meta-agent-mobile

内容来源：README.md（说明文档） · 原始地址 · 查看安装指南

原始内容

# Meta Agent Mobile

A comprehensive mobile application for conducting meta-analyses with AI assistance. Built with React Native, Expo, and powered by Glass AI.

![License](https://img.shields.io/badge/license-MIT-blue.svg)
![Platform](https://img.shields.io/badge/platform-iOS%20%7C%20Android%20%7C%20Web-lightgrey.svg)
![Tests](https://img.shields.io/badge/tests-901%20passing-brightgreen.svg)

## Features

### Core Meta-Analysis Capabilities

- **Data Entry & Validation** - Spreadsheet-based data entry with real-time validation for binary, continuous, and pre-calculated effect sizes
- **Automatic Data Type Detection** - MVP Orchestrator detects your data type and suggests appropriate analyses
- **R Code Generation** - Generates publication-ready R code using metafor, mada, and netmeta packages
- **Forest Plots** - Interactive forest plot visualization with customizable options
- **Funnel Plots** - Publication bias assessment with Egger's test and trim-and-fill analysis

### Glass AI Assistant

- **13+ AgentSkills** - Specialized skills for TSA, Bayesian meta-analysis, network meta-analysis, diagnostic accuracy, ROB assessment, GRADE, and more
- **Natural Language Interface** - Ask questions and get guidance in plain language
- **Slash Commands** - Quick access via `/analyze`, `/detect`, `/suggest`, `/generate-code`, `/explain`
- **Multilingual Support** - English, Portuguese, and Spanish

### Research Tools

- **PubMed/CrossRef Integration** - Search and import study metadata directly from literature databases
- **PlotDigitizer** - Extract data points from published figures with axis calibration
- **PROSPERO Integration** - Import systematic review protocols
- **Citation Formatting** - Automatic APA citation generation

### Collaboration & Sync

- **Cloud Backup** - Automatic backup of spreadsheets and progress
- **Real-time Collaboration** - Share projects with collaborators
- **Offline Support** - Works without internet connection

## Architecture

```
meta-agent-mobile/
├── app/                          # Expo Router screens
│   ├── (tabs)/                   # Tab-based navigation
│   │   ├── index.tsx             # Home screen with Glass chat
│   │   ├── my-progress.tsx       # Learning progress tracking
│   │   └── settings.tsx          # App settings
│   ├── onboarding.tsx            # First-launch onboarding
│   └── tutorial.tsx              # Interactive tutorials
├── components/
│   ├── glass/                    # Glass AI mascot and animations
│   ├── spreadsheet/              # Data entry components
│   ├── visualization/            # Charts and plots
│   ├── digitizer/                # PlotDigitizer tool
│   └── terminal/                 # Chat interface
├── lib/
│   ├── agent/                    # Glass AI agent system
│   │   ├── skills/               # AgentSkills definitions
│   │   └── commands.ts           # Slash command handlers
│   ├── glass/
│   │   └── orchestrator/         # MVP Orchestrator
│   │       ├── data-type-detector.ts
│   │       ├── analysis-suggester.ts
│   │       ├── r-code-generator.ts
│   │       └── use-orchestrator.ts
│   ├── spreadsheet/              # Spreadsheet logic
│   └── literature-search/        # PubMed/CrossRef APIs
├── hooks/                        # React hooks
├── server/                       # Backend API (tRPC)
├── drizzle/                      # Database schema
├── agentskills/                  # SKILL.md files for Glass
└── __tests__/                    # Test suites
    ├── e2e/                      # End-to-end tests
    ├── orchestrator/             # Orchestrator tests
    └── ...
```

## Getting Started

### Prerequisites

- Node.js 18+
- pnpm 9+
- Expo CLI
- iOS Simulator or Android Emulator (optional)

### Installation

```bash
# Clone the repository
git clone https://github.com/yourusername/meta-agent-mobile.git
cd meta-agent-mobile

# Install dependencies
pnpm install

# Start the development server
pnpm dev
```

### Running on Different Platforms

```bash
# Web
pnpm dev:metro

# iOS Simulator
pnpm ios

# Android Emulator
pnpm android

# Generate QR code for Expo Go
pnpm qr
```

## Testing

The project includes comprehensive test coverage with 901 tests across unit, integration, and E2E test suites.

```bash
# Run all tests
pnpm test

# Run specific test suite
pnpm test -- __tests__/e2e/

# Run tests in watch mode
pnpm test -- --watch

# Run tests with coverage
pnpm test -- --coverage
```

### Test Structure

| Suite | Tests | Description |
|-------|-------|-------------|
| E2E | 90 | Full workflow tests |
| Orchestrator | 75 | Data detection & analysis |
| Glass Integration | 49 | AI assistant tests |
| Commands | 18 | Slash command tests |
| R Integration | 50+ | R code generation |
| Spreadsheet | 100+ | Data entry & validation |

## MVP Orchestrator

The Orchestrator automatically detects data types and suggests appropriate analyses:

