Files
breakpilot-lehrer/edu-search-service/internal/orchestrator/audiences.go
Benjamin Boenisch 414e0f5ec0
All checks were successful
CI / go-lint (push) Has been skipped
CI / python-lint (push) Has been skipped
CI / nodejs-lint (push) Has been skipped
CI / test-go-school (push) Successful in 28s
CI / test-go-edu-search (push) Successful in 27s
CI / test-python-klausur (push) Successful in 1m45s
CI / test-python-agent-core (push) Successful in 16s
CI / test-nodejs-website (push) Successful in 21s
feat: edu-search-service migriert, voice-service/geo-service entfernt
- edu-search-service von breakpilot-pwa nach breakpilot-lehrer kopiert (ohne vendor)
- opensearch + edu-search-service in docker-compose.yml hinzugefuegt
- voice-service aus docker-compose.yml entfernt (jetzt in breakpilot-core)
- geo-service aus docker-compose.yml entfernt (nicht mehr benoetigt)
- CI/CD: edu-search-service zu Gitea Actions und Woodpecker hinzugefuegt
  (Go lint, test mit go mod download, build, SBOM)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-15 18:36:38 +01:00

425 lines
13 KiB
Go

// Package orchestrator implements multi-phase university crawling with queue management
package orchestrator
import (
"context"
"encoding/json"
"fmt"
"time"
"github.com/google/uuid"
)
// Audience represents a target audience filter configuration
type Audience struct {
ID uuid.UUID `json:"id"`
Name string `json:"name"`
Description string `json:"description,omitempty"`
Filters AudienceFilters `json:"filters"`
MemberCount int `json:"member_count"`
LastCountUpdate *time.Time `json:"last_count_update,omitempty"`
CreatedBy string `json:"created_by,omitempty"`
IsActive bool `json:"is_active"`
CreatedAt time.Time `json:"created_at"`
UpdatedAt time.Time `json:"updated_at"`
}
// AudienceFilters defines the filter criteria for an audience
type AudienceFilters struct {
PositionTypes []string `json:"position_types,omitempty"` // professor, researcher, lecturer
SubjectAreas []uuid.UUID `json:"subject_areas,omitempty"` // Subject area UUIDs
States []string `json:"states,omitempty"` // BW, BY, etc.
UniTypes []string `json:"uni_types,omitempty"` // UNI, PH, HAW
Universities []uuid.UUID `json:"universities,omitempty"` // University UUIDs
HasEmail *bool `json:"has_email,omitempty"`
IsActive *bool `json:"is_active,omitempty"`
Keywords []string `json:"keywords,omitempty"` // Keywords in name/research
}
// AudienceExport tracks exports of audience data
type AudienceExport struct {
ID uuid.UUID `json:"id"`
AudienceID uuid.UUID `json:"audience_id"`
ExportType string `json:"export_type"` // csv, json, email_list
RecordCount int `json:"record_count"`
FilePath string `json:"file_path,omitempty"`
ExportedBy string `json:"exported_by,omitempty"`
Purpose string `json:"purpose,omitempty"`
CreatedAt time.Time `json:"created_at"`
}
// AudienceMember represents a staff member in an audience preview
type AudienceMember struct {
ID uuid.UUID `json:"id"`
Name string `json:"name"`
Email string `json:"email,omitempty"`
Position string `json:"position,omitempty"`
University string `json:"university"`
Department string `json:"department,omitempty"`
SubjectArea string `json:"subject_area,omitempty"`
PublicationCount int `json:"publication_count"`
}
// AudienceRepository extends Repository with audience operations
type AudienceRepository interface {
// Audience CRUD
CreateAudience(ctx context.Context, audience *Audience) error
GetAudience(ctx context.Context, id uuid.UUID) (*Audience, error)
ListAudiences(ctx context.Context, activeOnly bool) ([]Audience, error)
UpdateAudience(ctx context.Context, audience *Audience) error
DeleteAudience(ctx context.Context, id uuid.UUID) error
// Audience members
GetAudienceMembers(ctx context.Context, id uuid.UUID, limit, offset int) ([]AudienceMember, int, error)
UpdateAudienceCount(ctx context.Context, id uuid.UUID) (int, error)
// Exports
CreateExport(ctx context.Context, export *AudienceExport) error
ListExports(ctx context.Context, audienceID uuid.UUID) ([]AudienceExport, error)
}
// ============================================================================
// POSTGRES IMPLEMENTATION
// ============================================================================
