breno-madruga
/
dib-covid-dataset
mirror of https://github.com/breno-madruga/dib-covid-dataset.git


  
1

	
2

	
3

	
4

	
5

	
6

	
7

	
8

	
9

	
10

	
11

	
12

	
13

	
14

	
15

	
16

	
17

	
18

	
19

	
20

	
21

	
22

	
23

	
24

	
25

	
26

	
27

	
28

	
29

	
30

	
31

	
32

	
33

	
34

	
35

	
36

	
37

	
38

	
39

	
40

	
41

	
42

	
43

	
44

	
45

	
46

	
47

	
48

	
49

	
50

	
51

	
52

	
53

	
54

	
55

	
56

	
57

	
58

	
59

	
60

	
61

	
62

	
63

	
64

	
65

	
66

	
67

	
68

	
69

	
70

	
71

	
72

	
73

	
74

	
75

	
76

	
77

	
78

	
79

	
80

	
81

	
82

	
83

	
84

	
85

	
86

	
87

	
88

	
89

	
90

	
91

	
92

	
93

	
94

	
95

	
96

	
97

	
98

	
99

	
100

	
101

	
102

	
103

	
104

	
105

	
106

	
107

	
108

	
109

	
110

	
111

	
112

	
113

	
114

	
115

	
116

	
117

	
118

	
119

	
120

	
121

	
122

	
123

	
124

	
125

	
126

	
127

	
128

	
129

	
130

	
131

	
132

	
133

	
134

	
135

	
136

	
137

	
138

	
139

	
140

	
141

	
142

	
143

	
144

	
145

	
146

	
147

	
148

	
149

	
150

	
151

	
152

	
153

	
154

	
155

	
156

	
157

	
158

	
159

	
160

	
161

	
162

	
163

	
164

	
165

	
166

	
167

	
168

	
169

	
170

	
171

	
172

	
173

	
174

	
175

	
176

	
177

	
178

	
179

	
180

	
181

	
182

	
183

	
184

	
185

	
186

	
187

	
188

	
189

	
190

	
191

	
192

	
193

	
194

	
195

	
196

	
197

	
            ##############################################################################################
############ Cleaning and Preprocessing the Scopus publications related to COVID-19 ##########
##############################################################################################

# For collecting the Scopus publications related to COVID-19, we used the "pybliometrics"
# library. It is avaliable on https://pypi.org/project/pybliometrics/.

########################################################################
# Uncomment to install the library.
# %pip install pylatexenc
########################################################################

########################################################################
# Importing the required libraries.
import re, pandas as pd, numpy as np
from pylatexenc.latex2text import LatexNodes2Text
from preprocess import Preprocess
########################################################################

class ProcessScopus(Preprocess):

    # Defining the function "clean_text" to clean and preprocess any text.
    @staticmethod
    def __clean_text(text, has_latex=False):
        if text:
            text = re.sub(r"\u2fff(s|\s)", r"'\1", re.sub(r"\s+", " ", re.sub(r"\ufeff\.?", "",
                re.sub(r"\\\\(\’\s)?", "", str(text))))).replace("\u200b", "").replace(
                "\ue001", "").replace("\ue061", "").replace("\u202f", "").replace(
                "\u2060", "").replace("\u200f", "").replace("\u200e", "").replace(
                "\u202c", "").replace("&#x2013;", "-").replace("&quot", "\"\"").replace(
                "\u200c", "").replace("\\u0019", "").replace("\\s", "s").replace(
                "\u202a", "").replace("\u202d", "-").replace("\u0383", "-").replace(
                "\u20f3", "ó").replace("\u20fa", "ú").replace("\u2fff", "-").strip()
            text = text.replace("TNF-alpha induced", "TNF-α induced").replace(
                "TNF-Alpha induced", "TNF-α induced").replace(
                "TNF- ␣ induced", "TNF-α induced").replace("TNF-αinduced", "TNF-α induced").replace(
                "via NF- \u242c B pathway", "via NF-κB pathway").replace(
                "via NF-kappaB pathway", "via NF-κB pathway").strip()
            if has_latex:
                text = LatexNodes2Text().latex_to_text(re.sub("\\?%", "@PER@CENT@", text)).replace(
                    "@PER@CENT@", "%")
            text = re.sub(r"\s+", " ", re.sub(r"\-{2,}", "-",
                re.sub(r"\s?\xad(\s|\-)?", "-", text))).replace("\\", "").replace(
                "\\%", "%").replace("()", "").replace("[]", "").strip()
            return text
        else:
            return None

