voice-mode 4.0.1__tar.gz → 4.1.0__tar.gz

{voice_mode-4.0.1 → voice_mode-4.1.0}/CHANGELOG.md

@@ -7,6 +7,20 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [4.1.0] - 2025-09-01
+
+### Added
+- **Pronunciation middleware for TTS/STT text processing**
+  - Configurable pronunciation rules system that processes text before TTS and after STT
+  - Regex-based text substitution rules with YAML configuration
+  - Separate TTS and STT rule sets for bidirectional corrections
+  - Privacy support - rules can be marked private to hide from LLM tool listings
+  - Default rules for common patterns (3M, PoE, GbE, etc.)
+  - Full CLI interface for managing pronunciation rules
+  - MCP tool for LLM-based rule management with `pronounce` tool
+  - Integrated into converse tool for automatic text processing
+  - New configuration file: `voice_mode/data/default_pronunciation.yaml`
+
 ## [4.0.1] - 2025-09-01
 
 ### Removed

{voice_mode-4.0.1 → voice_mode-4.1.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: voice-mode
-Version: 4.0.1
+Version: 4.1.0
 Summary: VoiceMode - Voice interaction capabilities for AI assistants (formerly voice-mcp)
 Project-URL: Homepage, https://github.com/mbailey/voicemode
 Project-URL: Repository, https://github.com/mbailey/voicemode

{voice_mode-4.0.1 → voice_mode-4.1.0}/voice_mode/__version__.py

@@ -1,3 +1,3 @@
 # This file is automatically updated by 'make release'
 # Do not edit manually
-__version__ = "4.0.1"
+__version__ = "4.1.0"

{voice_mode-4.0.1 → voice_mode-4.1.0}/voice_mode/cli.py

@@ -1360,13 +1360,16 @@ def cli():
 # Import subcommand groups
 from voice_mode.cli_commands import exchanges as exchanges_cmd
 from voice_mode.cli_commands import transcribe as transcribe_cmd
+from voice_mode.cli_commands import pronounce_commands
 
 # Add subcommands to legacy CLI
 cli.add_command(exchanges_cmd.exchanges)
 cli.add_command(transcribe_cmd.transcribe)
+cli.add_command(pronounce_commands.pronounce_group)
 
 # Add exchanges to main CLI
 voice_mode_main_cli.add_command(exchanges_cmd.exchanges)
+voice_mode_main_cli.add_command(pronounce_commands.pronounce_group)
 
