment
2. Manual Configuration Required: Users must manually configure each module after installation
3. No Default Settings: There's no way to automatically set up sensible defaults for AI tools
4. Inconsistent Experience: Each functionality requires different manual configuration steps that aren't documented or automated

The current approach forces users to:

• Manually configure each AI module after installation
• Set up content types, fields, and configurations specific to AI functionalities
• Configure search indexes, vector storage, and other complex integrations
• Apply security and permission settings manually

Describe your solution

Implement Drupal Recipe integration alongside the current module-based configuration system by leveraging the existing Drupal recipes ecosystem. The solution should:

1. Recipe Mapping

Utilize existing Drupal recipes or create new ones for each functionality defined in functionalities.yaml:

• vector-search → Use/create vector search recipe with pre-configured Search API, content types, and pgvector integration
• content-generation → Use/create content generation recipe with editorial workflows and AI content types
• image-analysis → Use/create image analysis recipe with media handling and AI vision integration
• qa-system → Use/create Q&A system recipe with knowledge base structure and search integration
• code-assistant → Use/create code assistant recipe with code snippet management and syntax highlighting
• document-processing → Use/create document processing recipe with file handling and extraction workflows

2. Enhanced Configuration Structure

Extend functionalities.yaml to include recipe references:

functionalities:
  vector-search:
    name: "Vector Search & Embeddings"
    description: "Semantic search using vector databases"
    required_addons: ["robertoperuzzo/ddev-pgvector"]
    required_capabilities: ["embeddings"]
    drupal_modules: ["ai", "ai_search", "search_api"]  # Keep existing for module installation
    drupal_recipes: ["drupal/ai-vector-search"]        # New: recipes to apply
    recipe_config_overrides:                           # New: custom config merging
      search_api:
        default_server: "pgvector_server"

3. Recipe Ecosystem Integration

• First: Research and leverage existing recipes from the Drupal recipes ecosystem
• Second: Create new recipes only where gaps exist in the ecosystem
• Third: Contribute new recipes back to the community

4. Merge Configuration Logic

During ddev drupal-ai setup, combine:

• Recipe-provided default configurations (from ecosystem recipes)
• User-selected customizations from the CLI wizard
• Environment-specific settings (API keys, endpoints, etc.)

5. Enhanced CLI Wizard

Extend the setup process to:

• Install the appropriate Drupal recipe(s) based on selected functionalities
• Apply recipe configurations as a baseline
• Overlay user customizations on top of recipe defaults
• Handle configuration conflicts gracefully

Describe alternatives

Alternative approaches considered:

1. Custom Configuration Profiles

Create our own configuration management system instead of using Drupal recipes

• Pros: Full control over configuration structure
• Cons: Reinventing the wheel, not leveraging community standards, maintenance burden

2. Enhanced Module Approach

Extend modules to include more opinionated defaults

• Pros: Simpler implementation, no recipe dependencies
• Cons: Still requires manual configuration, doesn't leverage Drupal's configuration management best practices

3. Recipe-Only Approach

Replace module configuration entirely with recipes

• Pros: Cleaner architecture, full configuration management
• Cons: May not provide granular control, dependency on recipe availability

Additional context

Implementation Strategy

1. Recipe Research Phase

• Audit existing Drupal recipes ecosystem for AI-related recipes
• Identify gaps where new recipes need to be created
• Document recipe dependencies and compatibility

2. Recipe Development Phase

• Create missing recipes for AI functionalities
• Ensure recipes follow Drupal community best practices
• Include comprehensive configuration for each AI use case

3. Integration Phase

• Extend functionalities.yaml structure to include recipe references
• Modify CLI wizard to handle recipe installation
• Implement configuration merging logic

4. Community Contribution

• Submit new recipes to appropriate Drupal recipe repositories
• Document recipe usage patterns for the community

Expected Benefits

• Faster Setup: Users get fully configured AI environments out of the box
• Community Standards: Leverage and contribute to Drupal recipes ecosystem
• Better Defaults: Use community-vetted configurations and best practices
• Reduced Errors: Less manual configuration means fewer setup mistakes
• Improved UX: Seamless transition from infrastructure setup to working AI tools
• Ecosystem Growth: Contribute back to Drupal AI recipe ecosystem

Related Files to Modify

• drupal-ai/configs/functionalities.yaml - Add recipe references
• commands/web/drupal-ai - Enhance setup wizard with recipe installation
• README.md - Update documentation to reflect recipe integration
• Template files in drupal-ai/templates/ - Add recipe-specific templates

Summary

This enhancement would transform the setup from "install modules and configure manually" to "install recipes and get a working AI-powered Drupal site with sensible defaults immediately."

[robertoperuzzo]

According to this issue we could implement a way to discover which are the AI recipes available supposing that each recipe represent an AI functionality.

This API might be useful, but I don't know how to filter by "recipe" only projects.

curl "https://www.drupal.org/api-d7/node.json?type=project_module&taxonomy_vocabulary_3=204588&limit=50"

• Found taxonomy term "Artificial Intelligence (AI)" — tid: 204588 (vocabulary id 3).
• Sample project_module node JSON shows the taxonomy field appears as taxonomy_vocabulary_3.

[robertoperuzzo]

reference to an existing AI recipe https://www.drupal.org/project/ai_ecosystem