### Supported Data Types

| Type | Detection Pattern | Suggested Analysis |
|------|-------------------|-------------------|
| Binary | events_*, n_* columns | OR, RR, RD with metafor |
| Continuous | mean_*, sd_*, n_* columns | SMD, MD with metafor |
| Pre-calculated | yi, vi or effect_size, se columns | Generic with metafor |
| Diagnostic | tp, fp, fn, tn columns | Bivariate with mada |
| Correlation | r, n columns | Fisher's z transformation |
| Hazard Ratio | hr, ci_lower, ci_upper columns | Log HR with metafor |

### Slash Commands

| Command | Description |
|---------|-------------|
| `/analyze` | Run full analysis workflow |
| `/detect` | Detect data type from spreadsheet |
| `/suggest` | Get analysis recommendations |
| `/generate-code` | Generate R code |
| `/explain [topic]` | Explain effect measures or concepts |
| `/help` | Show all available commands |

## AgentSkills

Glass AI has access to 13+ specialized skills:

- `trial-sequential-analysis` - TSA with monitoring boundaries
- `bayesian-meta-analysis` - Bayesian random effects models
- `network-meta-analysis` - NMA with netmeta package
- `diagnostic-meta-analysis` - Bivariate models for DTA
- `risk-of-bias` - RoB 2, ROBINS-I, NOS, QUADAS-2
- `grade-assessment` - GRADE certainty of evidence
- `ipd-meta-analysis` - Individual participant data
- `publication-bias` - Funnel plots, Egger's test, trim-and-fill
- `heterogeneity` - I², τ², prediction intervals
- `sensitivity-analysis` - Leave-one-out, influence diagnostics
- `subgroup-analysis` - Moderator analysis
- `meta-regression` - Continuous moderators
- `prisma-flowchart` - PRISMA 2020 flow diagrams

## Database Schema

The app uses PostgreSQL with Drizzle ORM:

```typescript
// Key tables
- users              // User accounts
- spreadsheets       // Meta-analysis data
- collaborators      // Shared access
- spreadsheet_history // Version control
- user_progress      // Learning tracking
- rob_assessments    // Risk of bias data
```

## API Routes

Backend API built with tRPC:

```typescript
// Key routes
- auth.*             // Authentication
- spreadsheet.*      // CRUD operations
- sync.*             // Cloud sync
- collaboration.*    // Real-time editing
- progress.*         // Learning progress
```

## Contributing

1. Fork the repository
2. Create a feature branch (`git checkout -b feature/amazing-feature`)
3. Commit your changes (`git commit -m 'Add amazing feature'`)
4. Push to the branch (`git push origin feature/amazing-feature`)
5. Open a Pull Request

### Development Guidelines

- Write tests for new features
- Follow the existing code style
- Update documentation as needed
- Ensure all tests pass before submitting PR

## License

This project is licensed under the MIT License - see the [LICENSE](https://github.com/matheus-rech/meta-agent-mobile/tree/HEAD/LICENSE) file for details.

## Acknowledgments

- [Expo](https://expo.dev/) - React Native framework
- [metafor](https://www.metafor-project.org/) - R package for meta-analysis
- [mada](https://cran.r-project.org/package=mada) - R package for diagnostic accuracy
- [netmeta](https://cran.r-project.org/package=netmeta) - R package for network meta-analysis
- [Cochrane Handbook](https://training.cochrane.org/handbook) - Meta-analysis methodology

## Support

For questions or issues, please open a GitHub issue or contact the maintainers.

Meta Agent Mobile

A comprehensive mobile application for conducting meta-analyses with AI assistance. Built with React Native, Expo, and powered by Glass AI.

Features

Core Meta-Analysis Capabilities

Data Entry & Validation - Spreadsheet-based data entry with real-time validation for binary, continuous, and pre-calculated effect sizes
Automatic Data Type Detection - MVP Orchestrator detects your data type and suggests appropriate analyses
R Code Generation - Generates publication-ready R code using metafor, mada, and netmeta packages
Forest Plots - Interactive forest plot visualization with customizable options
Funnel Plots - Publication bias assessment with Egger's test and trim-and-fill analysis

Glass AI Assistant

13+ AgentSkills - Specialized skills for TSA, Bayesian meta-analysis, network meta-analysis, diagnostic accuracy, ROB assessment, GRADE, and more
Natural Language Interface - Ask questions and get guidance in plain language
Slash Commands - Quick access via /analyze, /detect, /suggest, /generate-code, /explain
Multilingual Support - English, Portuguese, and Spanish

Research Tools

PubMed/CrossRef Integration - Search and import study metadata directly from literature databases
PlotDigitizer - Extract data points from published figures with axis calibration
PROSPERO Integration - Import systematic review protocols
Citation Formatting - Automatic APA citation generation

Collaboration & Sync

Cloud Backup - Automatic backup of spreadsheets and progress
Real-time Collaboration - Share projects with collaborators
Offline Support - Works without internet connection

