Database Models
===============

Overview
--------

Database models are organized into domain-specific modules under ``ai4drpm.db.models``:

* **auth/** - User and authentication models (User, RefreshToken)
* **resources/** - Legal resources and provisions (LegalResource, LegalProvision, LegalProvisionClassification)
* **analyses/** - Analysis types (DescriptionAnalysis, DataAnalysis, SolutionsAnalysis, InteroperabilityAnalysis, DataFlowsAnalysis)
* **categories/** - Hierarchical category taxonomy
* **classifiers/** - ML classifier metadata and training data
* **prompts/** - LLM prompt templates
* **pipelines/** - Haystack pipeline definitions and executions
* **queries/** - SPARQL query storage
* **tasks/** - Celery task tracking
* **token_usage** - LLM token usage tracking

Model Architecture
------------------

Base Model
~~~~~~~~~~

All models inherit from SQLAlchemy's declarative base defined in ``ai4drpm.db.base``.

Relationships
~~~~~~~~~~~~~

Models use SQLAlchemy relationships for:

* Foreign key relationships
* One-to-many and many-to-one associations
* Many-to-many through association tables
* Lazy and eager loading configurations

Association Tables
~~~~~~~~~~~~~~~~~~

Many-to-many relationships are managed through association tables defined alongside their related models.

Model Domains
-------------

.. toctree::
   :maxdepth: 2

   auth
   resources
   analyses
   categories
   classifiers
   prompts
   pipelines
   tasks