SERiF/utils/defaultConfig/generateDefaultConfig.py

import os
import re
import yaml
from io import StringIO

from tree_sitter import Language, Parser
import tree_sitter_cpp as tscpp

CPP_LANGUAGE = Language(tscpp.language())

def remove_comments_cpp(source_code):
    parser = Parser(CPP_LANGUAGE)
    tree = parser.parse(source_code.encode())
    def extract_code(node):
        if node.type in ['comment']:
            return ''
        elif node.children:
            return ''.join(extract_code(child) for child in node.children)
        return source_code[node.start_byte:node.end_byte]
    extractedCode = extract_code(tree.root_node)
    return extractedCode

# Regular expression to match .get<T>("a:b:c", x)
pattern = re.compile(r'\.get<([^>]+)>\("([^"]+)",\s*([^);]+)\)')

def parse_value(value, type_hint):
    """Converts values to appropriate types."""
    value = value.strip()
    
    # Handle common types
    if type_hint in {"int", "long"}:
        return (int(value), type_hint)
    elif type_hint in {"float", "double"}:
        return float(value)
    elif value == "defaultDataDir":
        return "data" # special case for defaultDataDir
        return (float(value), type_hint)
    elif type_hint == "bool":
        return (value.lower() in {"true", "1"}, type_hint)
    elif value.startswith('"') and value.endswith('"'):
        return (value.strip('"'), type_hint)  # Remove quotes for string literals
    elif value.startswith("'") and value.endswith("'"):
        return (value.strip("'"), type_hint)  # Remove single quotes
        return value.strip("'")  # Remove single quotes
    elif value.startswith('"') and value.endswith('"'):
        return (value.strip('"'). type_hint)# Remove quotes for string literals
    
    return (value, type_hint)  # Return as-is if unsure

def insert_into_dict(hierarchy_dict, keys, value, type_hint):
    """Recursively inserts values into a nested dictionary."""
    key = keys.pop(0)
    if not keys:  # If last key, store value
        hierarchy_dict[key] = parse_value(value, type_hint)
    else:
        if key not in hierarchy_dict:
            hierarchy_dict[key] = {}
        insert_into_dict(hierarchy_dict[key], keys, value, type_hint)

def scan_files(directory):
    """Scans all .cpp, .h, .c, and .hpp files in a directory recursively."""
    hierarchy_dict = {}

    for root, _, files in os.walk(directory):
        for file in files:
            if file.endswith(('.cpp', '.h', '.c', '.hpp')):
                file_path = os.path.join(root, file)
                with open(file_path, 'r', encoding='utf-8', errors='ignore') as f:
                    noCommentSourceCode = remove_comments_cpp(f.read())
                    matchs = pattern.finditer(noCommentSourceCode)
                    for match in matchs:
                        print(f"Match: {match.group()}")
                        type_hint, hierarchy, value = match.groups()
                        keys = hierarchy.split(":")
                        if keys[0] == "Data" and keys[1] == "Dir":
                            continue # Skip Data:Dir as it is a special case
                        insert_into_dict(hierarchy_dict, keys, value, type_hint)

    return hierarchy_dict

class QuotedString(str):
    pass

def represent_quoted_string(dumper, data):
    return dumper.represent_scalar('tag:yaml.org,2002:str', data, style='"')


def split_dict_recursive(originalDict):
    dataDict = {}
    typeDict = {}

    for key, value in originalDict.items():
        if isinstance(value, dict):  # If the value is another dictionary, recurse
            dataDict[key], typeDict[key] = split_dict_recursive(value)
        elif isinstance(value, tuple) and len(value) == 2:  # If it's a tuple, split it
            dataDict[key] = value[0]  # First element is data
            typeDict[key] = value[1]  # Second element is type hint
        else:
            raise ValueError(f"Unexpected structure in dictionary at key '{key}': {value}")

    return dataDict, typeDict

def save_yaml(data, output_file):
    """Saves the nested dictionary to a YAML file."""
    yaml.add_representer(QuotedString, represent_quoted_string)

    def quote_strings(data):
        if isinstance(data, dict):
            return {k: quote_strings(v) for k, v in data.items()}
        elif isinstance(data, list):
            return [quote_strings(item) for item in data]
        elif isinstance(data, str):
            return QuotedString(data)
        else:
            return data

    data = quote_strings(data)
    options, types = split_dict_recursive(data)
    with open(output_file, 'w', encoding='utf-8') as f:
        yaml.dump(options, f, default_flow_style=False, sort_keys=False, indent=4)
    with open(output_file, 'a') as f:
        f.write("\n")
        f.write("# This file was auto-generated by generateDefaultConfig.py\n")
        f.write("# Do not modify this file directly.\n")
        f.write('\n#')
        f.write('='*35 + " TYPE HINTS " + '='*35 + '\n')
        f.write('\n')

    type_file = StringIO()
    yaml.dump(types, type_file, default_flow_style=False, sort_keys=False, indent=4)
    # write types yaml out put with # in front of each line
    with open(output_file, 'a') as f:
        for line in type_file.getvalue().split('\n'):
            f.write(f"# {line}\n")


if __name__ == "__main__":
    import argparse
    parser = argparse.ArgumentParser(description="Generate a YAML file with default configuration values.")
    parser.add_argument("directory", help="Directory to scan for .cpp, .h, .c, and .hpp files.")
    parser.add_argument("-o", "--output", help="Output file name.", default="defaults.yaml")
    args = parser.parse_args()

    extracted_data = scan_files(args.directory)
    print(extracted_data)
    save_yaml(extracted_data, args.output)

    print(f"✅ YAML file generated: {args.output}")