Architecture

meta-agent-mobile/
├── app/                          # Expo Router screens
│   ├── (tabs)/                   # Tab-based navigation
│   │   ├── index.tsx             # Home screen with Glass chat
│   │   ├── my-progress.tsx       # Learning progress tracking
│   │   └── settings.tsx          # App settings
│   ├── onboarding.tsx            # First-launch onboarding
│   └── tutorial.tsx              # Interactive tutorials
├── components/
│   ├── glass/                    # Glass AI mascot and animations
│   ├── spreadsheet/              # Data entry components
│   ├── visualization/            # Charts and plots
│   ├── digitizer/                # PlotDigitizer tool
│   └── terminal/                 # Chat interface
├── lib/
│   ├── agent/                    # Glass AI agent system
│   │   ├── skills/               # AgentSkills definitions
│   │   └── commands.ts           # Slash command handlers
│   ├── glass/
│   │   └── orchestrator/         # MVP Orchestrator
│   │       ├── data-type-detector.ts
│   │       ├── analysis-suggester.ts
│   │       ├── r-code-generator.ts
│   │       └── use-orchestrator.ts
│   ├── spreadsheet/              # Spreadsheet logic
│   └── literature-search/        # PubMed/CrossRef APIs
├── hooks/                        # React hooks
├── server/                       # Backend API (tRPC)
├── drizzle/                      # Database schema
├── agentskills/                  # SKILL.md files for Glass
└── __tests__/                    # Test suites
    ├── e2e/                      # End-to-end tests
    ├── orchestrator/             # Orchestrator tests
    └── ...

Getting Started

Prerequisites

Node.js 18+
pnpm 9+
Expo CLI
iOS Simulator or Android Emulator (optional)

Installation

# Clone the repository
git clone https://github.com/yourusername/meta-agent-mobile.git
cd meta-agent-mobile

# Install dependencies
pnpm install

# Start the development server
pnpm dev

Running on Different Platforms

# Web
pnpm dev:metro

# iOS Simulator
pnpm ios

# Android Emulator
pnpm android

# Generate QR code for Expo Go
pnpm qr

Testing

The project includes comprehensive test coverage with 901 tests across unit, integration, and E2E test suites.

# Run all tests
pnpm test

# Run specific test suite
pnpm test -- __tests__/e2e/

# Run tests in watch mode
pnpm test -- --watch

# Run tests with coverage
pnpm test -- --coverage

Test Structure

Suite	Tests	Description
E2E	90	Full workflow tests
Orchestrator	75	Data detection & analysis
Glass Integration	49	AI assistant tests
Commands	18	Slash command tests
R Integration	50+	R code generation
Spreadsheet	100+	Data entry & validation

MVP Orchestrator

The Orchestrator automatically detects data types and suggests appropriate analyses:

Supported Data Types

Type	Detection Pattern	Suggested Analysis
Binary	events_, n_ columns	OR, RR, RD with metafor
Continuous	mean_, sd_, n_* columns	SMD, MD with metafor
Pre-calculated	yi, vi or effect_size, se columns	Generic with metafor
Diagnostic	tp, fp, fn, tn columns	Bivariate with mada
Correlation	r, n columns	Fisher's z transformation
Hazard Ratio	hr, ci_lower, ci_upper columns	Log HR with metafor

Slash Commands

Command	Description
`/analyze`	Run full analysis workflow
`/detect`	Detect data type from spreadsheet
`/suggest`	Get analysis recommendations
`/generate-code`	Generate R code
`/explain [topic]`	Explain effect measures or concepts
`/help`	Show all available commands

AgentSkills

Glass AI has access to 13+ specialized skills:

trial-sequential-analysis - TSA with monitoring boundaries
bayesian-meta-analysis - Bayesian random effects models
network-meta-analysis - NMA with netmeta package
diagnostic-meta-analysis - Bivariate models for DTA
risk-of-bias - RoB 2, ROBINS-I, NOS, QUADAS-2
grade-assessment - GRADE certainty of evidence
ipd-meta-analysis - Individual participant data
publication-bias - Funnel plots, Egger's test, trim-and-fill
heterogeneity - I², τ², prediction intervals
sensitivity-analysis - Leave-one-out, influence diagnostics
subgroup-analysis - Moderator analysis
meta-regression - Continuous moderators
prisma-flowchart - PRISMA 2020 flow diagrams

Database Schema

The app uses PostgreSQL with Drizzle ORM:

// Key tables
- users              // User accounts
- spreadsheets       // Meta-analysis data
- collaborators      // Shared access
- spreadsheet_history // Version control
- user_progress      // Learning tracking
- rob_assessments    // Risk of bias data

API Routes

Backend API built with tRPC:

// Key routes
- auth.*             // Authentication
- spreadsheet.*      // CRUD operations
- sync.*             // Cloud sync
- collaboration.*    // Real-time editing
- progress.*         // Learning progress

Contributing

Fork the repository
Create a feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

Development Guidelines

Write tests for new features
Follow the existing code style
Update documentation as needed
Ensure all tests pass before submitting PR

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

Expo - React Native framework
metafor - R package for meta-analysis
mada - R package for diagnostic accuracy
netmeta - R package for network meta-analysis
Cochrane Handbook - Meta-analysis methodology

Support

For questions or issues, please open a GitHub issue or contact the maintainers.