-- Sample data for testing SanadCheck LLM API
-- Execute after creating tables

-- Insert sample users (these will be created through the auth system)
-- This is just for reference - actual users will be created via the API

-- Sample narrator extractions for testing
INSERT INTO public.narrator_extractions (
    id,
    user_id,
    hadith_text,
    extracted_narrators,
    sanad_chain,
    success,
    processing_time_ms,
    created_at
) VALUES 
-- Note: Replace with actual user UUIDs after creating users through the API
-- (
--     uuid_generate_v4(),
--     'user-uuid-here',
--     'حدثنا محمد بن إسماعيل قال حدثنا عبد الله بن موسى قال أخبرنا إسرائيل عن أبي إسحاق عن البراء قال...',
--     ARRAY['محمد بن إسماعيل', 'عبد الله بن موسى', 'إسرائيل', 'أبو إسحاق', 'البراء'],
--     'محمد بن إسماعيل ← عبد الله بن موسى ← إسرائيل ← أبي إسحاق ← البراء',
--     true,
--     1500,
--     NOW() - INTERVAL '1 day'
-- );

-- Create some utility functions for data analysis

-- Function to get extraction statistics
CREATE OR REPLACE FUNCTION public.get_extraction_stats()
RETURNS TABLE (
    total_extractions BIGINT,
    successful_extractions BIGINT,
    failed_extractions BIGINT,
    success_rate NUMERIC,
    avg_processing_time NUMERIC,
    total_users BIGINT
) AS $$
BEGIN
    RETURN QUERY
    SELECT 
        COUNT(*) as total_extractions,
        COUNT(*) FILTER (WHERE success = true) as successful_extractions,
        COUNT(*) FILTER (WHERE success = false) as failed_extractions,
        ROUND(
            (COUNT(*) FILTER (WHERE success = true)::NUMERIC / COUNT(*)::NUMERIC) * 100, 
            2
        ) as success_rate,
        ROUND(AVG(processing_time_ms), 2) as avg_processing_time,
        COUNT(DISTINCT user_id) as total_users
    FROM public.narrator_extractions;
END;
$$ LANGUAGE plpgsql;

-- Function to get most analyzed narrators
CREATE OR REPLACE FUNCTION public.get_popular_narrators(limit_count INTEGER DEFAULT 10)
RETURNS TABLE (
    narrator_name TEXT,
    analysis_count BIGINT,
    avg_reliability_grade TEXT,
    most_common_grade TEXT
) AS $$
BEGIN
    RETURN QUERY
    SELECT 
        na.narrator_name,
        COUNT(*) as analysis_count,
        MODE() WITHIN GROUP (ORDER BY na.reliability_grade) as most_common_grade,
        ROUND(AVG(
            CASE 
                WHEN na.reliability_grade = 'Thiqah' THEN 5
                WHEN na.reliability_grade = 'Saduq' THEN 4
                WHEN na.reliability_grade = 'Da''if' THEN 2
                WHEN na.reliability_grade = 'Matruk' THEN 1
                ELSE 0
            END
        ), 2)::TEXT as avg_reliability_grade
    FROM public.narrator_analyses na
    WHERE na.success = true
    GROUP BY na.narrator_name
    ORDER BY analysis_count DESC
    LIMIT limit_count;
END;
$$ LANGUAGE plpgsql;

-- Function to get user activity summary
CREATE OR REPLACE FUNCTION public.get_user_activity(user_uuid UUID)
RETURNS TABLE (
    total_extractions BIGINT,
    successful_extractions BIGINT,
    total_analyses BIGINT,
    successful_analyses BIGINT,
    first_activity TIMESTAMP WITH TIME ZONE,
    last_activity TIMESTAMP WITH TIME ZONE,
    avg_processing_time NUMERIC
) AS $$
BEGIN
    RETURN QUERY
    SELECT 
        COUNT(ne.id) as total_extractions,
        COUNT(ne.id) FILTER (WHERE ne.success = true) as successful_extractions,
        COUNT(na.id) as total_analyses,
        COUNT(na.id) FILTER (WHERE na.success = true) as successful_analyses,
        MIN(LEAST(ne.created_at, na.created_at)) as first_activity,
        MAX(GREATEST(ne.created_at, na.created_at)) as last_activity,
        ROUND(AVG(COALESCE(ne.processing_time_ms, na.processing_time_ms)), 2) as avg_processing_time
    FROM public.narrator_extractions ne
    FULL OUTER JOIN public.narrator_analyses na ON ne.user_id = na.user_id
    WHERE ne.user_id = user_uuid OR na.user_id = user_uuid;
END;
$$ LANGUAGE plpgsql;

-- Create indexes for common queries
CREATE INDEX IF NOT EXISTS idx_narrator_extractions_text_search ON public.narrator_extractions 
USING gin(to_tsvector('arabic', hadith_text));

CREATE INDEX IF NOT EXISTS idx_narrator_analyses_name_search ON public.narrator_analyses 
USING gin(to_tsvector('arabic', narrator_name));

-- Grant execute permissions on functions
GRANT EXECUTE ON FUNCTION public.get_extraction_stats() TO authenticated;
GRANT EXECUTE ON FUNCTION public.get_popular_narrators(INTEGER) TO authenticated;
GRANT EXECUTE ON FUNCTION public.get_user_activity(UUID) TO authenticated;

-- Create a cleanup function for old sessions
CREATE OR REPLACE FUNCTION public.cleanup_expired_sessions()
RETURNS INTEGER AS $$
DECLARE
    deleted_count INTEGER;
BEGIN
    -- Delete sessions that have been expired for more than 7 days
    DELETE FROM public.user_sessions 
    WHERE expires_at < NOW() - INTERVAL '7 days';
    
    GET DIAGNOSTICS deleted_count = ROW_COUNT;
    RETURN deleted_count;
END;
$$ LANGUAGE plpgsql;

-- Grant execute permission
GRANT EXECUTE ON FUNCTION public.cleanup_expired_sessions() TO authenticated;

-- Comments for documentation
COMMENT ON TABLE public.users IS 'Extended user profiles linked to Supabase auth.users';
COMMENT ON TABLE public.user_sessions IS 'JWT session tracking for security and analytics';
COMMENT ON TABLE public.narrator_extractions IS 'Records of hadith narrator extraction requests and results';
COMMENT ON TABLE public.narrator_analyses IS 'Records of individual narrator analysis requests and results';

COMMENT ON FUNCTION public.get_extraction_stats() IS 'Returns overall statistics about extraction operations';
COMMENT ON FUNCTION public.get_popular_narrators(INTEGER) IS 'Returns most frequently analyzed narrators';
COMMENT ON FUNCTION public.get_user_activity(UUID) IS 'Returns activity summary for a specific user';
COMMENT ON FUNCTION public.cleanup_expired_sessions() IS 'Removes old expired sessions to keep the database clean';