BidiRag/fix_fstrings.py

"""
Script to convert f-strings to .format() for Python 3.5 compatibility
"""
import re
import os

files_to_fix = [
    "bdirag/document_processor.py",
    "bdirag/embedding_models.py",
    "bdirag/vector_stores.py",
    "bdirag/rag_methods.py",
    "bdirag/benchmark.py",
    "bdirag/config.py",
    "examples/test_bm25.py",
    "examples/benchmark_all_methods.py",
    "examples/quick_demo.py",
    "examples/benchmark_retrieval_speed.py",
    "examples/bid_field_extraction_demo.py",
]


def convert_fstring_in_line(line):
    # Find f"..." or f'...' patterns on a single line
    # Handle double quotes
    pattern_dq = r'f"([^"]*?\{[^}]*\}[^"]*?)"'
    pattern_sq = r"f'([^']*?\{[^}]*\}[^']*?)'"

    for pattern, quote in [(pattern_dq, '"'), (pattern_sq, "'")]:
        while True:
            match = re.search(pattern, line)
            if not match:
                break

            inner = match.group(1)
            values = []
            new_str = ""
            i = 0
            while i < len(inner):
                if inner[i] == "{" and i + 1 < len(inner) and inner[i + 1] != "{":
                    j = inner.index("}", i)
                    expr = inner[i + 1:j].strip()
                    values.append(expr)
                    new_str += "{" + str(len(values) - 1) + "}"
                    i = j + 1
                elif inner[i:i + 2] == "{{":
                    new_str += "{{"
                    i += 2
                elif inner[i:i + 2] == "}}":
                    new_str += "}}"
                    i += 2
                else:
                    new_str += inner[i]
                    i += 1

            format_call = ".format(" + ", ".join(values) + ")"
            replacement = new_str + format_call

            line = line[:match.start()] + quote + replacement + quote + line[match.end():]

    return line


for filepath in files_to_fix:
    if not os.path.exists(filepath):
        print("Not found: " + filepath)
        continue

    with open(filepath, "r", encoding="utf-8") as f:
        content = f.read()

    lines = content.split("\n")
    new_lines = []
    changed = False

    for line in lines:
        if "f'" in line or 'f"' in line:
            new_line = convert_fstring_in_line(line)
            if new_line != line:
                changed = True
            new_lines.append(new_line)
        else:
            new_lines.append(line)

    if changed:
        new_content = "\n".join(new_lines)
        with open(filepath, "w", encoding="utf-8") as f:
            f.write(new_content)
        print("Fixed: " + filepath)
    else:
        print("No changes: " + filepath)