utils.py

from sklearn.metrics import accuracy_score
import typing

def find_common_characters(strings):
    '''
    find common characters occuring in all the strings
    '''
    common_chars = set(strings[0])

    for string in strings[1:]:
        common_chars &= set(string)
    
    return common_chars

def remove_common_characters(strings):
    '''
    remove common characters occuring in all the stirings
    '''
    new_strings = []
    
    common_chars = find_common_characters(strings)
    
    for string in strings:
        # Remove all common characters from the string
        new_string = ''.join(char for char in string if char not in common_chars)
        new_strings.append(new_string)

    return new_strings
        
def print_accuracy(y_valid, y_pred, title : str = 'model') :
    print(f"Accuracy for {title} is {accuracy_score(y_valid, y_pred)}")

def get_char_columns(df):
    '''
    transorm strings into columns, where ith column represents the ith character of the string
    '''
    for i in range(3):
        df[f'c_{i+1}'] = df['input_emoticon'].apply(lambda x, _i=i: x[_i])
        
    columns = df.columns.to_list()
    if 'label' in columns:
        columns.remove('label')
    columns.remove('input_emoticon')
    
    return df[columns + (['label'] if 'label' in df.columns else []) ]

def process_strings(strs: typing.List[str]) -> typing.List[str]:
    '''
    remove common substrings occuring in all the strings, which are encodings of the common emojis 
    '''
    strs = [x.lstrip("0") for x in strs]

    #FIXME : evaluate repeating substrings from the strings, not using the encodding mapping
    repeat_emo_code = {
        '🙼' : '284',
        '🛐' : '464', 
        '🙯' : '262',
        '😛' : '15436', 
        '😣' : '614',
        '😑' : '1596', 
        '🚼' : '422'
    }

    # Example usage
    substrings = repeat_emo_code.values()

    # Remove the substrings
    strs = [remove_substrings(input_string, substrings) for input_string in strs]

    padded_strs = []

    for s in strs:
        if len(s) < 15:
            s = s + "0" * (15 - len(s))
        padded_strs.append(s)

    return padded_strs


def remove_substrings(input_string, substrings):
    """
    Removes all occurrences of substrings from the input string.
    """
    for substring in substrings:
        input_string = input_string.replace(substring, "")
    return input_string

def get_columns(df, num_feat):
    '''
    for the text sequences, convert into columns, where ith columns contains ith character of the string
    '''
    for i in range(num_feat):
        df[f"c_{i}"] = df["input_str"].apply(lambda x: x[i])
    return df.drop(columns=["input_str"])

def print_delimiter() :
    print("=========================================")