// CreateAudience creates a new audience
func (r *PostgresRepository) CreateAudience(ctx context.Context, audience *Audience) error {
filtersJSON, err := json.Marshal(audience.Filters)
if err != nil {
return fmt.Errorf("failed to marshal filters: %w", err)
}
query := `
INSERT INTO audiences (name, description, filters, created_by, is_active)
VALUES ($1, $2, $3, $4, $5)
RETURNING id, member_count, created_at, updated_at
`
return r.pool.QueryRow(ctx, query,
audience.Name,
audience.Description,
filtersJSON,
audience.CreatedBy,
audience.IsActive,
).Scan(&audience.ID, &audience.MemberCount, &audience.CreatedAt, &audience.UpdatedAt)
}
// GetAudience retrieves an audience by ID
func (r *PostgresRepository) GetAudience(ctx context.Context, id uuid.UUID) (*Audience, error) {
query := `
SELECT id, name, description, filters, member_count, last_count_update,
created_by, is_active, created_at, updated_at
FROM audiences
WHERE id = $1
`
var audience Audience
var filtersJSON []byte
err := r.pool.QueryRow(ctx, query, id).Scan(
&audience.ID, &audience.Name, &audience.Description, &filtersJSON,
&audience.MemberCount, &audience.LastCountUpdate,
&audience.CreatedBy, &audience.IsActive,
&audience.CreatedAt, &audience.UpdatedAt,
)
if err != nil {
return nil, err
}
if err := json.Unmarshal(filtersJSON, &audience.Filters); err != nil {
return nil, fmt.Errorf("failed to unmarshal filters: %w", err)
}
return &audience, nil
}
// ListAudiences lists all audiences
func (r *PostgresRepository) ListAudiences(ctx context.Context, activeOnly bool) ([]Audience, error) {
query := `
SELECT id, name, description, filters, member_count, last_count_update,
created_by, is_active, created_at, updated_at
FROM audiences
`
if activeOnly {
query += ` WHERE is_active = TRUE`
}
query += ` ORDER BY created_at DESC`
rows, err := r.pool.Query(ctx, query)
if err != nil {
return nil, fmt.Errorf("failed to query audiences: %w", err)
}
defer rows.Close()
var audiences []Audience
for rows.Next() {
var audience Audience
var filtersJSON []byte
if err := rows.Scan(
&audience.ID, &audience.Name, &audience.Description, &filtersJSON,
&audience.MemberCount, &audience.LastCountUpdate,
&audience.CreatedBy, &audience.IsActive,
&audience.CreatedAt, &audience.UpdatedAt,
); err != nil {
return nil, fmt.Errorf("failed to scan audience: %w", err)
}
if err := json.Unmarshal(filtersJSON, &audience.Filters); err != nil {
return nil, fmt.Errorf("failed to unmarshal filters: %w", err)
}
audiences = append(audiences, audience)
}
return audiences, rows.Err()
}
// UpdateAudience updates an existing audience
func (r *PostgresRepository) UpdateAudience(ctx context.Context, audience *Audience) error {
filtersJSON, err := json.Marshal(audience.Filters)
if err != nil {
return fmt.Errorf("failed to marshal filters: %w", err)
}
query := `
UPDATE audiences
SET name = $2, description = $3, filters = $4, is_active = $5, updated_at = NOW()
WHERE id = $1
RETURNING updated_at
`
return r.pool.QueryRow(ctx, query,
audience.ID,
audience.Name,
audience.Description,
filtersJSON,
audience.IsActive,
).Scan(&audience.UpdatedAt)
}
// DeleteAudience soft-deletes an audience (sets is_active = false)
func (r *PostgresRepository) DeleteAudience(ctx context.Context, id uuid.UUID) error {
query := `UPDATE audiences SET is_active = FALSE, updated_at = NOW() WHERE id = $1`
_, err := r.pool.Exec(ctx, query, id)
return err
}
// GetAudienceMembers retrieves members matching the audience filters
func (r *PostgresRepository) GetAudienceMembers(ctx context.Context, id uuid.UUID, limit, offset int) ([]AudienceMember, int, error) {
// First get the audience filters
audience, err := r.GetAudience(ctx, id)
if err != nil {
return nil, 0, fmt.Errorf("failed to get audience: %w", err)
}
// Build dynamic query based on filters
query, args := r.buildAudienceMemberQuery(audience.Filters, limit, offset, false)
countQuery, countArgs := r.buildAudienceMemberQuery(audience.Filters, 0, 0, true)
// Get total count
var totalCount int
if err := r.pool.QueryRow(ctx, countQuery, countArgs...).Scan(&totalCount); err != nil {
return nil, 0, fmt.Errorf("failed to count members: %w", err)
}
// Get members
rows, err := r.pool.Query(ctx, query, args...)
