Email signatures clutter AI analysis and waste tokens. We need an intelligent signature stripper that detects and removes signature blocks while preserving them separately for potential future use (e.g.,contact extraction).
Acceptance Criteria
• Create backend/src/services/parser/signature-stripper.service.ts
• Detect patterns: --\n, “Best regards”, “Sent from my iPhone”, “Confidentiality notice”
• Handle multi-line signatures with names, titles, phones, social links
• Support 3 sensitivity levels: strict, moderate, lenient
• Return { cleanedBody: string, signature: string | null }
• Add signature field to Email model (Prisma schema + migration)
• Must NOT strip legitimate content containing signature-like phrases
• Deterministic: same input → same output
Technical Notes
• Pure function, no database calls
• Handle signatures in multiple languages
• Test with 50+ real email samples
Email signatures clutter AI analysis and waste tokens. We need an intelligent signature stripper that detects and removes signature blocks while preserving them separately for potential future use (e.g.,contact extraction).
Acceptance Criteria
• Create backend/src/services/parser/signature-stripper.service.ts
• Detect patterns: --\n, “Best regards”, “Sent from my iPhone”, “Confidentiality notice”
• Handle multi-line signatures with names, titles, phones, social links
• Support 3 sensitivity levels: strict, moderate, lenient
• Return { cleanedBody: string, signature: string | null }
• Add signature field to Email model (Prisma schema + migration)
• Must NOT strip legitimate content containing signature-like phrases
• Deterministic: same input → same output
Technical Notes
• Pure function, no database calls
• Handle signatures in multiple languages
• Test with 50+ real email samples