Source code for autogluon.features.generators.text_special

import logging
import re
from typing import List

import numpy as np
import pandas as pd
from pandas import DataFrame, Series

from autogluon.core.features.types import S_IMAGE_PATH, S_TEXT, S_TEXT_SPECIAL

from .abstract import AbstractFeatureGenerator
from .binned import BinnedFeatureGenerator

logger = logging.getLogger(__name__)


[docs]class TextSpecialFeatureGenerator(AbstractFeatureGenerator): """ TextSpecialFeatureGenerator generates text specific features from incoming raw text features. These include word counts, character counts, symbol counts, capital letter ratios, and much more. Features generated by this generator will have 'text_special' as a special type. Parameters ---------- symbols : List[str], optional List of string symbols to compute counts and ratios for as features. If not specified, defaults to ['!', '?', '@', '%', '$', '*', '&', '#', '^', '.', ':', ' ', '/', ';', '-', '='] min_occur_ratio : float, default 0.01 Minimum ratio of symbol occurrence to consider as a feature. If a symbol appears in fewer than 1 in 1/min_occur_ratio samples, it will not be used as a feature. min_occur_offset : int, default 10 Minimum symbol occurrences to consider as a feature. This is added to the threshold calculated from min_occur_ratio. bin_features : bool, default True If True, adds a BinnedFeatureGenerator to the front of post_generators such that all features generated from this generator are then binned. This is useful for 'text_special' features because it lowers the chance models will overfit on the features and reduces their memory usage. post_drop_duplicates : bool, default True Identical to AbstractFeatureGenerator's post_drop_duplicates, except it is defaulted to True instead of False. This helps to clean the output of this generator when symbols aren't present in the data. **kwargs : Refer to AbstractFeatureGenerator documentation for details on valid key word arguments. """ def __init__(self, symbols: List[str] = None, min_occur_ratio=0.01, min_occur_offset=10, bin_features: bool = True, post_drop_duplicates: bool = True, **kwargs): super().__init__(post_drop_duplicates=post_drop_duplicates, **kwargs) if symbols is None: symbols = ['!', '?', '@', '%', '$', '*', '&', '#', '^', '.', ':', ' ', '/', ';', '-', '='] self._symbols = symbols # Symbols to generate count and ratio features for. self._symbols_per_feature = dict() self._min_occur_ratio = min_occur_ratio self._min_occur_offset = min_occur_offset if bin_features: self._post_generators = [BinnedFeatureGenerator()] + self._post_generators def _fit_transform(self, X: DataFrame, **kwargs) -> (DataFrame, dict): self._symbols_per_feature = self._filter_symbols(X, self._symbols) self._feature_names_dict = self._compute_feature_names_dict() X_out = self._transform(X) type_family_groups_special = { S_TEXT_SPECIAL: list(X_out.columns) } return X_out, type_family_groups_special def _transform(self, X: DataFrame) -> DataFrame: return self._generate_features_text_special(X) def _compute_feature_names_dict(self) -> dict: feature_names = dict() for feature in self.features_in: feature_names_cur = dict() for feature_name_base in ['char_count', 'word_count', 'capital_ratio', 'lower_ratio', 'digit_ratio', 'special_ratio']: feature_names_cur[feature_name_base] = f'{feature}.{feature_name_base}' symbols = self._symbols_per_feature[feature] for symbol in symbols: feature_names_cur[symbol] = {} feature_names_cur[symbol]['count'] = f'{feature}.symbol_count.{symbol}' feature_names_cur[symbol]['ratio'] = f'{feature}.symbol_ratio.{symbol}' feature_names[feature] = feature_names_cur return feature_names @staticmethod def get_default_infer_features_in_args() -> dict: return dict(required_special_types=[S_TEXT], invalid_special_types=[S_IMAGE_PATH]) def _filter_symbols(self, X: DataFrame, symbols: list): symbols_per_feature = dict() if self.features_in: num_samples = len(X) occur_threshold = min(np.ceil(self._min_occur_offset + num_samples * self._min_occur_ratio), np.ceil(num_samples / 2)) for nlp_feature in self.features_in: symbols_feature = [] for symbol in symbols: occur_symbol = np.sum([value.count(symbol) != 0 for value in X[nlp_feature]]) if occur_symbol >= occur_threshold: symbols_feature.append(symbol) symbols_per_feature[nlp_feature] = np.array(symbols_feature) return symbols_per_feature def _generate_features_text_special(self, X: DataFrame) -> DataFrame: if self.features_in: X_text_special_combined = dict() for nlp_feature in self.features_in: X_text_special_combined = self._generate_text_special(X[nlp_feature], nlp_feature, symbols=self._symbols_per_feature[nlp_feature], X_dict=X_text_special_combined) X_text_special_combined = pd.DataFrame(X_text_special_combined, index=X.index) else: X_text_special_combined = pd.DataFrame(index=X.index) return X_text_special_combined def _generate_text_special(self, X: Series, feature: str, symbols: list, X_dict: dict) -> dict: fn = self._feature_names_dict[feature] X_dict[fn['char_count']] = np.array([self.char_count(value) for value in X], dtype=np.uint32) X_dict[fn['word_count']] = np.array([self.word_count(value) for value in X], dtype=np.uint32) X_dict[fn['capital_ratio']] = np.array([self.capital_ratio(value) for value in X], dtype=np.float32) X_dict[fn['lower_ratio']] = np.array([self.lower_ratio(value) for value in X], dtype=np.float32) X_dict[fn['digit_ratio']] = np.array([self.digit_ratio(value) for value in X], dtype=np.float32) X_dict[fn['special_ratio']] = np.array([self.special_ratio(value) for value in X], dtype=np.float32) char_count = X_dict[fn['char_count']] char_count_valid = char_count != 0 shape = char_count.shape for symbol in symbols: X_dict[fn[symbol]['count']] = np.array([value.count(symbol) for value in X], dtype=np.uint32) X_dict[fn[symbol]['ratio']] = np.divide(X_dict[fn[symbol]['count']], char_count, out=np.zeros(shape, dtype=np.float32), where=char_count_valid) return X_dict @staticmethod def word_count(string: str) -> int: return len(string.split()) @staticmethod def char_count(string: str) -> int: return len(string) @staticmethod def special_ratio(string: str) -> float: string = string.replace(' ', '') if not string: return 0 new_str = re.sub(r'[\w]+', '', string) return len(new_str) / len(string) @staticmethod def digit_ratio(string: str) -> float: string = string.replace(' ', '') if not string: return 0 return sum(c.isdigit() for c in string) / len(string) @staticmethod def lower_ratio(string: str) -> float: string = string.replace(' ', '') if not string: return 0 return sum(c.islower() for c in string) / len(string) @staticmethod def capital_ratio(string: str) -> float: string = string.replace(' ', '') if not string: return 0 return sum(1 for c in string if c.isupper()) / len(string) def _remove_features_in(self, features: list): super()._remove_features_in(features) if self._symbols_per_feature: for feature in features: if feature in self._symbols_per_feature: self._symbols_per_feature.pop(feature)