Add project and language analyzers

src/contextgen/analyzers/language_detector.py

@@ -0,0 +1,102 @@
+"""Language detector using file extensions and patterns."""
+
+from pathlib import Path
+from typing import Any
+
+from contextgen.analyzers.base import BaseAnalyzer
+
+
+LANGUAGE_EXTENSIONS: dict[str, list[str]] = {
+    "Python": [".py", ".pyi"],
+    "JavaScript": [".js", ".mjs", ".cjs"],
+    "TypeScript": [".ts", ".tsx"],
+    "Go": [".go"],
+    "Rust": [".rs"],
+    "Java": [".java"],
+    "C": [".c", ".h"],
+    "C++": [".cpp", ".cc", ".cxx", ".hpp"],
+    "C#": [".cs"],
+    "Ruby": [".rb"],
+    "PHP": [".php"],
+    "Swift": [".swift"],
+    "Kotlin": [".kt", ".kts"],
+    "Scala": [".scala"],
+    "HTML": [".html", ".htm"],
+    "CSS": [".css", ".scss", ".sass", ".less"],
+    "JSON": [".json"],
+    "YAML": [".yaml", ".yml"],
+    "XML": [".xml"],
+    "Shell": [".sh", ".bash", ".zsh"],
+    "Markdown": [".md", ".markdown"],
+    "SQL": [".sql"],
+    "Dockerfile": ["Dockerfile"],
+    "Vue": [".vue"],
+    "Svelte": [".svelte"],
+    "Solid": [".tsx"],
+}
+
+LANGUAGE_PATTERNS: dict[str, list[tuple[str, str]]] = {
+    "Python": [("import ", "from "), ("def ", "class "), ("if __name__ == ", "print(")],
+    "JavaScript": [("function ", "const "), ("export ", "import "), ("=>", "console.log(")],
+    "TypeScript": [("interface ", "type "), ("export ", "import "), (": string", ": number")],
+    "Go": [("package ", "func "), ("import (", "type "), ("struct ", "func (")],
+    "Rust": [("fn ", "struct "), ("impl ", "pub "), ("use ", "mod ")],
+    "Java": [("public class", "private "), ("import ", "public "), ("System.out.", "void ")],
+    "Ruby": [("def ", "class "), ("require ", "module "), ("puts ", "end")],
+    "PHP": [("<?php", "function "), ("class ", "public "), ("require ", "namespace ")],
+}
+
+
+class LanguageDetector(BaseAnalyzer):
+    """Detects programming languages used in the project."""
+
+    def analyze(self) -> dict[str, Any]:
+        """Detect languages and return statistics."""
+        language_counts: dict[str, int] = {}
+        language_files: dict[str, list[str]] = {}
+        
+        for ext, lang in self._extension_to_language().items():
+            files = list(self.project_path.rglob(f"*{ext}"))
+            filtered_files = [f for f in files if not self._is_ignored(f)]
+            
+            count = len(filtered_files)
+            if count > 0:
+                language_counts[lang] = language_counts.get(lang, 0) + count
+                language_files.setdefault(lang, []).extend(
+                    [str(f.relative_to(self.project_path)) for f in filtered_files[:10]]
+                )
+        
+        primary_language = max(language_counts, key=language_counts.get) if language_counts else None
+        
+        return {
+            "languages": language_counts,
+            "language_files": language_files,
+            "primary_language": primary_language,
+            "language_details": self._get_language_details(language_counts),
+        }
+
+    def _extension_to_language(self) -> dict[str, str]:
+        """Create a mapping from extensions to languages."""
+        mapping: dict[str, str] = {}
+        for lang, exts in LANGUAGE_EXTENSIONS.items():
+            for ext in exts:
+                mapping[ext] = lang
+        return mapping
+
+    def _is_ignored(self, path: Path) -> bool:
+        """Check if file should be ignored."""
+        ignore_names = [".git", "__pycache__", "node_modules", ".venv", "venv", "dist", "build", ".tox", ".nox"]
+        return any(part in path.parts for part in ignore_names)
+
+    def _get_language_details(self, language_counts: dict[str, int]) -> dict[str, dict[str, Any]]:
+        """Get detailed information about each detected language."""
+        details: dict[str, dict[str, Any]] = {}
+        
+        for lang, count in language_counts.items():
+            details[lang] = {
+                "file_count": count,
+                "extensions": LANGUAGE_EXTENSIONS.get(lang, []),
+                "common_patterns": LANGUAGE_PATTERNS.get(lang, []),
+            }
+        
+        return details