token_positives about groundinglmm HOT 2 CLOSED

def find_word_indices(caption, word):
    """
    Finds the start and end token indices of the first occurrence of a word in a caption.

    Parameters:
    - caption (str): The caption in which to search for the word.
    - word (str): The word to search for.

    Returns:
    - tuple: A tuple containing the start and end token indices of the word. If the word is not found, returns (-1, -1).
    """
    # Convert the caption and word to lowercase to make the search case-insensitive.
    caption_lower = caption.lower()
    word_lower = word.lower()

    # Find the start index of the word in the caption.
    start_idx = caption_lower.find(word_lower)

    # If the word is not found, check for its plural form by adding an 's' at the end.
    if start_idx == -1:
        new_word = word_lower.split('-')[0] + 's'
        start_idx = caption_lower.find(new_word)
        if start_idx == -1:
            return -1, -1, new_word
        else:
            return start_idx, start_idx + len(new_word), new_word

    # If the word is not found, return (-1, -1).
    if start_idx == -1:
        return -1, -1, word

    # Find the end index of the word in the caption.
    end_idx = start_idx + len(word)

    return (start_idx, end_idx, word)

start_idx, end_idx, word = find_word_indices(caption, word)
token_positives = [start_idx, end_idx]