tracking_policy_agendas.classifiers.meta_clf

Complexity

Total Complexity

13

Size/Duplication

Total Lines	116
Duplicated Lines	75.86 %

Test Coverage

Coverage

60.66%

Importance

Changes

0

7 Methods

Rating	Name	Duplication	Size	Complexity
A	MetaClf.load_model()	12	12	3
A	MetaClf.predict()	9	9	1
A	MetaClf.prep_scaler()	9	9	1
A	MetaClf.__init__()	22	22	3
A	MetaClf.save_model()	12	12	3
A	MetaClf.__getitem__()	7	7	1
A	MetaClf.fit()	10	10	1

"""
Meta Class for Classifiers

....................................................................................................
MIT License
Copyright (c) 2021-2023 AUT Iran, Mohammad H Forouhesh
Copyright (c) 2021-2022 MetoData.ai, Mohammad H Forouhesh
....................................................................................................
This module abstracts classifiers.
"""

import os
import pickle
from typing import List

import numpy as np

Unused numpy imported as np

import pandas as pd
from tqdm import tqdm
from sklearn.preprocessing import MinMaxScaler
from sklearn.metrics import classification_report
from sklearn.model_selection import train_test_split

from ..api import get_resources
from ..preprocess.preprocessing import remove_redundant_characters, remove_emoji
from ..word2vec.w2v_emb import W2VEmb


class MetaClf:

This code seems to be duplicated in your project.

Missing class docstring

    def __init__(self, classifier_instance, text_array: list = None, embedding_doc: list = None, labels: list = None, load_path: str = None):

Too many arguments (6/5)

This line is too long as per the coding-style (141/100).

        if not isinstance(text_array, pd.Series): text_array = pd.Series(text_array)

More than one statement on a single line

        self.clf = classifier_instance
        self.emb = W2VEmb()
        self.scaler = None
        self.dir_path = os.path.dirname(
            os.path.dirname(
                os.path.dirname(
                    os.path.realpath(__file__)))) + "/"
        if load_path is not None:
            get_resources(self.dir_path, resource_name=load_path)
            self.load_model(load_path)
        else:
            assert text_array is not None and labels is not None
            text_array.fillna('', inplace=True)
            self.emb = W2VEmb(embedding_doc)

            encoded = list(map(self.emb.encode, tqdm(text_array)))
            self.labels = list(labels)
            self.scaler = self.prep_scaler(encoded)
            self.encoded_input = self.scaler.transform(encoded)

    def prep_scaler(self, encoded: List[int]) -> MinMaxScaler:

This method could be written as a function/class method.

        """
        Fitting a Min-Max Scaler to use in the pipeline
        :param encoded:     An array of numbers.
        :return:            A MinMaxScaler
        """
        scaler = MinMaxScaler()
        scaler.fit(encoded)
        return scaler

    def fit(self):

Missing function or method docstring

        X_train, X_test, y_train, y_test = train_test_split(self.encoded_input, self.labels, test_size=0.2,

This line is too long as per the coding-style (107/100).

Variable name "X_train" doesn't conform to snake_case naming style ('([^\\W\\dA-Z][^\\WA-Z]2,|_[^\\WA-Z]*|__[^\\WA-Z\\d_][^\\WA-Z]+__)$' pattern)

Variable name "X_test" doesn't conform to snake_case naming style ('([^\\W\\dA-Z][^\\WA-Z]2,|_[^\\WA-Z]*|__[^\\WA-Z\\d_][^\\WA-Z]+__)$' pattern)

                                                            random_state=42, stratify=self.labels)
        self.clf.fit(X_train, y_train)
        print('score: ', self.clf.score(X_test, y_test))
        print('============================trian============================')
        print(classification_report(y_train, self.clf.predict(X_train)))
        print('=============================test============================')
        print(classification_report(y_test, self.clf.predict(X_test)))
        return self.clf

    def load_model(self, load_path: str) -> None:
        """
        A tool to load model from disk.
        :param load_path:   Model path.
        :return:            None
        """

        loading_prep = lambda string: f'model_dir/{load_path}/{string}'
        self.clf.load_model(loading_prep('model.json'))
        self.emb.load(loading_prep('emb.pkl'))
        with open(loading_prep('scaler.pkl'), 'rb') as f:

Variable name "f" doesn't conform to snake_case naming style ('([^\\W\\dA-Z][^\\WA-Z]2,|_[^\\WA-Z]*|__[^\\WA-Z\\d_][^\\WA-Z]+__)$' pattern)

            self.scaler = pickle.load(f)

    def save_model(self, save_path: str):
        """
        A tool to save model to disk
        :param save_path:   Saving path.
        :return:            None.
        """
        os.makedirs(f'model_dir/{save_path}', exist_ok=True)
        saving_prep = lambda string: f'model_dir/{save_path}/{string}'
        self.clf.save_model(saving_prep('model.json'))
        self.emb.save(saving_prep('emb.pkl'))
        with open(saving_prep('scaler.pkl'), 'wb') as f:

Variable name "f" doesn't conform to snake_case naming style ('([^\\W\\dA-Z][^\\WA-Z]2,|_[^\\WA-Z]*|__[^\\WA-Z\\d_][^\\WA-Z]+__)$' pattern)

            pickle.dump(self.scaler, f, pickle.HIGHEST_PROTOCOL)

    def __getitem__(self, item: str) -> int:
        """
        getitem overwritten
        :param item:    Input text
        :return:        Predicted class (0, 1).
        """
        return self.predict(item)

    def predict(self, input_text: str) -> int:
        """
        Prediction method.
        :param input_text:  input text, string
        :return:            predicted class. (0, 1)
        """
        prep_text = remove_redundant_characters(remove_emoji(input_text))
        vector = self.scaler.transform(self.emb.encode(prep_text).reshape(1, -1))
        return self.clf.predict(vector)[0]