breakpilot-lehrer/edu-search-service/internal/database/repository.go

package database

import (
	"context"

	"github.com/google/uuid"
	"github.com/jackc/pgx/v5"
)

// Repository provides database operations for staff and publications
type Repository struct {
	db *DB
}

// NewRepository creates a new repository
func NewRepository(db *DB) *Repository {
	return &Repository{db: db}
}

// ============================================================================
// UNIVERSITIES
// ============================================================================

// CreateUniversity creates a new university
func (r *Repository) CreateUniversity(ctx context.Context, u *University) error {
	query := `
		INSERT INTO universities (name, short_name, url, state, uni_type, staff_page_pattern)
		VALUES ($1, $2, $3, $4, $5, $6)
		ON CONFLICT (url) DO UPDATE SET
			name = EXCLUDED.name,
			short_name = EXCLUDED.short_name,
			state = EXCLUDED.state,
			uni_type = EXCLUDED.uni_type,
			staff_page_pattern = EXCLUDED.staff_page_pattern,
			updated_at = NOW()
		RETURNING id, created_at, updated_at
	`
	return r.db.Pool.QueryRow(ctx, query,
		u.Name, u.ShortName, u.URL, u.State, u.UniType, u.StaffPagePattern,
	).Scan(&u.ID, &u.CreatedAt, &u.UpdatedAt)
}

// GetUniversity retrieves a university by ID
func (r *Repository) GetUniversity(ctx context.Context, id uuid.UUID) (*University, error) {
	query := `SELECT id, name, short_name, url, state, uni_type, staff_page_pattern, created_at, updated_at
		FROM universities WHERE id = $1`

	u := &University{}
	err := r.db.Pool.QueryRow(ctx, query, id).Scan(
		&u.ID, &u.Name, &u.ShortName, &u.URL, &u.State, &u.UniType,
		&u.StaffPagePattern, &u.CreatedAt, &u.UpdatedAt,
	)
	if err == pgx.ErrNoRows {
		return nil, nil
	}
	if err != nil {
		return nil, err
	}
	return u, nil
}

// GetUniversityByID is an alias for GetUniversity (for interface compatibility)
func (r *Repository) GetUniversityByID(ctx context.Context, id uuid.UUID) (*University, error) {
	return r.GetUniversity(ctx, id)
}

// GetUniversityByURL retrieves a university by URL
func (r *Repository) GetUniversityByURL(ctx context.Context, url string) (*University, error) {
	query := `SELECT id, name, short_name, url, state, uni_type, staff_page_pattern, created_at, updated_at
		FROM universities WHERE url = $1`

	u := &University{}
	err := r.db.Pool.QueryRow(ctx, query, url).Scan(
		&u.ID, &u.Name, &u.ShortName, &u.URL, &u.State, &u.UniType,
		&u.StaffPagePattern, &u.CreatedAt, &u.UpdatedAt,
	)
	if err != nil {
		return nil, err
	}
	return u, nil
}

// ListUniversities lists all universities
func (r *Repository) ListUniversities(ctx context.Context) ([]University, error) {
	query := `SELECT id, name, short_name, url, state, uni_type, staff_page_pattern, created_at, updated_at
		FROM universities ORDER BY name`

	rows, err := r.db.Pool.Query(ctx, query)
	if err != nil {
		return nil, err
	}
	defer rows.Close()

	var universities []University
	for rows.Next() {
		var u University
		if err := rows.Scan(
			&u.ID, &u.Name, &u.ShortName, &u.URL, &u.State, &u.UniType,
			&u.StaffPagePattern, &u.CreatedAt, &u.UpdatedAt,
		); err != nil {
			return nil, err
		}
		universities = append(universities, u)
	}
	return universities, rows.Err()
}

// ============================================================================
// DEPARTMENTS
// ============================================================================

// CreateDepartment creates or updates a department
func (r *Repository) CreateDepartment(ctx context.Context, d *Department) error {
	query := `
		INSERT INTO departments (university_id, name, name_en, url, category, parent_id)
		VALUES ($1, $2, $3, $4, $5, $6)
		ON CONFLICT (university_id, name) DO UPDATE SET
			name_en = EXCLUDED.name_en,
			url = EXCLUDED.url,
			category = EXCLUDED.category,
			parent_id = EXCLUDED.parent_id,
			updated_at = NOW()
		RETURNING id, created_at, updated_at
	`
	return r.db.Pool.QueryRow(ctx, query,
		d.UniversityID, d.Name, d.NameEN, d.URL, d.Category, d.ParentID,
	).Scan(&d.ID, &d.CreatedAt, &d.UpdatedAt)
}

// GetDepartmentByName retrieves a department by university and name
func (r *Repository) GetDepartmentByName(ctx context.Context, uniID uuid.UUID, name string) (*Department, error) {
	query := `SELECT id, university_id, name, name_en, url, category, parent_id, created_at, updated_at
		FROM departments WHERE university_id = $1 AND name = $2`

	d := &Department{}
	err := r.db.Pool.QueryRow(ctx, query, uniID, name).Scan(
		&d.ID, &d.UniversityID, &d.Name, &d.NameEN, &d.URL, &d.Category,
		&d.ParentID, &d.CreatedAt, &d.UpdatedAt,
	)
	if err != nil {
		return nil, err
	}
	return d, nil
}

