75e7b9da3d
Dashboard Features: - 8 navigation sections: Overview, Outcomes, Poor CX, FCR, Churn, Agent, Call Explorer, Export - Beyond Brand Identity styling (colors #6D84E3, Outfit font) - RCA Sankey diagram (Driver → Outcome → Churn Risk flow) - Correlation heatmaps (driver co-occurrence, driver-outcome) - Outcome Deep Dive (root causes, correlation, duration analysis) - Export functionality (Excel, HTML, JSON) Blueprint Compliance: - FCR: 4 categories (Primera Llamada/Rellamada × Sin/Con Riesgo de Fuga) - Churn: Binary view (Sin Riesgo de Fuga / En Riesgo de Fuga) - Agent: Talento Para Replicar / Oportunidades de Mejora - Fixed FCR rate calculation (only FIRST_CALL counts as success) Technical: - Streamlit + Plotly for interactive visualizations - Light theme configuration (.streamlit/config.toml) - Fixed Plotly colorbar titlefont deprecation Documentation: - Updated PROJECT_CONTEXT.md, TODO.md, CHANGELOG.md - Added 4 new technical decisions (TD-014 to TD-017) - Created TROUBLESHOOTING.md with 10 common issues Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
5.3 KiB
5.3 KiB
TODO.md
Lista priorizada de tareas pendientes
Checkpoints Completados
CP1: Project Setup & Contracts ✅
- Crear estructura de carpetas
- Inicializar repo Git
- Crear requirements.txt
- Crear .env.example
- Crear README.md
- Crear config/rca_taxonomy.yaml
- Crear config/settings.yaml
- Crear schemas Pydantic
CP2: Transcription Module ✅
- Implementar Transcriber interface
- Implementar AssemblyAITranscriber
- Implementar modelos (Transcript, SpeakerTurn)
- Tests unitarios
- Notebook 01_transcription_validation.ipynb
CP3: RCA Schemas & Data Contracts ✅
- Implementar CallAnalysis
- Implementar RCALabel, EvidenceSpan
- Implementar Event
- Separar observed vs inferred
- Crear data/examples/
CP4: Feature & Event Extraction ✅
- Implementar event_detector.py
- Implementar turn_metrics.py
- Tests unitarios
CP5: Inference Engine ✅
- Crear prompt MAP único
- Implementar LLMClient con JSON strict
- Implementar BatchInference con resume
- Tests de evidence obligatorio
- Notebook 02_inference_validation.ipynb
CP6: Transcript Compression ✅
- Implementar CompressedTranscript
- Validar reducción >60% tokens
- Integrar en inference
- Notebook 03_compression_validation.ipynb
CP7: Aggregation & RCA Trees ✅
- Implementar statistics.py
- Definir severity_score con reglas explícitas
- Implementar RCATreeBuilder
- Notebook 04_aggregation_validation.ipynb
CP8: End-to-End Pipeline ✅
- Implementar CXInsightsPipeline
- Implementar manifests por stage
- Implementar resume
- Implementar exports (JSON, Excel, PDF)
- CLI principal
- Notebook 05_full_pipeline_test.ipynb
CP-GAPS: v2.0 Blueprint Alignment ✅ (2026-01-19)
- Gap Analysis vs BeyondCX Blueprints (4 docs Word)
- Update rca_taxonomy.yaml with new driver categories
- churn_risk drivers
- fcr_failure drivers
- agent_skills (positive + improvement_needed)
- Update call_analysis.py models with new fields
- FCRStatus enum
- ChurnRisk enum
- AgentClassification enum
- DriverOrigin enum
- AgentSkillIndicator model
- Enhanced RCALabel with origin, corrective_action, replicable_practice
- Updated CallAnalysis with new fields
- Create prompt v2.0 (config/prompts/call_analysis/v2.0/)
- system.txt
- user.txt
- schema.json
- Update versions.yaml to active v2.0
- Update prompt_manager.py with TaxonomyTexts
- Update analyzer.py to parse new fields
- Update aggregation models and statistics for v2.0
- Update tests for v2.0 compatibility
CP-DASH: Streamlit Dashboard ✅ (2026-01-19)
- Create dashboard structure (app.py, config.py, data_loader.py, components.py)
- Implement Beyond Brand Identity styling
- Colors: Black #000000, Blue #6D84E3, Grey #B1B1B0
- Light theme configuration (.streamlit/config.toml)
- Custom CSS with Outfit font
- Implement 8 dashboard sections
- Overview (KPIs, outcomes, drivers, FCR, churn)
- Outcomes Analysis
- Poor CX Analysis
- FCR Analysis
- Churn Risk Analysis
- Agent Performance
- Call Explorer
- Export Insights
- Advanced visualizations
- RCA Sankey Diagram (Driver → Outcome → Churn Risk)
- Correlation Heatmaps (co-occurrence, driver-outcome)
- Outcome Deep Dive (root causes, correlation, duration)
- Export functionality
- Excel multi-sheet workbook
- HTML executive summary report
- JSON raw data export
- Blueprint terminology compliance
- FCR: 4 categorías (Primera Llamada/Rellamada × Sin/Con Riesgo)
- Churn: Sin Riesgo de Fuga / En Riesgo de Fuga
- Agent: Talento Para Replicar / Oportunidades de Mejora
Alta prioridad (Pendiente)
- Run real benchmark with v2.0 - Ejecutar pipeline con 50-100 llamadas reales
- Measure actual costs - Documentar costes reales STT + LLM
- Validate v2.0 RCA accuracy - Manual review de 20 llamadas con nuevos campos
- Documentation - Completar stubs en docs/ ✅
- Test v2.0 with real transcripts - Validado con batch test-07 (30 llamadas) ✅
- Update exports for v2.0 - Dashboard incluye todos los campos nuevos ✅
- Dashboard Streamlit - Implementado con Beyond branding ✅
Media prioridad (CP9 - Optional)
- Caching por hash de transcript
- Batch size benchmarks (encontrar óptimo)
- Comparar STT providers (Whisper, Google)
- Comparar LLM providers (Claude vs GPT-4o)
- DuckDB para analytics de grandes batches
Baja prioridad (Fase 2)
- Dashboard Streamlit ✅ (completado 2026-01-19)
- Docker containerization
- CI/CD pipeline
- API REST (FastAPI)
- Multi-idioma support
- Real-time processing
- Integración BeyondDiagnosticPrototipo
- Campaign tracking (Blueprint KPI 2)
- Customer value analysis (Blueprint Pilar 4)
- Sales cycle optimization analysis
Backlog (Ideas)
- Automatic prompt tuning based on validation results
- A/B testing de prompts
- Confidence calibration
- Active learning loop
- Cost anomaly detection
Última actualización: 2026-01-19 (v2.1 Dashboard + Blueprint Compliance completed)