 # Add transcribe to main CLI
 voice_mode_main_cli.add_command(transcribe_cmd.transcribe)

voice_mode-4.1.0/voice_mode/cli_commands/pronounce_commands.py

@@ -0,0 +1,223 @@
+"""CLI commands for managing pronunciation rules."""
+
+import click
+import yaml
+import json
+from pathlib import Path
+from typing import Optional
+
+from voice_mode.pronounce import get_manager
+
+
+@click.group(name='pronounce')
+def pronounce_group():
+    """Manage pronunciation rules for TTS and STT."""
+    pass
+
+
+@pronounce_group.command(name='list')
+@click.option('--direction', '-d', type=click.Choice(['tts', 'stt', 'all']), default='all',
+              help='Filter by direction (tts/stt/all)')
+@click.option('--enabled-only', '-e', is_flag=True, help='Show only enabled rules')
+@click.option('--show-private', '-p', is_flag=True, help='Include private rules')
+@click.option('--format', '-f', type=click.Choice(['table', 'yaml', 'json']), default='table',
+              help='Output format')
+def list_rules(direction: str, enabled_only: bool, show_private: bool, format: str):
+    """List pronunciation rules."""
+    manager = get_manager()
+    
+    # Get rules
+    if direction == 'all':
+        rules = manager.list_rules(include_private=show_private)
+    else:
+        rules = manager.list_rules(direction=direction, include_private=show_private)
+    
+    # Filter if needed
+    if enabled_only:
+        rules = [r for r in rules if r['enabled']]
+    
+    # Format output
+    if format == 'table':
+        if not rules:
+            click.echo("No rules found.")
+            return
+        
+        # Count private rules that were hidden
+        all_rules = manager.list_rules(include_private=True)
+        private_count = len(all_rules) - len(rules)
+        
+        # Simple table format without tabulate
+        click.echo("\nPronunciation Rules:")
+        click.echo("=" * 80)
+        
+        for rule in rules:
+            status = '✓' if rule['enabled'] else '✗'
+            click.echo(f"\n{status} [{rule['direction'].upper()}] {rule['name']} (order: {rule['order']})")
+            click.echo(f"  Pattern:  {rule['pattern']}")
+            click.echo(f"  Replace:  {rule['replacement']}")
+            if rule['description']:
+                click.echo(f"  Desc:     {rule['description']}")
+        
+        if private_count > 0 and not show_private:
+            click.echo(f"\n({private_count} private rules hidden. Use --show-private to display)")
+    
+    elif format == 'yaml':
+        import yaml
+        click.echo(yaml.dump(rules, default_flow_style=False))
+    
+    elif format == 'json':
+        import json
+        click.echo(json.dumps(rules, indent=2))
+
+
+@pronounce_group.command(name='test')
+@click.argument('text')
+@click.option('--direction', '-d', type=click.Choice(['tts', 'stt']), default='tts',
+              help='Test direction (tts/stt)')
+def test_rule(text: str, direction: str):
+    """Test pronunciation rules on text."""
+    manager = get_manager()
+    result = manager.test_rule(text, direction)
+    
+    if text != result:
+        click.echo(f"Original: {text}")
+        click.echo(f"Modified: {result}")
+        
+        # Show which rules were applied if logging is enabled
+        import os
+        if os.environ.get('VOICEMODE_PRONUNCIATION_LOG_SUBSTITUTIONS', '').lower() == 'true':
+            click.echo("\n(Check logs for applied rules)")
+    else:
+        click.echo(f"No changes: {text}")
+
+
+@pronounce_group.command(name='add')
+@click.option('--direction', '-d', type=click.Choice(['tts', 'stt']), required=True,
+              help='Rule direction (tts/stt)')
+@click.option('--pattern', '-p', required=True, help='Regex pattern to match')
+@click.option('--replacement', '-r', required=True, help='Replacement text')
+@click.option('--name', '-n', help='Rule name (auto-generated if not provided)')
+@click.option('--description', help='Rule description')
+@click.option('--order', type=int, default=100, help='Processing order (lower = earlier)')
+@click.option('--disabled', is_flag=True, help='Create rule as disabled')