// ============================================================================
// CRAWL STATUS
// ============================================================================

// UpdateCrawlStatus updates crawl status for a university
func (r *Repository) UpdateCrawlStatus(ctx context.Context, status *UniversityCrawlStatus) error {
	query := `
		INSERT INTO university_crawl_status (
			university_id, last_staff_crawl, staff_crawl_status, staff_count, staff_errors,
			last_pub_crawl, pub_crawl_status, pub_count, pub_errors,
			next_scheduled_crawl, crawl_priority
		) VALUES ($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11)
		ON CONFLICT (university_id) DO UPDATE SET
			last_staff_crawl = EXCLUDED.last_staff_crawl,
			staff_crawl_status = EXCLUDED.staff_crawl_status,
			staff_count = EXCLUDED.staff_count,
			staff_errors = EXCLUDED.staff_errors,
			last_pub_crawl = EXCLUDED.last_pub_crawl,
			pub_crawl_status = EXCLUDED.pub_crawl_status,
			pub_count = EXCLUDED.pub_count,
			pub_errors = EXCLUDED.pub_errors,
			next_scheduled_crawl = EXCLUDED.next_scheduled_crawl,
			crawl_priority = EXCLUDED.crawl_priority,
			updated_at = NOW()
	`
	_, err := r.db.Pool.Exec(ctx, query,
		status.UniversityID, status.LastStaffCrawl, status.StaffCrawlStatus, status.StaffCount, status.StaffErrors,
		status.LastPubCrawl, status.PubCrawlStatus, status.PubCount, status.PubErrors,
		status.NextScheduledCrawl, status.CrawlPriority,
	)
	return err
}

// GetCrawlStatus retrieves crawl status for a university
func (r *Repository) GetCrawlStatus(ctx context.Context, uniID uuid.UUID) (*UniversityCrawlStatus, error) {
	query := `SELECT * FROM university_crawl_status WHERE university_id = $1`

	s := &UniversityCrawlStatus{}
	err := r.db.Pool.QueryRow(ctx, query, uniID).Scan(
		&s.UniversityID, &s.LastStaffCrawl, &s.StaffCrawlStatus, &s.StaffCount, &s.StaffErrors,
		&s.LastPubCrawl, &s.PubCrawlStatus, &s.PubCount, &s.PubErrors,
		&s.NextScheduledCrawl, &s.CrawlPriority, &s.CreatedAt, &s.UpdatedAt,
	)
	if err == pgx.ErrNoRows {
		return nil, nil
	}
	if err != nil {
		return nil, err
	}
	return s, nil
}

// ============================================================================
// STATS
// ============================================================================

// GetStaffStats retrieves statistics about staff data
func (r *Repository) GetStaffStats(ctx context.Context) (*StaffStats, error) {
	stats := &StaffStats{
		ByState:        make(map[string]int),
		ByUniType:      make(map[string]int),
		ByPositionType: make(map[string]int),
	}

	// Basic counts
	queries := []struct {
		query string
		dest  *int
	}{
		{"SELECT COUNT(*) FROM university_staff WHERE is_active = true", &stats.TotalStaff},
		{"SELECT COUNT(*) FROM university_staff WHERE is_professor = true AND is_active = true", &stats.TotalProfessors},
		{"SELECT COUNT(*) FROM publications", &stats.TotalPublications},
		{"SELECT COUNT(*) FROM universities", &stats.TotalUniversities},
	}

	for _, q := range queries {
		if err := r.db.Pool.QueryRow(ctx, q.query).Scan(q.dest); err != nil {
			return nil, err
		}
	}

	// By state
	rows, err := r.db.Pool.Query(ctx, `
		SELECT COALESCE(u.state, 'unknown'), COUNT(*)
		FROM university_staff s
		JOIN universities u ON s.university_id = u.id
		WHERE s.is_active = true
		GROUP BY u.state
	`)
	if err != nil {
		return nil, err
	}
	defer rows.Close()

	for rows.Next() {
		var state string
		var count int
		if err := rows.Scan(&state, &count); err != nil {
			return nil, err
		}
		stats.ByState[state] = count
	}

	// By uni type
	rows2, err := r.db.Pool.Query(ctx, `
		SELECT COALESCE(u.uni_type, 'unknown'), COUNT(*)
		FROM university_staff s
		JOIN universities u ON s.university_id = u.id
		WHERE s.is_active = true
		GROUP BY u.uni_type
	`)
	if err != nil {
		return nil, err
	}
	defer rows2.Close()

	for rows2.Next() {
		var uniType string
		var count int
		if err := rows2.Scan(&uniType, &count); err != nil {
			return nil, err
		}
		stats.ByUniType[uniType] = count
	}

	// By position type
	rows3, err := r.db.Pool.Query(ctx, `
		SELECT COALESCE(position_type, 'unknown'), COUNT(*)
		FROM university_staff
		WHERE is_active = true
		GROUP BY position_type
	`)
	if err != nil {
		return nil, err
	}
	defer rows3.Close()

	for rows3.Next() {
		var posType string
		var count int
		if err := rows3.Scan(&posType, &count); err != nil {
			return nil, err
		}
		stats.ByPositionType[posType] = count
	}

	return stats, nil
}