Add package.json-based repo language detection

src/codegen/sdk/utils.py

@@ -5,6 +5,7 @@
 from collections.abc import Iterable
 from contextlib import contextmanager
 from pathlib import Path
+from typing import Literal
 from xml.dom.minidom import parseString
 
 import dicttoxml
@@ -242,54 +243,65 @@ def get_language_file_extensions(language: ProgrammingLanguage):
         return set(TSFile.get_extensions())
 
 
-def determine_project_language(folder_path: str):
-    from codegen.sdk.python import PyFile
-    from codegen.sdk.typescript.file import TSFile
+def determine_project_language(folder_path: str, strategy: Literal["most_common", "package_json"] = "package_json") -> ProgrammingLanguage:
+    if strategy == "most_common":
+        # Analyzes a folder to determine the primary programming language based on file extensions.
+        # Returns the language with the most matching files.
+        from codegen.sdk.python import PyFile
+        from codegen.sdk.typescript.file import TSFile
 
-    EXTENSIONS = {
-        ProgrammingLanguage.PYTHON: PyFile.get_extensions(),
-        ProgrammingLanguage.TYPESCRIPT: TSFile.get_extensions(),
-    }
+        EXTENSIONS = {
+            ProgrammingLanguage.PYTHON: PyFile.get_extensions(),
+            ProgrammingLanguage.TYPESCRIPT: TSFile.get_extensions(),
+        }
 
-    """
-    Analyzes a folder to determine the primary programming language based on file extensions.
-    Returns the language with the most matching files.
+        """
+        Analyzes a folder to determine the primary programming language based on file extensions.
+        Returns the language with the most matching files.
 
-    Args:
-        folder_path (str): Path to the folder to analyze
+        Args:
+            folder_path (str): Path to the folder to analyze
 
-    Returns:
-        Optional[ProgrammingLanguage]: The dominant programming language, or None if no matching files found
-    """
-    folder = Path(folder_path)
-    if not folder.exists() or not folder.is_dir():
-        msg = f"Invalid folder path: {folder_path}"
-        raise ValueError(msg)
-
-    # Initialize counters for each language
-    language_counts = Counter()
-
-    # Walk through the directory
-    for file_path in folder.rglob("*"):
-        # Skip directories and hidden files
-        if file_path.is_dir() or file_path.name.startswith("."):
-            continue
-
-        # Skip common directories to ignore
-        if any(ignore in str(file_path) for ignore in [".git", "node_modules", "__pycache__", "venv", ".env"]):
-            continue
-
-        # Count files for each language based on extensions
-        for language, exts in EXTENSIONS.items():
-            if file_path.suffix in exts:
-                language_counts[language] += 1
-
-    # If no files found, return None
-    if not language_counts:
-        return ProgrammingLanguage.UNSUPPORTED
-
-    # Return the language with the highest count
-    return language_counts.most_common(1)[0][0]
+        Returns:
+            Optional[ProgrammingLanguage]: The dominant programming language, or None if no matching files found
+        """
+        folder = Path(folder_path)
+        if not folder.exists() or not folder.is_dir():
+            msg = f"Invalid folder path: {folder_path}"
+            raise ValueError(msg)
+
+        # Initialize counters for each language
+        language_counts = Counter()
+
+        # Walk through the directory
+        for file_path in folder.rglob("*"):
+            # Skip directories and hidden files
+            if file_path.is_dir() or file_path.name.startswith("."):
+                continue
+
+            # Skip common directories to ignore
+            if any(ignore in str(file_path) for ignore in [".git", "node_modules", "__pycache__", "venv", ".env"]):
+                continue
+
+            # Count files for each language based on extensions
+            for language, exts in EXTENSIONS.items():
+                if file_path.suffix in exts:
+                    language_counts[language] += 1
+
+        # If no files found, return None
+        if not language_counts:
+            return ProgrammingLanguage.UNSUPPORTED
+
+        # Return the language with the highest count
+        return language_counts.most_common(1)[0][0]
+    elif strategy == "package_json":
+        # TODO: Hacky implementation that checks for package.json.
+        # Faster but less accurate than the most_common strategy.
+        package_json_path = Path(folder_path) / "package.json"
+        if package_json_path.exists():
+            return ProgrammingLanguage.TYPESCRIPT
+        else:
+            return ProgrammingLanguage.PYTHON
 
 
 def split_git_path(filepath: str) -> tuple[str, str | None]: