Monlam Doccano

Custom Doccano annotation platform with Tibetan language support and Monlam AI enhancements.

Features

🇹🇴 Tibetan Language Support

• Full Tibetan (བོད་སྐད་) UI translations
• Monlam Unicode font support
• RTL-compatible text handling

🎨 Monlam Branding

• Custom color scheme (gold/navy)
• Monlam logo and favicon
• Clean, distraction-free interface (GitHub buttons hidden)

🎙️ Speech-to-Text Enhancements

• JSONL import with external audio URLs - Import audio from S3/MinIO without uploading
• Auto TextLabel creation - Transcripts are automatically created as annotations
• Pre-filled labels - Annotators can review/edit instead of transcribing from scratch
• Correct audio URL export - Exports include the original audio URLs
• Auto-loop audio playback - Audio automatically loops by default with toggle controls (📖 Setup Guide)

🖼️ Image Classification Enhancements

• JSONL import with external image URLs - Import images from S3/MinIO
• Pre-filled category labels - Labels are pre-populated for review/approval workflow

✅ UI Improvements

• Review button styling: 🔴 Red Circle for "not done", 🟢 Green Check for "done"

📊 Custom Annotation Status Tracking System

• Per-annotator completion status - Track each annotator's progress on every example
• Per-approver approval status - Track each approver's reviews with notes
• Visual indicators - Color-coded status badges and progress bars in UI
• Project Manager role - New role with full visibility of completion matrix
• Admin dashboard - Comprehensive dashboard showing completion matrix for all team members
• Real-time tracking - Auto-updating status indicators
• Export functionality - Download completion data as CSV

🎨 UI Integration Features:

• Enhanced Members Progress Page - Color-coded progress bars with hover tooltips and detailed stats modal
• Dataset Table Completion Columns - Annotator and Approver status columns with badges and usernames
• Non-invasive design - JavaScript-based enhancements, no core file modifications
• Real-time updates - Automatic refresh on pagination and filtering

📖 View Completion Tracking Documentation
📖 View UI Integration Guide

Quick Start

1. Clone and Configure

git clone https://github.com/monlam-ai/monlam-doccano.git
cd monlam-doccano
cp .env.example .env
# Edit .env with your settings

2. Build and Run

docker-compose up -d --build

3. Access

Open http://localhost:8000 (or your configured port)

Default credentials (change in .env):

• Username: admin
• Password: changeme123

JSONL Import Formats

Speech-to-Text

{"filename": "https://s3.example.com/audio/file1.wav", "text": "Transcript here", "label": "Transcript here"}
{"filename": "https://s3.example.com/audio/file2.wav", "text": "Another transcript", "label": "Another transcript"}

Field	Description
filename	Audio URL (S3, MinIO, or any HTTP URL)
text	Transcript text (stored in example)
label	Pre-filled annotation (shown to annotator)
meta	Optional metadata object

Image Classification

{"filename": "https://s3.example.com/images/cat.jpg", "label": ["cat", "animal"]}
{"filename": "https://s3.example.com/images/dog.jpg", "label": ["dog", "pet"]}

Field	Description
filename	Image URL (S3, MinIO, or any HTTP URL)
label	Array of category labels (pre-filled)
meta	Optional metadata object

Project Structure

monlam-doccano/
├── Dockerfile              # Custom Doccano image
├── docker-compose.yml      # Container orchestration
├── .env.example            # Environment template
├── branding/
│   ├── i18n/               # Tibetan translations
│   │   ├── bo/             # བོད་སྐད་ locale files
│   │   └── index.js        # Locale registry
│   └── static/
│       ├── logo.png        # Monlam logo
│       └── favicon.png     # Browser favicon
└── patches/
    ├── assignment/         # Custom assignment & completion tracking
    │   ├── models_separate.py          # Assignment models
    │   ├── completion_tracking.py      # Completion tracking models
    │   ├── roles.py                    # Project Manager role
    │   ├── completion_views.py         # Completion matrix API
    │   ├── views.py                    # Assignment API
    │   ├── serializers.py              # REST serializers
    │   ├── urls.py                     # URL routing
    │   ├── migrations/
    │   │   ├── 0001_initial.py
    │   │   └── 0002_completion_tracking.py
    │   ├── COMPLETION_TRACKING_README.md
    │   ├── INSTALLATION_GUIDE.md
    │   └── QUICK_START.md
    ├── backend/
    │   ├── celery_tasks.py # Auto TextLabel creation
    │   ├── serializers.py  # External URL handling
    │   ├── export_models.py# Correct export URLs
    │   ├── catalog.py      # JSONL import options
    │   └── datasets.py     # Custom dataset classes
    ├── frontend/
    │   ├── index.html              # UI customizations
    │   ├── 200.html                # SPA fallback
    │   ├── completion-matrix.html  # Completion dashboard
    │   └── status-indicators.js    # Status UI components
    └── examples/
        ├── speech_to_text/
        │   └── example.jsonl
        └── image_classification/
            └── example.jsonl

Development

Live Frontend Updates

Uncomment the volume mounts in docker-compose.yml to enable live frontend updates:

volumes:
  - ./patches/frontend/index.html:/doccano/backend/client/dist/index.html:ro
  - ./patches/frontend/200.html:/doccano/backend/client/dist/200.html:ro

Rebuilding

After modifying backend patches:

docker-compose build --no-cache
docker-compose up -d

Integration with Traefik

For production with Traefik reverse proxy:

labels:
  - "traefik.enable=true"
  - "traefik.http.routers.doccano.rule=Host(`annotate.example.com`)"
  - "traefik.http.routers.doccano.entrypoints=websecure"
  - "traefik.http.routers.doccano.tls=true"
  - "traefik.http.services.doccano.loadbalancer.server.port=8000"

License

Based on Doccano (MIT License).

Monlam customizations © 2024 Monlam AI.

Support

For issues with Monlam customizations, please open an issue in this repository.

For core Doccano issues, please refer to the upstream repository.