aboutsummaryrefslogtreecommitdiffstats
path: root/python/fatcat_tools/normal.py
blob: 7b4bd19c6b681fb51182673f9391fcd92f592924 (plain)
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
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219

"""
A bunch of helpers to parse and normalize strings: external identifiers,
free-form input, titles, etc.
"""

import re

DOI_REGEX = re.compile("^10.\d{3,6}/\S+$")

def clean_doi(raw):
    """
    Removes any:
    - padding whitespace
    - 'doi:' prefix
    - URL prefix

    Does not try to un-URL-encode

    Returns None if not a valid DOI
    """
    if not raw:
        return None
    raw = raw.strip()
    if '\u2013' in raw:
        # Do not attempt to normalize "en dash" and since FC does not allow
        # unicode in DOI, treat this as invalid.
        return None
    if len(raw.split()) != 1:
        return None
    if raw.startswith("doi:"):
        raw = raw[4:]
    if raw.startswith("http://"):
        raw = raw[7:]
    if raw.startswith("https://"):
        raw = raw[8:]
    if raw.startswith("doi.org/"):
        raw = raw[8:]
    if raw.startswith("dx.doi.org/"):
        raw = raw[11:]
    if raw[7:9] == "//":
        raw = raw[:8] + raw[9:]
    if not raw.startswith("10."):
        return None
    if not DOI_REGEX.fullmatch(raw):
        return None
    return raw

def test_clean_doi():
    assert clean_doi("10.1234/asdf ") == "10.1234/asdf"
    assert clean_doi("10.1037//0002-9432.72.1.50") == "10.1037/0002-9432.72.1.50"
    assert clean_doi("10.1037/0002-9432.72.1.50") == "10.1037/0002-9432.72.1.50"
    assert clean_doi("10.23750/abm.v88i2 -s.6506") == None
    assert clean_doi("10.17167/mksz.2017.2.129–155") == None
    assert clean_doi("http://doi.org/10.1234/asdf ") == "10.1234/asdf"
    assert clean_doi("https://dx.doi.org/10.1234/asdf ") == "10.1234/asdf"
    assert clean_doi("doi:10.1234/asdf ") == "10.1234/asdf"
    assert clean_doi("doi:10.1234/ asdf ") == None

ARXIV_ID_REGEX = re.compile("^(\d{4}.\d{4,5}|[a-z\-]+(\.[A-Z]{2})?/\d{7})(v\d+)?$")

def clean_arxiv_id(raw):
    """
    Removes any:
    - 'arxiv:' prefix

    Works with versioned or un-versioned arxiv identifiers.
    """
    if not raw:
        return None
    raw = raw.strip()
    if raw.lower().startswith("arxiv:"):
        raw = raw[6:]
    if raw.lower().startswith("https://arxiv.org/abs/"):
        raw = raw[22:]
    if not ARXIV_ID_REGEX.fullmatch(raw):
        return None
    return raw

def test_clean_arxiv_id():
    assert clean_arxiv_id("0806.2878v1") == "0806.2878v1"
    assert clean_arxiv_id("0806.2878") == "0806.2878"
    assert clean_arxiv_id("1501.00001v1") == "1501.00001v1"
    assert clean_arxiv_id("1501.00001") == "1501.00001"
    assert clean_arxiv_id("hep-th/9901001v1") == "hep-th/9901001v1"
    assert clean_arxiv_id("hep-th/9901001") == "hep-th/9901001"
    assert clean_arxiv_id("math.CA/0611800v2") == "math.CA/0611800v2"
    assert clean_arxiv_id("math.CA/0611800") == "math.CA/0611800"
    assert clean_arxiv_id("0806.2878v1 ") == "0806.2878v1"

    assert clean_arxiv_id("https://arxiv.org/abs/0806.2878v1") == "0806.2878v1"
    assert clean_arxiv_id("arxiv:0806.2878v1") == "0806.2878v1"
    assert clean_arxiv_id("arXiv:0806.2878v1") == "0806.2878v1"

    assert clean_arxiv_id("hep-TH/9901001v1") == None
    assert clean_arxiv_id("hßp-th/9901001v1") == None
    assert clean_arxiv_id("math.CA/06l1800v2") == None
    assert clean_arxiv_id("mßth.ca/0611800v2") == None
    assert clean_arxiv_id("MATH.CA/0611800v2") == None
    assert clean_arxiv_id("0806.2878v23") == "0806.2878v23"  # ?
    assert clean_arxiv_id("0806.2878v") == None
    assert clean_arxiv_id("0806.2878") == "0806.2878"
    assert clean_arxiv_id("006.2878v1") == None
    assert clean_arxiv_id("0806.v1") == None
    assert clean_arxiv_id("08062878v1") == None

