shell-history-alias-generator/shell_alias_gen/command_analyzer.py

"""Command analysis and alias generation module."""

from collections import Counter
from dataclasses import dataclass, field
from typing import List, Dict, Optional, Tuple
from .parsers import ParsedCommand


@dataclass
class AliasSuggestion:
    """Represents a suggested alias."""
    alias_name: str
    original_command: str
    frequency: int
    score: float
    
    def to_shell(self, shell: str = 'bash') -> str:
        """Generate alias string for specified shell."""
        if shell == 'fish':
            return f"alias {self.alias_name}='{self.original_command}'"
        return f"alias {self.alias_name}='{self.original_command}'"


class CommandAnalyzer:
    """Analyzes shell commands and generates alias suggestions."""
    
    MIN_COMMAND_LENGTH = 15
    MIN_FREQUENCY = 2
    SCORE_WEIGHTS = {
        'frequency': 1.0,
        'length': 0.5,
        'complexity': 0.3,
    }
    
    def __init__(self):
        self.used_aliases: Dict[str, int] = {}
    
    def analyze_commands(self, commands: List[ParsedCommand]) -> List[AliasSuggestion]:
        """Analyze commands and return alias suggestions."""
        command_counter = Counter()
        
        for cmd in commands:
            normalized = cmd.normalized
            if len(normalized) >= self.MIN_COMMAND_LENGTH:
                command_counter[normalized] += 1
        
        suggestions = []
        for command, frequency in command_counter.most_common():
            if frequency >= self.MIN_FREQUENCY:
                score = self._calculate_score(command, frequency)
                alias_name = self._generate_alias_name(command, frequency)
                
                if self._is_valid_alias(alias_name, command):
                    suggestion = AliasSuggestion(
                        alias_name=alias_name,
                        original_command=command,
                        frequency=frequency,
                        score=score,
                    )
                    suggestions.append(suggestion)
        
        suggestions.sort(key=lambda x: x.score, reverse=True)
        return suggestions[:50]
    
    def _calculate_score(self, command: str, frequency: int) -> float:
        """Calculate a score for an alias suggestion."""
        length_score = min(len(command) / 100, 1.0)
        freq_score = min(frequency / 10, 1.0)
        complexity = self._calculate_complexity(command)
        complexity_score = min(complexity / 20, 1.0)
        
        return (
            freq_score * self.SCORE_WEIGHTS['frequency'] +
            length_score * self.SCORE_WEIGHTS['length'] +
            complexity_score * self.SCORE_WEIGHTS['complexity']
        )
    
    def _calculate_complexity(self, command: str) -> int:
        """Calculate command complexity based on various factors."""
        complexity = 0
        words = command.split()
        
        complexity += len(words) * 2
        complexity += sum(1 for w in words if len(w) > 5)
        complexity += sum(1 for c in command if c in '-/')
        complexity += command.count('--') * 2
        
        flags = [w for w in words if w.startswith('-')]
        complexity += len(flags) * 2
        
        return complexity
    
    def _generate_alias_name(self, command: str, frequency: int) -> str:
        """Generate a short alias name from a command."""
        words = command.split()
        
        if not words:
            return 'alias_cmd'
        
        candidates = []
        
        first_letters = ''.join(w[0] for w in words if w and len(w) > 0)
        if len(first_letters) >= 2:
            candidates.append(first_letters[:8].lower())
        
        main_words = [w for w in words if not w.startswith('-') and not w.startswith('$')]
        if main_words:
            meaningful = [w.rstrip(',.;:') for w in main_words if len(w) > 2]
            if meaningful:
                base = meaningful[0]
                candidates.append(base[:10].lower())
                if len(meaningful) > 1:
                    candidates.append((meaningful[0][:4] + meaningful[1][:4]).lower())
        
        keywords = ['git', 'docker', 'npm', 'pip', 'python', 'ssh', 'curl', 'grep']
        for kw in keywords:
            if kw in command.lower():
                candidates.insert(0, kw)
                break
        
        base_name = candidates[0] if candidates else 'cmd'
        
        suffix = 0
        final_name = base_name
        while final_name in self.used_aliases or not self._is_valid_alias_name(final_name):
            suffix += 1
            final_name = f"{base_name}{suffix}"
            if suffix > 100:
                final_name = f"alias_{hash(command) % 10000}"
                break
        
        self.used_aliases[final_name] = frequency
        return final_name
    
    def _is_valid_alias_name(self, name: str) -> bool:
        """Check if name is a valid shell alias identifier."""
        if not name:
            return False
        if not name[0].isalpha() and name[0] != '_':
            return False
        return all(c.isalnum() or c == '_' for c in name)
    
    def _is_valid_alias(self, alias_name: str, command: str) -> bool:
        """Validate alias is suitable for creation."""
        if not self._is_valid_alias_name(alias_name):
            return False
        if len(alias_name) > 64:
            return False
        if len(command) > 1000:
            return False
        return True
    
    def filter_suggestions(
        self,
        suggestions: List[AliasSuggestion],
        min_score: float = 0.1,
        max_count: int = 20
    ) -> List[AliasSuggestion]:
        """Filter suggestions by score and count."""
        filtered = [s for s in suggestions if s.score >= min_score]
        return filtered[:max_count]