MOPC-App/src/server/services/anonymization.ts

/**
 * Data Anonymization Service
 *
 * Strips PII (names, emails, etc.) from data before sending to AI services.
 * Returns ID mappings for de-anonymization of results.
 */

export interface AnonymizedJuror {
  anonymousId: string
  expertiseTags: string[]
  currentAssignmentCount: number
  maxAssignments: number | null
}

export interface AnonymizedProject {
  anonymousId: string
  title: string
  description: string | null
  tags: string[]
  teamName: string | null
}

export interface JurorMapping {
  anonymousId: string
  realId: string
}

export interface ProjectMapping {
  anonymousId: string
  realId: string
}

export interface AnonymizationResult {
  jurors: AnonymizedJuror[]
  projects: AnonymizedProject[]
  jurorMappings: JurorMapping[]
  projectMappings: ProjectMapping[]
}

/**
 * Juror data from database
 */
interface JurorInput {
  id: string
  name?: string | null
  email: string
  expertiseTags: string[]
  maxAssignments?: number | null
  _count?: {
    assignments: number
  }
}

/**
 * Project data from database
 */
interface ProjectInput {
  id: string
  title: string
  description?: string | null
  tags: string[]
  teamName?: string | null
}

/**
 * Anonymize juror and project data for AI processing
 *
 * This function:
 * 1. Strips all PII (names, emails) from juror data
 * 2. Replaces real IDs with sequential anonymous IDs
 * 3. Keeps only expertise tags and assignment counts
 * 4. Returns mappings for de-anonymization
 */
export function anonymizeForAI(
  jurors: JurorInput[],
  projects: ProjectInput[]
): AnonymizationResult {
  const jurorMappings: JurorMapping[] = []
  const projectMappings: ProjectMapping[] = []

  // Anonymize jurors
  const anonymizedJurors: AnonymizedJuror[] = jurors.map((juror, index) => {
    const anonymousId = `juror_${(index + 1).toString().padStart(3, '0')}`

    jurorMappings.push({
      anonymousId,
      realId: juror.id,
    })

    return {
      anonymousId,
      expertiseTags: juror.expertiseTags,
      currentAssignmentCount: juror._count?.assignments ?? 0,
      maxAssignments: juror.maxAssignments ?? null,
    }
  })

  // Anonymize projects (keep content but replace IDs)
  const anonymizedProjects: AnonymizedProject[] = projects.map(
    (project, index) => {
      const anonymousId = `project_${(index + 1).toString().padStart(3, '0')}`

      projectMappings.push({
        anonymousId,
        realId: project.id,
      })

      return {
        anonymousId,
        title: sanitizeText(project.title),
        description: project.description
          ? sanitizeText(project.description)
          : null,
        tags: project.tags,
        // Replace specific team names with generic identifier
        teamName: project.teamName ? `Team ${index + 1}` : null,
      }
    }
  )

  return {
    jurors: anonymizedJurors,
    projects: anonymizedProjects,
    jurorMappings,
    projectMappings,
  }
}

/**
 * De-anonymize AI results back to real IDs
 */
export function deanonymizeResults<T extends { jurorId: string; projectId: string }>(
  results: T[],
  jurorMappings: JurorMapping[],
  projectMappings: ProjectMapping[]
): (T & { realJurorId: string; realProjectId: string })[] {
  const jurorMap = new Map(
    jurorMappings.map((m) => [m.anonymousId, m.realId])
  )
  const projectMap = new Map(
    projectMappings.map((m) => [m.anonymousId, m.realId])
  )

  return results.map((result) => ({
    ...result,
    realJurorId: jurorMap.get(result.jurorId) || result.jurorId,
    realProjectId: projectMap.get(result.projectId) || result.projectId,
  }))
}

/**
 * Sanitize text to remove potential PII patterns
 * Removes emails, phone numbers, and URLs from text
 */
function sanitizeText(text: string): string {
  // Remove email addresses
  let sanitized = text.replace(
    /[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}/g,
    '[email removed]'
  )

  // Remove phone numbers (various formats)
  sanitized = sanitized.replace(
    /(\+?\d{1,3}[-.\s]?)?\(?\d{3}\)?[-.\s]?\d{3}[-.\s]?\d{4}/g,
    '[phone removed]'
  )

  // Remove URLs
  sanitized = sanitized.replace(
    /https?:\/\/[^\s]+/g,
    '[url removed]'
  )

  return sanitized
}

/**
 * Validate that data has been properly anonymized
 * Returns true if no PII patterns are detected
 */
export function validateAnonymization(data: AnonymizationResult): boolean {
  const piiPatterns = [
    /[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}/, // Email
    /(\+?\d{1,3}[-.\s]?)?\(?\d{3}\)?[-.\s]?\d{3}[-.\s]?\d{4}/, // Phone
  ]

  const checkText = (text: string | null | undefined): boolean => {
    if (!text) return true
    return !piiPatterns.some((pattern) => pattern.test(text))
  }

  // Check jurors (they should only have expertise tags)
  for (const juror of data.jurors) {
    // Jurors should not have any text fields that could contain PII
    // Only check expertiseTags
    for (const tag of juror.expertiseTags) {
      if (!checkText(tag)) return false
    }
  }

  // Check projects
  for (const project of data.projects) {
    if (!checkText(project.title)) return false
    if (!checkText(project.description)) return false
    for (const tag of project.tags) {
      if (!checkText(tag)) return false
    }
  }

  return true
}