Rob200/LegacyCodeDocumentAgent

1"""
2File handling utilities for the Legacy Code Documentation Agent.
3Reads code files and detects their language.
4"""
5
6import os
7
8# Map file extensions to language names
9EXTENSION_MAP = {
10    ".sql": "SQL",
11    ".py": "Python",
12    ".cpp": "C++",
13    ".h": "C++",
14    ".dax": "DAX",
15    ".m": "DAX",  # Power BI measure files sometimes use .m
16}
17
18
19def get_language_from_extension(filepath: str) -> str:
20    """
21    Determine the programming language based on file extension.
22
23    Args:
24        filepath: Path to the code file
25
26    Returns:
27        Language name as a string (e.g., "SQL", "Python")
28    """
29    _, extension = os.path.splitext(filepath)
30    extension = extension.lower()
31
32    return EXTENSION_MAP.get(extension, "Unknown")
33
34
35def read_code_file(filepath: str) -> dict:
36    """
37    Read a code file and return its contents with metadata.
38
39    Args:
40        filepath: Path to the code file
41
42    Returns:
43        Dictionary with keys: 'filename', 'language', 'content', 'success', 'error'
44    """
45    result = {
46        "filename": os.path.basename(filepath),
47        "language": None,
48        "content": None,
49        "success": False,
50        "error": None
51    }
52
53    # Check if file exists
54    if not os.path.exists(filepath):
55        result["error"] = f"File not found: {filepath}"
56        return result
57
58    # Get the language
59    result["language"] = get_language_from_extension(filepath)
60
61    # Read the file contents
62    try:
63        with open(filepath, "r", encoding="utf-8") as f:
64            result["content"] = f.read()
65        result["success"] = True
66    except Exception as e:
67        result["error"] = f"Error reading file: {str(e)}"
68
69    return result
70
71
72def get_code_files_from_folder(folder_path: str) -> list:
73    """
74    Scan a folder and return a list of supported code files.
75
76    Args:
77        folder_path: Path to the folder to scan
78
79    Returns:
80        List of file paths for supported code files
81    """
82    supported_files = []
83
84    # Check if folder exists
85    if not os.path.isdir(folder_path):
86        return supported_files
87
88    # Walk through the folder
89    for root, dirs, files in os.walk(folder_path):
90        for filename in files:
91            filepath = os.path.join(root, filename)
92            language = get_language_from_extension(filepath)
93
94            # Only include files we know how to handle
95            if language != "Unknown":
96                supported_files.append(filepath)
97
98    return supported_files