Features

This page provides a comprehensive overview of all features available in the Lakra annotation and evaluation system.

Core Annotation Features

Interactive Text Highlighting

Select and highlight portions of translated text to mark errors:

Click and drag to select text
Error classification on each highlight
Multiple highlights per sentence
Overlapping regions support
Color-coded by error type
Edit and delete existing highlights

Error Types:

MI_ST: Minor Syntax errors
MI_SE: Minor Semantic errors
MA_ST: Major Syntax errors
MA_SE: Major Semantic errors

Quality Scoring

Multi-dimensional quality assessment:

Fluency Rating (1-5)

Measures how natural and grammatically correct the translation reads.

Adequacy Rating (1-5)

Assesses how well the translation preserves the source meaning.

Overall Quality Rating (1-5)

Holistic judgment of translation quality.

Features:

Interactive sliders for easy rating
Real-time score validation
Historical score tracking
Comparative analytics

Error Classification and Documentation

Detailed error tracking:

Error type selection
Error descriptions for each highlight
Severity indicators
Suggested corrections
Error location tracking (character positions)

Voice Recording

Audio annotation support for detailed explanations:

Built-in voice recorder
Record, pause, resume functionality
Playback before submission
Re-record if needed
Secure storage in Supabase
Signed URLs for access control
Format: WebM or compatible audio formats

Use Cases:

Explain complex corrections
Provide pronunciation guidance
Give detailed feedback
Demonstrate proper phrasing

Comments and Suggestions

Text-based feedback:

General comments on translation
Specific error explanations
Correction suggestions
Improvement recommendations
Markdown support for formatting (where available)

Quality Assessment Features

AI-Powered Evaluation

Automated quality assessment with machine learning:

AI Quality Scoring

Automated fluency assessment
Adequacy evaluation
Overall quality prediction
Confidence levels for each score
Explanation generation for ratings

AI Error Detection

Syntax error identification
Semantic error detection
Error localization (highlighting)
Severity classification
Improvement suggestions

AI Explanations

Detailed reasoning for scores
Specific issue descriptions
Pattern recognition across similar errors
Context-aware analysis

Human-in-the-Loop Validation

Evaluators review and validate AI assessments:

Confirm AI suggestions
Reject incorrect assessments
Modify partial agreements
Add human insight not captured by AI
Flag edge cases for review

Quality Assessment Workflow

AI generates initial assessment
Evaluator reviews AI findings
Evaluator validates or corrects
Final assessment combines AI and human judgment
Feedback improves AI over time

User Management Features

Multi-Role System

Role-based access control:

Administrator Role

Full system access
User management
Content management
System configuration
Analytics and reporting

Annotator Role

Create annotations
View own work
Track personal progress
Access annotation interface

Evaluator Role

Review annotations
Quality assessments
Provide feedback
Access evaluation interface

Multi-Role Support

Users can have multiple roles
Role-specific dashboards
Granular permissions

Authentication

Flexible sign-in options:

Email authentication: Sign in with email and password
Username authentication: Sign in with username and password
Password reset: Self-service password recovery
Session management: Secure token-based authentication
Remember me: Persistent sessions (optional)

User Profiles

Comprehensive user information:

Profile details: Name, email, username
Role assignment: Current role(s)
Language preferences: Default language settings
Activity tracking: Last active, registration date
Statistics: Annotation counts, evaluation counts
Quality metrics: Average scores, agreement rates

Onboarding System

Quality assurance through testing:

Role-specific tests: Different tests for annotators and evaluators
Language-pair specific: Tests in relevant languages
Passing requirements: Minimum scores to qualify
Multiple attempts: Retake if needed
Performance tracking: Test scores and history
Certification status: Pass/fail indicators

Progress Tracking Features

Personal Dashboard

Individual progress monitoring:

Annotation count: Total completed annotations
Evaluation count: Total completed evaluations
Completion rate: Percentage of assigned work done
Average scores: Quality metrics over time
Recent activity: Latest annotations and evaluations
Streak tracking: Consecutive days active (if enabled)

System-Wide Analytics

For administrators:

User statistics: Active users, role distribution
Annotation metrics: Total annotations, completion rates
Quality trends: Quality scores over time
Error distributions: Common error types and patterns
Evaluator agreement: Inter-rater reliability metrics
Language pair coverage: Distribution across languages

Historical Data

Track progress over time:

Annotation history: All past annotations by user
Evaluation history: All past evaluations
Score trends: Quality improvements or regressions
Time tracking: Time spent on annotations
Productivity metrics: Annotations per day/week/month

Content Management Features

Sentence Management

Organize translation content:

Add sentences individually or in bulk
Edit existing sentences
Activate/deactivate sentences
Delete sentences (with confirmation)
Search and filter by language, domain, status
Annotation status: Track which sentences are done

