The start-curriculum command now runs interactively by default! Simply run:
uv run start-curriculum
The script has sensible defaults set at the top:
DEFAULT_DOMAIN = "biochemistry"
DEFAULT_ENTITY = "karl_friston"
DEFAULT_LANGUAGE = "Spanish"
When you run the script, you’ll see prompts like:
🎯 Active Inference Curriculum Generator
==================================================
Press Enter to use default values, or type your custom choice.
📚 Select Domain (default: biochemistry)
Available domains: biochemistry, neuroscience, artificial_intelligence, psychology...
💡 Tip: You can also enter a new domain name to create a custom domain!
Domain [biochemistry]:
👤 Select Target Entity (default: karl_friston)
Available entities: karl_friston, tulsi_gabbard, elon_musk...
💡 Tip: You can also enter a new entity name to create a custom target audience!
Entity [karl_friston]:
🌍 Select Target Language (default: Spanish)
Available languages: Spanish, French, Chinese, Arabic, Hindi...
💡 Tip: You can also enter a new language name to create a custom translation target!
Language [Spanish]:
✅ Selected Configuration:
Domain: biochemistry
Entity: karl_friston
Language: Spanish
Proceed with this configuration? [Y/n]:
The script now allows you to create custom domains, entities, and languages on the fly!
Domain [biochemistry]: art
🆕 'art' is not in the existing domains.
Create this as a new custom domain? [y/N]: y
✅ Created custom domain: 'art'
Entity [karl_friston]: artist
🆕 'artist' is not in the existing entities.
Create this as a new custom entity? [y/N]: y
Enter a brief description for 'artist': Professional visual artists working in galleries
✅ Created custom entity: 'artist' - Professional visual artists working in galleries
Language [Spanish]: Portuguese
🆕 'Portuguese' is not in the existing languages.
Create this as a new custom language? [y/N]: y
✅ Created custom language: 'Portuguese'
🔤 Note: Translation quality may vary for custom languages not in the AI model's training.
uv run start-curriculum
# Just press Enter three times, then Y to confirm
uv run start-curriculum
# Type: neuroscience
# Press Enter for entity (karl_friston)
# Press Enter for language (Spanish)
# Type: Y to confirm
uv run start-curriculum
# Type: artificial_intelligence
# Type: elon_musk
# Type: Chinese
# Type: Y to confirm
uv run start-curriculum
# Type: art (new custom domain)
# Type: y to create custom domain
# Type: curator (new custom entity)
# Type: y to create custom entity
# Type: Museum curator specializing in contemporary art
# Type: Portuguese (new custom language)
# Type: y to create custom language
# Type: Y to confirm and proceed
You can still use command-line arguments to skip interactive mode:
# Use specific configuration
uv run start-curriculum --domains neuroscience --entities tulsi_gabbard --languages French
# Force non-interactive with defaults
uv run start-curriculum --non-interactive
# Force interactive mode even with args
uv run start-curriculum --interactive
After confirming your configuration, the script will:
The script reads from these configuration files:
data/config/domains.yaml - Available domainsdata/config/entities.yaml - Available entities/audiencesdata/config/languages.yaml - Available target languagesexport PERPLEXITY_API_KEY="your-key"
export OPENROUTER_API_KEY="your-key"
Check Available Options: The script shows available choices from your config files
Start Simple: Use defaults first to test the pipeline, then customize
Monitor Progress: The script provides detailed logging throughout the process
data/domain_research/ - Domain analysisdata/audience_research/ - Entity analysisdata/written_curriculums/ - Generated curriculadata/visualizations/ - Charts and diagramsdata/translated_curriculums/ - TranslationsInvalid Domain/Entity/Language: The script validates your inputs against available options in the config files. If you get an error, check the available options shown in the prompt.
API Errors: Ensure your API keys are properly set and have sufficient credits.
Configuration is required: The script reads domains, entities, and languages from data/config/ and reports missing or malformed files.
Enjoy creating personalized Active Inference curricula! 🎯