+def add_rule(direction: str, pattern: str, replacement: str, name: Optional[str],
+             description: str, order: int, disabled: bool):
+    """Add a new pronunciation rule."""
+    manager = get_manager()
+    
+    success = manager.add_rule(
+        direction=direction,
+        pattern=pattern,
+        replacement=replacement,
+        name=name,
+        description=description or "",
+        enabled=not disabled,
+        order=order,
+        private=False  # CLI-created rules are not private
+    )
+    
+    if success:
+        click.echo(f"✓ Rule added successfully")
+    else:
+        click.echo("✗ Failed to add rule (check pattern validity)", err=True)
+
+
+@pronounce_group.command(name='remove')
+@click.option('--direction', '-d', type=click.Choice(['tts', 'stt']), required=True,
+              help='Rule direction (tts/stt)')
+@click.argument('name')
+def remove_rule(direction: str, name: str):
+    """Remove a pronunciation rule by name."""
+    manager = get_manager()
+    
+    success = manager.remove_rule(direction, name)
+    
+    if success:
+        click.echo(f"✓ Rule '{name}' removed")
+    else:
+        click.echo(f"✗ Rule '{name}' not found", err=True)
+
+
+@pronounce_group.command(name='enable')
+@click.option('--direction', '-d', type=click.Choice(['tts', 'stt']), required=True,
+              help='Rule direction (tts/stt)')
+@click.argument('name')
+def enable_rule(direction: str, name: str):
+    """Enable a pronunciation rule."""
+    manager = get_manager()
+    
+    success = manager.enable_rule(direction, name)
+    
+    if success:
+        click.echo(f"✓ Rule '{name}' enabled")
+    else:
+        click.echo(f"✗ Failed to enable rule '{name}' (not found or private)", err=True)
+
+
+@pronounce_group.command(name='disable')
+@click.option('--direction', '-d', type=click.Choice(['tts', 'stt']), required=True,
+              help='Rule direction (tts/stt)')
+@click.argument('name')
+def disable_rule(direction: str, name: str):
+    """Disable a pronunciation rule."""
+    manager = get_manager()
+    
+    success = manager.disable_rule(direction, name)
+    
+    if success:
+        click.echo(f"✓ Rule '{name}' disabled")
+    else:
+        click.echo(f"✗ Failed to disable rule '{name}' (not found or private)", err=True)
+
+
+@pronounce_group.command(name='reload')
+def reload_rules():
+    """Reload pronunciation rules from configuration files."""
+    manager = get_manager()
+    manager.reload_rules()
+    click.echo("✓ Pronunciation rules reloaded")
+
+
+@pronounce_group.command(name='edit')
+@click.option('--system', is_flag=True, help='Edit system default rules (requires sudo)')
+def edit_config(system: bool):
+    """Open pronunciation config in editor."""
+    import os
+    import subprocess
+    
+    if system:
+        # Edit system defaults
+        config_path = Path(__file__).parent.parent / 'data' / 'default_pronunciation.yaml'
+        if not config_path.exists():
+            click.echo(f"System config not found: {config_path}", err=True)
+            return
+        # Might need sudo
+        editor = os.environ.get('EDITOR', 'nano')
+        subprocess.run(['sudo', editor, str(config_path)])
+    else:
+        # Edit user config
+        config_path = Path.home() / '.voicemode' / 'config' / 'pronunciation.yaml'
+        if not config_path.exists():
+            # Create default config
+            config_path.parent.mkdir(parents=True, exist_ok=True)
+            default_config = {
+                'version': 1,
+                'tts_rules': [],
+                'stt_rules': []
+            }
+            with open(config_path, 'w') as f:
+                yaml.dump(default_config, f, default_flow_style=False)
+        
+        editor = os.environ.get('EDITOR', 'nano')
+        subprocess.run([editor, str(config_path)])
+    
+    # Reload after editing
+    manager = get_manager()
+    manager.reload_rules()
+    click.echo("✓ Configuration edited and reloaded")
+
+
+# Register the command group
+def register_commands(cli):
+    """Register pronunciation commands with the main CLI."""
+    cli.add_command(pronounce_group)