Language Pair Support

Multi-language capabilities:

Flexible language pairs: Any source → target combination
ISO language codes: Standardized language identification
Language-specific settings: Configure by language
RTL support: Right-to-left language handling (where applicable)

Domain Classification

Categorize content by subject:

Custom domains: Create your own categories
Domain filtering: Find sentences by domain
Domain-specific guidelines: Tailored annotation rules
Statistics by domain: Track quality per subject area

Bulk Import/Export

Efficient data management:

CSV Import

Bulk sentence upload: Add many sentences at once
User import: Create multiple users via CSV
Validation: Pre-import data checking
Error reporting: Identify and fix issues before import

Data Export

Export annotations: Download annotation data
Export evaluations: Download evaluation data
Export reports: Generate analytical reports
Multiple formats: CSV, JSON, Excel

Real-Time Features

Live Updates

Dynamic interface updates:

Real-time validation: Immediate feedback on input
Auto-save drafts: Prevent data loss (where enabled)
Session persistence: Resume work after disconnect
Optimistic updates: Responsive UI interactions

Notifications

Stay informed:

In-app notifications: Alerts within the interface
Email notifications: Important updates via email (if configured)
Feedback notifications: When evaluations are complete
System announcements: Administrator messages

Interface Features

Responsive Design

Works on all devices:

Desktop optimized: Full-featured experience on large screens
Tablet support: Touch-friendly interface
Mobile compatible: Core features on smartphones
Adaptive layout: Adjusts to screen size

Accessibility

Inclusive design:

Keyboard navigation: Full keyboard support
Screen reader compatibility: ARIA labels and semantics
High contrast: Support for visibility needs
Focus indicators: Clear keyboard focus
Skip links: Navigate efficiently

User Experience

Intuitive interface design:

Clean layout: Uncluttered, focused design
Contextual help: Info buttons and tooltips
Clear labeling: Descriptive form labels
Error messages: Helpful, specific feedback
Success confirmation: Clear completion indicators
Loading states: Progress indicators for long operations

Customization

Personalize your experience:

Language preferences: Set UI language
Theme support: Light/dark mode (if available)
Display preferences: Adjust text size, density
Notification settings: Control alert preferences

Security Features

Data Protection

Keep information secure:

Row Level Security: Database-level access control
Encrypted connections: HTTPS/TLS for all communication
Secure authentication: JWT token-based auth
Password hashing: bcrypt or similar for passwords
Session timeout: Automatic logout after inactivity

Privacy

Respect user privacy:

Data minimization: Collect only needed data
Access controls: Role-based data access
Audit logs: Track data access and changes
Data deletion: Complete removal on request
Export capability: Users can export their data

Integration Features

Supabase Integration

Comprehensive backend integration:

Authentication: Supabase Auth for user management
Database: PostgreSQL with real-time capabilities
Storage: Voice recordings in Supabase Storage
Edge Functions: Serverless for custom logic
Real-time subscriptions: Live data updates

API Access

Programmatic access:

REST API: Full CRUD operations via Supabase client
Authentication API: User sign in/up/out
Annotations API: Create and retrieve annotations
Evaluations API: Manage evaluations
Admin API: System management functions

Third-Party Integration

Extensibility:

Export formats: Standard formats for external tools
Webhooks: Event notifications (if configured)
API keys: Secure access for integrations
Custom extensions: Ability to add new features

Performance Features

Optimization

Fast and efficient:

Lazy loading: Load content as needed
Caching: Reduce redundant data fetching
Debouncing: Optimize search and input
Pagination: Handle large datasets
Indexing: Fast database queries

Scalability

Handle growth:

Horizontal scaling: Add more servers as needed
Database optimization: Efficient queries and indexes
CDN support: Fast static asset delivery
Load balancing: Distribute traffic

Reporting Features

Built-in Reports

Pre-configured analytics:

User activity: Login frequency, engagement metrics
Annotation quality: Quality scores distribution
Error analysis: Common error patterns
Progress reports: Completion tracking
Evaluator performance: Agreement rates, consistency

Custom Reports

Generate specific analyses:

Date range filtering: Analyze specific periods
User filtering: Focus on specific users or roles
Language pair filtering: Analyze specific language combinations
Export options: Download reports in various formats

Future Features

Features planned or in development:

Advanced AI models: Improved quality assessment
Collaborative annotation: Multiple annotators on same sentence
Annotation versioning: Track changes over time
Advanced analytics: Machine learning insights
Mobile apps: Native iOS/Android applications
API documentation: Interactive API documentation interface

Feature Requests

Have an idea for a new feature?

Contact your administrator
Submit via GitHub (if open source)
Participate in user surveys
Join user feedback sessions