[M2][Epic 2.3] Add NER, topic classification, summarization, and anomaly detection

[POWDER-RANGER]

Problem

Incrementally add ML capabilities beyond sentiment analysis.

Tasks

Phase 1 - NER

• Add POST /analyze/entities (extract people, orgs, locations)
• Store entities in new entities table, link to documents
• Add entity visualization to dashboard

Phase 2 - Topics

• Add POST /analyze/topics (classify into categories)
• Store topics in topics table
• Add topic filter to dashboard

Phase 3 - Summarization

• Add POST /analyze/summarize (generate summaries)
• Display summaries in document detail view

Phase 4 - Anomaly

• Add POST /analyze/anomaly (detect outliers)
• Create anomaly alerts

Acceptance Criteria

• ✅ Each new analysis has documented API endpoint
• ✅ DB schema for storing analysis results
• ✅ Frontend visualization for each analysis type
• ✅ Regression tests prevent model degradation

Priority: P2 - Medium
Labels: ml, features, M2, P2