def clean_pmid(raw):
    if not raw:
        return None
    raw = raw.strip()
    if len(raw.split()) != 1:
        return None
    if raw.isdigit():
        return raw
    return None

def test_clean_pmid():
    assert clean_pmid("1234") == "1234"
    assert clean_pmid("1234 ") == "1234"
    assert clean_pmid("PMC123") == None
    assert clean_sha1("qfba3") == None
    assert clean_sha1("") == None

def clean_pmcid(raw):
    if not raw:
        return None
    raw = raw.strip()
    if len(raw.split()) != 1:
        return None
    if raw.startswith("PMC") and raw[3:] and raw[3:].isdigit():
        return raw
    return None

def clean_sha1(raw):
    if not raw:
        return None
    raw = raw.strip().lower()
    if len(raw.split()) != 1:
        return None
    if len(raw) != 40:
        return None
    for c in raw:
        if c not in "0123456789abcdef":
            return None
    return raw

def test_clean_sha1():
    assert clean_sha1("0fba3fba0e1937aa0297de3836b768b5dfb23d7b") == "0fba3fba0e1937aa0297de3836b768b5dfb23d7b"
    assert clean_sha1("0fba3fba0e1937aa0297de3836b768b5dfb23d7b ") == "0fba3fba0e1937aa0297de3836b768b5dfb23d7b"
    assert clean_sha1("fba3fba0e1937aa0297de3836b768b5dfb23d7b") == None
    assert clean_sha1("qfba3fba0e1937aa0297de3836b768b5dfb23d7b") == None
    assert clean_sha1("0fba3fb a0e1937aa0297de3836b768b5dfb23d7b") == None

def clean_sha256(raw):
    raw = raw.strip().lower()
    if len(raw.split()) != 1:
        return None
    if len(raw) != 64:
        return None
    for c in raw:
        if c not in "0123456789abcdef":
            return None
    return raw

def test_clean_sha256():
    assert clean_sha256("6cc853f2ae75696b2e45f476c76b946b0fc2df7c52bb38287cb074aceb77bc7f") == "6cc853f2ae75696b2e45f476c76b946b0fc2df7c52bb38287cb074aceb77bc7f"
    assert clean_sha256("0fba3fba0e1937aa0297de3836b768b5dfb23d7b") == None

ISSN_REGEX = re.compile("^\d{4}-\d{3}[0-9X]$")

def clean_issn(raw):
    if not raw:
        return None
    raw = raw.strip().upper()
    if len(raw) != 9:
        return None
    if not ISSN_REGEX.fullmatch(raw):
        return None
    return raw

def test_clean_issn():
    assert clean_issn("1234-4567") == "1234-4567"
    assert clean_issn("1234-456X") == "1234-456X"
    assert clean_issn("134-4567") == None
    assert clean_issn("123X-4567") == None

ISBN13_REGEX = re.compile("^97(?:8|9)-\d{1,5}-\d{1,7}-\d{1,6}-\d$")

def clean_isbn13(raw):
    if not raw:
        return None
    raw = raw.strip()
    if not ISBN13_REGEX.fullmatch(raw):
        return None
    return raw

def test_clean_isbn13():
    assert clean_isbn13("978-1-56619-909-4") == "978-1-56619-909-4"
    assert clean_isbn13("978-1-4028-9462-6") == "978-1-4028-9462-6"
    assert clean_isbn13("978-1-56619-909-4 ") == "978-1-56619-909-4"
    assert clean_isbn13("9781566199094") == None

ORCID_REGEX = re.compile("^\d{4}-\d{4}-\d{4}-\d{3}[\dX]$")

def clean_orcid(raw):
    if not raw:
        return None
    raw = raw.strip()
    if not ORCID_REGEX.fullmatch(raw):
        return None
    return raw

def test_clean_orcid():
    assert clean_orcid("0123-4567-3456-6789") == "0123-4567-3456-6789"
    assert clean_orcid("0123-4567-3456-678X") == "0123-4567-3456-678X"
    assert clean_orcid("0123-4567-3456-6789 ") == "0123-4567-3456-6789"
    assert clean_orcid("01234567-3456-6780") == None
    assert clean_orcid("0x23-4567-3456-6780") == None