if err != nil {
return nil, 0, fmt.Errorf("failed to query members: %w", err)
}
defer rows.Close()
var members []AudienceMember
for rows.Next() {
var m AudienceMember
if err := rows.Scan(
&m.ID, &m.Name, &m.Email, &m.Position,
&m.University, &m.Department, &m.SubjectArea, &m.PublicationCount,
); err != nil {
return nil, 0, fmt.Errorf("failed to scan member: %w", err)
}
members = append(members, m)
}
return members, totalCount, rows.Err()
}
// buildAudienceMemberQuery constructs a SQL query for audience members
func (r *PostgresRepository) buildAudienceMemberQuery(filters AudienceFilters, limit, offset int, countOnly bool) (string, []interface{}) {
var args []interface{}
argNum := 1
var selectClause string
if countOnly {
selectClause = "SELECT COUNT(*)"
} else {
selectClause = `
SELECT
s.id,
COALESCE(s.title || ' ', '') || s.first_name || ' ' || s.last_name as name,
COALESCE(s.email, '') as email,
COALESCE(s.position_type, '') as position,
u.name as university,
COALESCE(d.name, '') as department,
COALESCE(sa.name, '') as subject_area,
(SELECT COUNT(*) FROM staff_publications sp WHERE sp.staff_id = s.id) as publication_count
`
}
query := selectClause + `
FROM university_staff s
JOIN universities u ON s.university_id = u.id
LEFT JOIN departments d ON s.department_id = d.id
LEFT JOIN subject_areas sa ON s.subject_area_id = sa.id
WHERE 1=1
`
// Position types filter
if len(filters.PositionTypes) > 0 {
query += fmt.Sprintf(" AND s.position_type = ANY($%d)", argNum)
args = append(args, filters.PositionTypes)
argNum++
}
// Subject areas filter
if len(filters.SubjectAreas) > 0 {
query += fmt.Sprintf(" AND s.subject_area_id = ANY($%d)", argNum)
args = append(args, filters.SubjectAreas)
argNum++
}
// States filter
if len(filters.States) > 0 {
query += fmt.Sprintf(" AND u.state = ANY($%d)", argNum)
args = append(args, filters.States)
argNum++
}
// Uni types filter
if len(filters.UniTypes) > 0 {
query += fmt.Sprintf(" AND u.uni_type = ANY($%d)", argNum)
args = append(args, filters.UniTypes)
argNum++
}
// Universities filter
if len(filters.Universities) > 0 {
query += fmt.Sprintf(" AND s.university_id = ANY($%d)", argNum)
args = append(args, filters.Universities)
argNum++
}
// Has email filter
if filters.HasEmail != nil && *filters.HasEmail {
query += " AND s.email IS NOT NULL AND s.email != ''"
}
// Is active filter
if filters.IsActive != nil && *filters.IsActive {
query += " AND s.is_active = TRUE"
}
// Keywords filter (search in name and research_areas)
if len(filters.Keywords) > 0 {
for _, keyword := range filters.Keywords {
query += fmt.Sprintf(" AND (s.first_name ILIKE $%d OR s.last_name ILIKE $%d OR s.research_areas ILIKE $%d)", argNum, argNum, argNum)
args = append(args, "%"+keyword+"%")
argNum++
}
}
if !countOnly {
query += " ORDER BY s.last_name, s.first_name"
if limit > 0 {
query += fmt.Sprintf(" LIMIT $%d", argNum)
args = append(args, limit)
argNum++
}
if offset > 0 {
query += fmt.Sprintf(" OFFSET $%d", argNum)
args = append(args, offset)
}
}
return query, args
}
// UpdateAudienceCount updates the cached member count for an audience
func (r *PostgresRepository) UpdateAudienceCount(ctx context.Context, id uuid.UUID) (int, error) {
// Get the audience filters
audience, err := r.GetAudience(ctx, id)
if err != nil {
return 0, fmt.Errorf("failed to get audience: %w", err)
}
// Count members
countQuery, countArgs := r.buildAudienceMemberQuery(audience.Filters, 0, 0, true)
var count int
if err := r.pool.QueryRow(ctx, countQuery, countArgs...).Scan(&count); err != nil {
return 0, fmt.Errorf("failed to count members: %w", err)
}
// Update the cached count
updateQuery := `
UPDATE audiences
SET member_count = $2, last_count_update = NOW(), updated_at = NOW()
WHERE id = $1
`
if _, err := r.pool.Exec(ctx, updateQuery, id, count); err != nil {
return 0, fmt.Errorf("failed to update count: %w", err)
}
return count, nil
}
// CreateExport creates a new export record
func (r *PostgresRepository) CreateExport(ctx context.Context, export *AudienceExport) error {
query := `
INSERT INTO audience_exports (audience_id, export_type, record_count, file_path, exported_by, purpose)
VALUES ($1, $2, $3, $4, $5, $6)
RETURNING id, created_at
`
return r.pool.QueryRow(ctx, query,
export.AudienceID,
export.ExportType,
export.RecordCount,
export.FilePath,
export.ExportedBy,
export.Purpose,
).Scan(&export.ID, &export.CreatedAt)
}
// ListExports lists exports for an audience
func (r *PostgresRepository) ListExports(ctx context.Context, audienceID uuid.UUID) ([]AudienceExport, error) {
query := `
SELECT id, audience_id, export_type, record_count, file_path, exported_by, purpose, created_at
FROM audience_exports
WHERE audience_id = $1
ORDER BY created_at DESC
`
rows, err := r.pool.Query(ctx, query, audienceID)
if err != nil {
return nil, fmt.Errorf("failed to query exports: %w", err)
}
defer rows.Close()
var exports []AudienceExport
for rows.Next() {
var e AudienceExport
if err := rows.Scan(
&e.ID, &e.AudienceID, &e.ExportType, &e.RecordCount,
&e.FilePath, &e.ExportedBy, &e.Purpose, &e.CreatedAt,
); err != nil {
return nil, fmt.Errorf("failed to scan export: %w", err)
}
exports = append(exports, e)
}
return exports, rows.Err()
}