Exportar a DOCX

The Word document bridges AI-generated content with human layout workflows. It must be clean, well-structured, and immediately usable by diagramadores.

Purpose

Generate branded Word documents from editorial content using python-docx, producing separate files for texto alumno and guia docente.

Procedure

1. Read brand-config.json for color tokens and typography settings. Fallback: JM Labs defaults (#6366F1 primary, #22D3EE accent, #EFEAE4 background, Inter font).

2. Collect all generated units (read from units/ directory or session state).

3. Generate texto alumno DOCX:

   • Cover page: Title, asignatura, grado, editorial branding, year
   • Table of contents (auto-generated from headings)
   • Per unit: Heading 1 (unit title), then sections as Heading 2 (Apertura, Desarrollo, Cierre, Evaluacion)
   • Activities formatted with numbered lists
   • Key vocabulary in bold
   • DZ-ref tags as small footnotes or marginalia
   • Tables for rubricas with header row styled
   • Page breaks between units

4. Generate guia docente DOCX:

   • Same cover page format but marked "Guia Docente"
   • Per unit: weekly planning table, methodological orientations, DUA adaptations (highlighted boxes), rubrics
   • Indicators linked to destrezas

5. Write DOCX files using python-docx script:

   from docx import Document
   from docx.shared import Inches, Pt, RGBColor
   from docx.enum.text import WD_ALIGN_PARAGRAPH
   

6. Save to exports/ directory.

Inputs

Name	Source	Required
Generated units	units/ directory	Yes
Generated guides	guides/ directory	For guia docente
brand-config.json	references/	Recommended

Outputs

Artifact	Format	Destination
Texto alumno	.docx	exports/
Guia docente	.docx	exports/

Acceptance Criteria

• DOCX opens correctly in Word/LibreOffice
• All units present with correct heading hierarchy
• Branding applied (colors, fonts where possible)
• Tables properly formatted (no broken cells)
• Page breaks between units
• DZ-ref tags preserved in document
• File size reasonable (<10MB per document)

Output Format

# Export DOCX — Report

## Archivos Generados
| Archivo | Tamano | Unidades | Estado |
|---------|--------|----------|--------|
| exports/{asig}_{grado}_texto_alumno.docx | N KB | N | OK |
| exports/{asig}_{grado}_guia_docente.docx | N KB | N | OK |

## Estructura del Documento
- Portada: {present/absent}
- Tabla de contenidos: {present/absent}
- Unidades: N/N incluidas
- Tablas de rubrica: N formateadas
- Saltos de pagina: {correct/issues}

## Branding
- Colores: {applied from brand-config / JM Labs defaults}
- Fuente: {Inter / fallback}

Error Handling

• python-docx not installed: Run pip install python-docx automatically
• No units generated: Error message suggesting /scriba:unit first
• Brand config missing: Use JM Labs defaults with warning
• Table content too wide: Auto-adjust column widths, warn if truncation occurs
• Special characters in content: Ensure UTF-8 encoding throughout

Anti-Patterns

• Don't embed images as base64 in DOCX — use proper image insertion via python-docx
• Don't generate a single DOCX with both texto alumno and guia docente — always separate files
• Don't hardcode styles — always read from brand-config.json first
• Don't skip the table of contents — it's critical for navigation in printed documents

References

• references/docx-generation-patterns.md — python-docx patterns and style templates