# Configuration

Edison provides several ways to configure its behavior to suit your needs. This guide explains the available configuration options and how to set them.

## Configuration File

The primary way to configure Edison is through the `edison.yaml` file. This file is located in the Edison installation directory.

```mermaid
flowchart LR
    A[edison.yaml] --> B[OpenAI Settings]
    A --> C[Safety Settings]
    A --> D[Shell Settings]
    
    B --> B1[Model]
    B --> B2[Temperature]
    B --> B3[Max Tokens]
    B --> B4[API Key]
    
    C --> C1[Safety Mode]
    
    D --> D1[Shell Type]
    
    style A fill:#f9d5e5,stroke:#333,stroke-width:2px
    style B fill:#eeeeee,stroke:#333,stroke-width:2px
    style C fill:#d3f6db,stroke:#333,stroke-width:2px
    style D fill:#d3f6f5,stroke:#333,stroke-width:2px
```

### Sample Configuration File

Here's an example `edison.yaml` file with all available options:

```yaml
# OpenAI API settings
model: gpt-4o-mini
temperature: 0
max_tokens: 500
openai_api_key: your_api_key_here  # Optional, can be set via environment variable
streaming: true  # Enable streaming command generation

# Safety settings
safety: true

# UI settings
show_generating_prefix: true  # Show "Generating command:" prefix during generation

# Shell settings
shell: bash  # Options: bash, zsh, powershell.exe, cmd.exe, etc.
```

## Configuration Options

### OpenAI API Settings

| Option | Description | Default | Possible Values |
|--------|-------------|---------|----------------|
| model | OpenAI model to use | gpt-4o-mini | gpt-4o-mini, gpt-3.5-turbo, gpt-4, etc. |
| temperature | Randomness of completions | 0 | 0.0 - 1.0 |
| max_tokens | Maximum tokens in response | 500 | 1 - 4096 |
| openai_api_key | Your OpenAI API key | - | Valid API key |
| streaming | Enable streaming command generation | true | true, false |

### Safety Settings

| Option | Description | Default | Possible Values |
|--------|-------------|---------|----------------|
| safety | Require confirmation before execution | true | true, false |

### UI Settings

| Option | Description | Default | Possible Values |
|--------|-------------|---------|----------------|
| show_generating_prefix | Show "Generating command:" prefix during generation | true | true, false |

### Shell Settings

| Option | Description | Default | Possible Values |
|--------|-------------|---------|----------------|
| shell | Shell to use for command execution | (system default) | bash, zsh, powershell.exe, cmd.exe, etc. |

## Setting Your OpenAI API Key

There are several ways to set your OpenAI API key, listed in order of precedence:

1. **Environment variable**:
   - Linux/macOS: `export OPENAI_API_KEY=<yourkey>`
   - Windows: `$env:OPENAI_API_KEY="<yourkey>"`

2. **API key file**:
   Create a file at `~/.openai.apikey` containing only your API key.

3. **Configuration file**:
   Add your key to the `edison.yaml` file:
   ```yaml
   openai_api_key: your_api_key_here
   ```

## Viewing Current Configuration

To view your current configuration:

```bash
edison -c
```

This will display all current settings, including where your OpenAI API key is being loaded from (but will not display the key itself for security reasons).

## Configuration Load Order

Edison loads configuration in this order:

1. Default values
2. Values from `edison.yaml`
3. Environment variables (for API key)
4. Command-line arguments

Later sources override earlier ones.

## Recommended Configurations

### For Beginners

```yaml
model: gpt-4o-mini
temperature: 0
max_tokens: 500
safety: true
```

### For Advanced Users

```yaml
model: gpt-4  # If you have access
temperature: 0.2  # Slightly more creative
max_tokens: 1000  # For more detailed responses
safety: true  # Keep safety on unless you really know what you're doing
```

### For Performance

```yaml
model: gpt-3.5-turbo  # Fastest option if you need maximum speed
temperature: 0
max_tokens: 300  # Lower limits can be faster
```

## Troubleshooting Configuration Issues

- **Configuration not applying**: Ensure your `edison.yaml` file is in the correct location
- **API key not recognized**: Check for extra whitespace or quotes in your key
- **Model not available**: Verify you have access to the specified model in your OpenAI account