    # Cleaning and preprocessing the dataframe.
    def _preprocess(self):
        # Removing the invalid articles.
        self._dataframe = self._dataframe.loc[
            self._dataframe.id.notnull() & self._dataframe.eid.notnull()]

        # Defining the "None" value for the "NaN" values.
        self._dataframe.replace({np.nan: None}, inplace=True)

        # Defining the "zero" value for the articles without numbers of citation and references.
        self._dataframe.citation_num.loc[self._dataframe.citation_num.isnull()] = 0
        self._dataframe.ref_count.loc[self._dataframe.ref_count.isnull()] = 0

        # Normalizing the feature "abstract".
        self._dataframe.abstract.loc[
            self._dataframe.abstract.isnull() & self._dataframe.description.notnull()
        ] = self._dataframe.description.loc[
            self._dataframe.abstract.isnull() & self._dataframe.description.notnull()]
        self._dataframe.abstract.loc[self._dataframe.abstract.notnull()] = \
            self._dataframe.abstract.loc[self._dataframe.abstract.notnull()].apply(
                lambda x: ProcessScopus.__clean_text(x, True))

        # Normalizing the feature "vehicle_name".
        self._dataframe.vehicle_name.loc[
            self._dataframe.conference_name.notnull() & self._dataframe.vehicle_name.notnull()
        ] = self._dataframe.conference_name.loc[
            self._dataframe.conference_name.notnull() & self._dataframe.vehicle_name.notnull()]
        self._dataframe.vehicle_name.loc[self._dataframe.vehicle_name.notnull()] = \
            self._dataframe.vehicle_name.loc[self._dataframe.vehicle_name.notnull()].apply(
                ProcessScopus.__clean_text)

        # Normalizing the feature "title".
        self._dataframe.title.loc[self._dataframe.title.notnull()] = self._dataframe.title.loc[
            self._dataframe.title.notnull()].apply(ProcessScopus.__clean_text)

        # Removing unnecessary columns.
        columns_drop = ["eid", "pii", "description", "isbn", "conf_location", "conference_name",
            "vehicle_address", "title_edition"]
        self._dataframe.drop(axis=1, columns=columns_drop, inplace=True)

        # Changing the type of some features.
        self._dataframe.loc[:, ["citation_num", "ref_count"]] = self._dataframe.loc[:,
            ["citation_num", "ref_count"]].astype(np.float32)
        self._dataframe.auth_keywords.loc[self._dataframe.auth_keywords.notnull()] = \
            self._dataframe.auth_keywords.loc[self._dataframe.auth_keywords.notnull()].apply(eval)
        self._dataframe.index_terms.loc[self._dataframe.index_terms.notnull()] = \
            self._dataframe.index_terms.loc[self._dataframe.index_terms.notnull()].apply(eval)
        self._dataframe.affiliations.loc[self._dataframe.affiliations.notnull()] = \
            self._dataframe.affiliations.loc[self._dataframe.affiliations.notnull()].apply(eval)
        self._dataframe.subject_areas.loc[self._dataframe.subject_areas.notnull()] = \
            self._dataframe.subject_areas.loc[self._dataframe.subject_areas.notnull()].apply(eval)
        self._dataframe.authors.loc[self._dataframe.authors.notnull()] = \
            self._dataframe.authors.loc[self._dataframe.authors.notnull()].apply(eval)
        self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()] = \
            self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()].apply(eval)
        self._dataframe.references.loc[self._dataframe.references.notnull()] = \
            self._dataframe.references.loc[self._dataframe.references.notnull()].apply(eval)
        self._dataframe.publication_date = pd.to_datetime(self._dataframe.publication_date)

        # Creating the feature "period" from the feature "publication_date".
        if "period" not in self._dataframe:
            self._dataframe["period"] = self._dataframe.publication_date.apply(
                lambda x: "{}-{}".format(x.year, x.month))

        # Normalizing the itens contained in the features "auth_keywords" and "index_terms".
        self._dataframe.auth_keywords.loc[self._dataframe.auth_keywords.notnull()] = \
            self._dataframe.auth_keywords.loc[self._dataframe.auth_keywords.notnull()].apply(
                lambda x: tuple([ProcessScopus.__clean_text(item) for item in x]))
        self._dataframe.index_terms.loc[self._dataframe.index_terms.notnull()] = \
            self._dataframe.index_terms.loc[self._dataframe.index_terms.notnull()].apply(
                lambda x: tuple([ProcessScopus.__clean_text(item) for item in x]))

