"""Set multilingual glossaries target_language to 'multi'

Revision ID: a1b2c3d4e5f6
Revises: e5b2c9d1f4a8
Create Date: 2026-05-31

Glossary templates that were enriched with multilingual translations
(via enrich_glossary_templates.py) contain translations for 11 languages
(de, es, it, pt, nl, ru, ja, ko, zh, ar, fa) in each term's translations
field. These should be marked as target_language='multi' instead of 'en'.

This migration detects glossaries whose terms have multilingual translations
and sets their target_language to 'multi'.
"""

from typing import Sequence, Union

from alembic import op
import sqlalchemy as sa

# revision identifiers
revision = "a1b2c3d4e5f6"
down_revision = "e5b2c9d1f4a8"
branch_labels: Union[str, Sequence[str], None] = None
depends_on: Union[str, Sequence[str], None] = None


def upgrade() -> None:
    # Glossaries with terms containing 5+ translation keys are multilingual templates
    # (enriched glossaries have 11 translations: de, es, it, pt, nl, ru, ja, ko, zh, ar, fa)
    op.execute("""
        UPDATE glossaries
        SET target_language = 'multi'
        WHERE id IN (
            SELECT DISTINCT g.id
            FROM glossaries g
            JOIN glossary_terms gt ON gt.glossary_id = g.id
            WHERE gt.translations IS NOT NULL
              AND jsonb_typeof(gt.translations) = 'object'
              AND (
                SELECT count(*)
                FROM jsonb_object_keys(gt.translations)
              ) >= 5
        )
    """)


def downgrade() -> None:
    # Revert multilingual glossaries back to 'en'
    op.execute("""
        UPDATE glossaries
        SET target_language = 'en'
        WHERE target_language = 'multi'
    """)