voice_mode-4.1.0/voice_mode/data/default_pronunciation.yaml

@@ -0,0 +1,268 @@
+version: 1
+tts_rules:
+  # Company names
+  - name: "3m_company"
+    order: 10
+    pattern: '\b3M\b'
+    replacement: 'three em'
+    enabled: true
+    description: "Pronounce 3M company name correctly"
+    private: false
+  
+  # Networking terms
+  - name: "poe_acronym"
+    order: 20
+    pattern: '\bPoE\+?\b'
+    replacement: 'P O E'
+    enabled: true
+    description: "Power over Ethernet acronym"
+    private: false
+  
+  - name: "gigabit_ethernet"
+    order: 30
+    pattern: '\b(\d+(?:\.\d+)?)\s*GbE\b'
+    replacement: '$1 gigabit ethernet'
+    enabled: true
+    description: "Expand GbE to gigabit ethernet with number"
+    private: false
+  
+  - name: "tcp_ip"
+    order: 40
+    pattern: '\bTCP/IP\b'
+    replacement: 'T C P I P'
+    enabled: true
+    description: "Pronounce TCP/IP as individual letters"
+    private: false
+  
+  - name: "ssh"
+    order: 50
+    pattern: '\bSSH\b'
+    replacement: 'S S H'
+    enabled: true
+    description: "Pronounce SSH as individual letters"
+    private: false
+  
+  - name: "https"
+    order: 60
+    pattern: '\bHTTPS\b'
+    replacement: 'H T T P S'
+    enabled: true
+    description: "Pronounce HTTPS as individual letters"
+    private: false
+  
+  # Model numbers and versions
+  - name: "unifi_u7"
+    order: 70
+    pattern: '\bU7\b'
+    replacement: 'U seven'
+    enabled: true
+    description: "UniFi U7 model number"
+    private: false
+  
+  - name: "wifi_6e"
+    order: 80
+    pattern: '\bWiFi 6E\b'
+    replacement: 'wifi six E'
+    enabled: true
+    description: "WiFi 6E standard"
+    private: false
+  
+  - name: "wifi_7"
+    order: 90
+    pattern: '\bWiFi 7\b'
+    replacement: 'wifi seven'
+    enabled: true
+    description: "WiFi 7 standard"
+    private: false
+  
+  # Units and measurements
+  - name: "gigahertz"
+    order: 100
+    pattern: '\b(\d+(?:\.\d+)?)\s*GHz\b'
+    replacement: '$1 gigahertz'
+    enabled: true
+    description: "Expand GHz to gigahertz"
+    private: false
+  
+  - name: "megahertz"
+    order: 110
+    pattern: '\b(\d+(?:\.\d+)?)\s*MHz\b'
+    replacement: '$1 megahertz'
+    enabled: true
+    description: "Expand MHz to megahertz"
+    private: false
+  
+  - name: "gigabytes"
+    order: 120
+    pattern: '\b(\d+(?:\.\d+)?)\s*GB\b'
+    replacement: '$1 gigabytes'
+    enabled: true
+    description: "Expand GB to gigabytes"
+    private: false
+  
+  - name: "terabytes"
+    order: 130
+    pattern: '\b(\d+(?:\.\d+)?)\s*TB\b'
+    replacement: '$1 terabytes'
+    enabled: true
+    description: "Expand TB to terabytes"
+    private: false
+  
+  - name: "milliseconds"
+    order: 140
+    pattern: '\b(\d+(?:\.\d+)?)\s*ms\b'
+    replacement: '$1 milliseconds'
+    enabled: true
+    description: "Expand ms to milliseconds"
+    private: false
+  
+  # Cloud providers
+  - name: "aws"
+    order: 150
+    pattern: '\bAWS\b'
+    replacement: 'A W S'
+    enabled: true
+    description: "Amazon Web Services"
+    private: false
+  
+  - name: "gcp"
+    order: 160
+    pattern: '\bGCP\b'
+    replacement: 'G C P'
+    enabled: true
+    description: "Google Cloud Platform"
+    private: false
+  
+  # Programming terms
+  - name: "json"
+    order: 170
+    pattern: '\bJSON\b'
+    replacement: 'jason'
+    enabled: true
+    description: "Pronounce JSON as Jason"
+    private: false
+  
+  - name: "yaml"
+    order: 180
+    pattern: '\bYAML\b'
+    replacement: 'yammel'
+    enabled: true
+    description: "Pronounce YAML as yammel"
+    private: false
+  
+  - name: "api"
+    order: 190
+    pattern: '\bAPI\b'
+    replacement: 'A P I'
+    enabled: true
+    description: "Pronounce API as individual letters"
+    private: false
+  
+  - name: "cli"
+    order: 200
+    pattern: '\bCLI\b'
+    replacement: 'C L I'
+    enabled: true
+    description: "Command Line Interface"
+    private: false
+  
+  # Project specific
+  - name: "mcp"
+    order: 210
+    pattern: '\bMCP\b'
+    replacement: 'M C P'
+    enabled: true
+    description: "Model Context Protocol"
+    private: false
+  
+  - name: "tts"
+    order: 220
+    pattern: '\bTTS\b'
+    replacement: 'T T S'
+    enabled: true
+    description: "Text to Speech"
+    private: false
+  
+  - name: "stt"
+    order: 230
+    pattern: '\bSTT\b'
+    replacement: 'S T T'
+    enabled: true
+    description: "Speech to Text"
+    private: false
+
+stt_rules:
+  # Common Whisper mishearings
+  - name: "metool_correction"
+    order: 10
+    pattern: '\bme tool\b'
+    replacement: 'metool'
+    enabled: true
+    description: "Correct Whisper hearing 'me tool' as two words"
+    private: false
+  
+  - name: "cora_seven"
+    order: 20
+    pattern: '\bcora 7\b'
+    replacement: 'Cora 7'
+    enabled: true
+    description: "Ensure proper capitalization of Cora 7"
+    private: false
+  
+  - name: "ai_cora"
+    order: 30
+    pattern: '\bAI cora\b'
+    replacement: 'ai-cora'
+    enabled: true
+    description: "GitHub username format"
+    private: false
+  
+  # Programming terms often misheard
+  - name: "python_capital"
+    order: 40
+    pattern: '\bpython\b'
+    replacement: 'Python'
+    enabled: true
+    description: "Capitalize Python language name"
+    private: false
+  
+  - name: "github_capital"
+    order: 50
+    pattern: '\bgithub\b'
+    replacement: 'GitHub'
+    enabled: true
+    description: "Correct GitHub capitalization"
+    private: false
+  
+  - name: "tmux_spelling"
+    order: 60
+    pattern: '\b(tee mux|t mux|teamux)\b'
+    replacement: 'tmux'
+    enabled: true
+    description: "Correct various tmux mishearings"
+    private: false
+  
+  - name: "neovim_spelling"
+    order: 70
+    pattern: '\b(neo vim|neovem|neovam)\b'
+    replacement: 'Neovim'
+    enabled: true
+    description: "Correct Neovim mishearings"
+    private: false
+  
+  # Common phrases
+  - name: "lets_contraction"
+    order: 80
+    pattern: '\blet us\b'
+    replacement: "let's"
+    enabled: true
+    description: "Contract 'let us' to let's"
+    private: false
+  
+  - name: "its_contraction"
+    order: 90
+    pattern: '\bit is\b'
+    replacement: "it's"
+    enabled: false  # Disabled by default as it can be wrong
+    description: "Contract 'it is' to it's (use carefully)"
+    private: false