diff --git a/src/gdiffer/language_detector.py b/src/gdiffer/language_detector.py new file mode 100644 index 0000000..344b600 --- /dev/null +++ b/src/gdiffer/language_detector.py @@ -0,0 +1,75 @@ +"""Language detection for code files.""" + +from typing import Optional + + +class LanguageDetector: + EXTENSION_MAP = { + 'py': 'python', 'pyw': 'python', 'pyx': 'python', + 'js': 'javascript', 'mjs': 'javascript', 'cjs': 'javascript', 'jsx': 'javascript', + 'ts': 'typescript', 'tsx': 'typescript', 'mts': 'typescript', 'cts': 'typescript', + 'java': 'java', 'kt': 'kotlin', 'kts': 'kotlin', + 'go': 'go', 'rs': 'rust', 'c': 'c', 'h': 'c', + 'cpp': 'cpp', 'cc': 'cpp', 'cxx': 'cpp', 'hpp': 'cpp', 'hxx': 'cpp', + 'cs': 'csharp', 'rb': 'ruby', 'erb': 'ruby', 'php': 'php', + 'swift': 'swift', 'm': 'objective-c', 'mm': 'objective-c', + 'scala': 'scala', 'sc': 'scala', 'jl': 'julia', + 'r': 'r', 'R': 'r', 'lua': 'lua', + 'pl': 'perl', 'pm': 'perl', 'sql': 'sql', + 'sh': 'bash', 'bash': 'bash', 'zsh': 'bash', 'fish': 'bash', + 'yaml': 'yaml', 'yml': 'yaml', 'json': 'json', + 'xml': 'xml', 'html': 'html', 'htm': 'html', + 'css': 'css', 'scss': 'scss', 'sass': 'sass', 'less': 'less', + 'md': 'markdown', 'markdown': 'markdown', + 'txt': 'text', 'dockerfile': 'dockerfile', 'Dockerfile': 'dockerfile', + } + + CONTENT_PATTERNS = { + 'python': [r'^import\s+\w+', r'^from\s+\w+\s+import', r'^def\s+\w+\s*\(', r'^class\s+\w+'], + 'javascript': [r'^const\s+\w+', r'^let\s+\w+', r'^var\s+\w+', r'^function\s+\w+', r'=>\s*\{'], + 'typescript': [r'^interface\s+\w+', r'^type\s+\w+', r':\s*(string|number|boolean)'], + 'java': [r'^package\s+[\w.]+;', r'^import\s+[\w.]+;', r'^public\s+class\s+\w+'], + 'go': [r'^package\s+\w+', r'^import\s+\(', r'func\s+\w+'], + 'rust': [r'^fn\s+\w+', r'^impl\s+\w+', r'^struct\s+\w+', r'^enum\s+\w+'], + 'c': [r'#include\s*<', r'#include\s*"', r'int\s+main\s*\('], + 'cpp': [r'#include\s*<', r'#include\s*"', r'class\s+\w+', r'std::\w+'], + 'ruby': [r'^require\s+', r'^class\s+\w+', r'^module\s+\w+', r'def\s+\w+'], + 'php': [r'<\?php', r'\$\w+\s*=', r'function\s+\w+', r'class\s+\w+'], + } + + def detect_from_filename(self, filename: str) -> Optional[str]: + if '.' not in filename: + return None + ext = filename.rsplit('.', 1)[-1].lower() + return self.EXTENSION_MAP.get(ext) + + def detect_from_content(self, content: str) -> Optional[str]: + first_lines = '\n'.join(content.splitlines()[:50]) + scores = {} + for lang, patterns in self.CONTENT_PATTERNS.items(): + import re + score = sum(len(re.findall(p, first_lines, re.MULTILINE)) for p in patterns) + if score > 0: + scores[lang] = score + return max(scores, key=scores.get) if scores else None + + def detect(self, filename: str, content: str = "") -> str: + ext_lang = self.detect_from_filename(filename) + if ext_lang and ext_lang not in ['text', 'markdown', 'json', 'yaml', 'xml', 'html', 'css', 'dockerfile']: + return ext_lang + if content: + content_lang = self.detect_from_content(content) + if content_lang: + return content_lang + return ext_lang or "text" + + def get_supported_languages(self) -> list[str]: + return sorted(set(self.EXTENSION_MAP.values())) + + def is_language_supported(self, language: str) -> bool: + return language in self.get_supported_languages() + + +def detect_language(filename: str, content: str = "") -> str: + detector = LanguageDetector() + return detector.detect(filename, content)