        # Checking there are invalid values in the features "auth_keywords", "index_terms" and "subject_areas".
        for column in ["auth_keywords", "index_terms", "subject_areas"]:
            count = self._dataframe.loc[self._dataframe[column].notnull(), column][
                    [np.any([item == None or item.lower() == "none" for item in items])
                    for items in self._dataframe.loc[self._dataframe[column].notnull(), column]]].size
            print("{}: {}".format(column, count))

        # Removing the invalid values in the features "auth_keywords", "index_terms" and "subject_areas".
        for column in ["auth_keywords", "index_terms", "subject_areas"]:
            self._dataframe.loc[self._dataframe[column].notnull(), column] = [
                tuple([item for item in items if item])
                for items in self._dataframe.loc[self._dataframe[column].notnull(), column]]
            self._dataframe.loc[self._dataframe[column].notnull(), column] = self._dataframe.loc[
                self._dataframe[column].notnull(), column].apply(lambda x: x if len(x) > 0 else None)

        # Normalizing the content contained in the features "authors", "affiliations" and "author_affil".
        self._dataframe.affiliations.loc[self._dataframe.affiliations.notnull()] = \
            self._dataframe.affiliations.loc[self._dataframe.affiliations.notnull()].apply(
                lambda x: tuple([{"id": item["id"],
                    "affiliation": ProcessScopus.__clean_text(item["affiliation"]),
                    "country": item["country"]} for item in x if item["id"]]))
        self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()] = \
            self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()].apply(
                lambda x: tuple([{"id": item["id"],
                    "name": ProcessScopus.__clean_text(item["name"]), "affil_id": item["affil_id"],
                    "affiliation": ProcessScopus.__clean_text(item["affiliation"]),
                    "country": item["country"]} for item in x if item["id"] or item["name"] or \
                        item["affil_id"] or item["affiliation"] or item["country"]]))
        self._dataframe.authors.loc[self._dataframe.authors.notnull()] = \
            self._dataframe.authors.loc[self._dataframe.authors.notnull()].apply(
                lambda x: tuple([{"id": item["id"],
                    "name": ProcessScopus.__clean_text(item["name"])} for item in x if item["id"]]))

        # Removing the invalid values in the features "authors", "affiliations" and "author_affil".
        for column in ["authors", "affiliations", "author_affil"]:
            self._dataframe.loc[self._dataframe[column].notnull(), column] = self._dataframe.loc[
                self._dataframe[column].notnull(), column].apply(lambda x: x if len(x) > 0 else None)

        # Creating the affiliations' and authors' IDs for those that have not a ID.
        self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()] = \
            self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()].apply(
                lambda x: tuple([{"id": item["id"] if item["id"] and item["name"] else \
                    str(hash("{} - {}".format(item["name"], "Scopus"))) if item["name"] else None,
                    "name": item["name"],
                    "affil_id": item["affil_id"] if item["affil_id"] and item["affiliation"] else \
                        str(hash("{} - {}".format(item["affiliation"], "Scopus"))) \
                            if item["affiliation"] else None,
                    "affiliation": item["affiliation"], "country": item["country"]}
                    for item in x]))

        # Removing duplicates within the list of affiliations and authors.
        self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()] = [
            set([(au["id"], au["name"], au["affil_id"],
                au["affiliation"], au["country"]) for au in row])
            for row in self._dataframe.author_affil[self._dataframe.author_affil.notnull()]]
        self._dataframe.author_affil.loc[self._dataframe.author_affil.notnull()] = [tuple([dict(zip(
                ["id", "name", "affil_id", "affiliation", "country"], au)) for au in row])
            for row in self._dataframe.author_affil[self._dataframe.author_affil.notnull()]]

        # Removing the duplicated records by feature "id".
        self._dataframe = self._dataframe.sort_values(by=["id", "period"]).drop_duplicates(
            "id", keep="first")

        # Removing the duplicated records by features "title" and "doi".
        self._dataframe = pd.concat([
            self._dataframe[self._dataframe.title.isnull() | self._dataframe.doi.isnull()],
            self._dataframe[self._dataframe.title.notnull() & self._dataframe.doi.notnull()].sort_values(
                by=["title", "citation_num", "publication_date"]).drop_duplicates(
                    ["title", "doi"], "last")], ignore_index=True)

        # Normalizing the feature "references".
        self._dataframe.references.loc[self._dataframe.references.notnull()] = \
            self._dataframe.references.loc[self._dataframe.references.notnull()].apply(
                lambda x: tuple([
                    {"id": ref["id"], "title": ProcessScopus.__clean_text(ref["title"], True),
                    "doi": ProcessScopus.__clean_text(ref["doi"]),
                    "authors": ProcessScopus.__clean_text(ref["authors"], True)} for ref in x]))