number of data transformers for sentiment analysis

transformers.py

import re
from sklearn.base import BaseEstimator, TransformerMixin


class RemoveEllipseTransformer(TransformerMixin):

    @staticmethod
    def _preprocess_data(data_series):
        """
        inspired from:
        https://raw.githubusercontent.com/youngsoul/ml-twitter-sentiment-analysis/develop/cleanup.py
        :param data_series:
        :return:
        """
        # remove ..., ...., ..... with space
        for remove in map(lambda r: re.compile(r), ["\.\.\.\.\.", "\.\.\.\.", "\.\.\."]):
            data_series.replace(remove, " ", inplace=True)

    def fit(self, X, y=None):
        return self

    def transform(self, X):
        """

        :param X: Series, aka column of data.
        :return:
        """
        RemoveEllipseTransformer._preprocess_data(X)
        return X


class RemoveNumbersTransformer(TransformerMixin):

    @staticmethod
    def _preprocess_data(data_series):
        """
        inspired from:
        https://raw.githubusercontent.com/youngsoul/ml-twitter-sentiment-analysis/develop/cleanup.py
        :param data_series:
        :return:
        """

        # remove numbers
        regex = re.compile(r"\s?[0-9]+\.?[0-9]*")
        data_series.replace(regex, "", inplace=True)

    def fit(self, X, y=None):
        return self

    def transform(self, X):
        """

        :param X: Series, aka column of data.
        :return:
        """
        RemoveNumbersTransformer._preprocess_data(X)
        return X


class RemoveSpecialCharactersTransformer(TransformerMixin):

    @staticmethod
    def _preprocess_data(data_series):
        """
        inspired from:
        https://raw.githubusercontent.com/youngsoul/ml-twitter-sentiment-analysis/develop/cleanup.py
        :param data_series:
        :return:
        """

        # remove special characters
        for remove in map(lambda r: re.compile(re.escape(r)), [",", ":", "\"", "=", "&", ";", "%", "$",
                                                               "@", "%", "^", "*", "(", ")", "{", "}",
                                                               "[", "]", "|", "/", "\\", ">", "<", "-",
                                                               "!", "?", ".", "'",
                                                               "--", "---", "#"]):
            data_series.replace(remove, "", inplace=True)

    def fit(self, X, y=None):
        return self

    def transform(self, X):
        """

        :param X: Series, aka column of data.
        :return:
        """
        RemoveSpecialCharactersTransformer._preprocess_data(X)
        return X


class RemoveHtmlEncodedTransformer(TransformerMixin):

    @staticmethod
    def _preprocess_data(data_series):
        """
        inspired from:
        https://raw.githubusercontent.com/youngsoul/ml-twitter-sentiment-analysis/develop/cleanup.py
        :param data_series:
        :return:
        """

        # html encoded characters
        for remove in map(lambda r: re.compile(r), ["&lt;", "&gt;", "&quot;", "&amp;", "w/o", "w/"]):
            data_series.replace(remove, "", inplace=True)

    def fit(self, X, y=None):
        return self

    def transform(self, X):
        """

        :param X: Series, aka column of data.
        :return:
        """
        RemoveHtmlEncodedTransformer._preprocess_data(X)
        return X


class RemoveUsernameTransformer(TransformerMixin):

    @staticmethod
    def _preprocess_data(data_series):
        """
        inspired from:
        https://raw.githubusercontent.com/youngsoul/ml-twitter-sentiment-analysis/develop/cleanup.py
        :param data_series:
        :return:
        """

        # remove user name
        regex = re.compile(r"@[^\s]+[\s]?")
        data_series.replace(regex, "", inplace=True)

    def fit(self, X, y=None):
        return self

    def transform(self, X):
        """

        :param X: Series, aka column of data.
        :return:
        """
        RemoveUsernameTransformer._preprocess_data(X)
        return X


class RemoveUrlsTransformer(TransformerMixin):

    @staticmethod
    def _preprocess_data(data_series):
        """
        inspired from:
        https://raw.githubusercontent.com/youngsoul/ml-twitter-sentiment-analysis/develop/cleanup.py
        :param data_series:
        :return:
        """

        # remove urls
        regex = re.compile(r"http.?://[^\s]+[\s]?")
        data_series.replace(regex, "", inplace=True)

    def fit(self, X, y=None):
        return self

    def transform(self, X):
        """

        :param X: Series, aka column of data.
        :return:
        """
        RemoveUrlsTransformer._preprocess_data(X)
        return X