1 #!/usr/bin/env python3
2 """
3 Utility for parsing HTML5 entity definitions available from:
4
5 https://html.spec.whatwg.org/entities.json
6 https://html.spec.whatwg.org/multipage/named-characters.html
7
8 The page now contains the following note:
9
10 "This list is static and will not be expanded or changed in the future."
11
12 Written by Ezio Melotti and Iuliia Proskurnia.
13 """
14
15 import os
16 import sys
17 import json
18 from urllib.request import urlopen
19 from html.entities import html5
20
21 SCRIPT_NAME = 'Tools/build/parse_html5_entities.py'
22 PAGE_URL = 'https://html.spec.whatwg.org/multipage/named-characters.html'
23 ENTITIES_URL = 'https://html.spec.whatwg.org/entities.json'
24 HTML5_SECTION_START = '# HTML5 named character references'
25
26 def get_json(url):
27 """Download the json file from the url and returns a decoded object."""
28 with urlopen(url) as f:
29 data = f.read().decode('utf-8')
30 return json.loads(data)
31
32 def create_dict(entities):
33 """Create the html5 dict from the decoded json object."""
34 new_html5 = {}
35 for name, value in entities.items():
36 new_html5[name.lstrip('&')] = value['characters']
37 return new_html5
38
39 def compare_dicts(old, new):
40 """Compare the old and new dicts and print the differences."""
41 added = new.keys() - old.keys()
42 if added:
43 print('{} entitie(s) have been added:'.format(len(added)))
44 for name in sorted(added):
45 print(' {!r}: {!r}'.format(name, new[name]))
46 removed = old.keys() - new.keys()
47 if removed:
48 print('{} entitie(s) have been removed:'.format(len(removed)))
49 for name in sorted(removed):
50 print(' {!r}: {!r}'.format(name, old[name]))
51 changed = set()
52 for name in (old.keys() & new.keys()):
53 if old[name] != new[name]:
54 changed.add((name, old[name], new[name]))
55 if changed:
56 print('{} entitie(s) have been modified:'.format(len(changed)))
57 for item in sorted(changed):
58 print(' {!r}: {!r} -> {!r}'.format(*item))
59
60 def write_items(entities, file=sys.stdout):
61 """Write the items of the dictionary in the specified file."""
62 # The keys in the generated dictionary should be sorted
63 # in a case-insensitive way, however, when two keys are equal,
64 # the uppercase version should come first so that the result
65 # looks like: ['Aacute', 'aacute', 'Aacute;', 'aacute;', ...]
66 # To do this we first sort in a case-sensitive way (so all the
67 # uppercase chars come first) and then sort with key=str.lower.
68 # Since the sorting is stable the uppercase keys will eventually
69 # be before their equivalent lowercase version.
70 keys = sorted(entities.keys())
71 keys = sorted(keys, key=str.lower)
72 print(HTML5_SECTION_START, file=file)
73 print(f'# Generated by {SCRIPT_NAME}\n'
74 f'# from {ENTITIES_URL} and\n'
75 f'# {PAGE_URL}.\n'
76 f'# Map HTML5 named character references to the '
77 f'equivalent Unicode character(s).', file=file)
78 print('html5 = {', file=file)
79 for name in keys:
80 print(f' {name!r}: {entities[name]!a},', file=file)
81 print('}', file=file)
82
83
84 if __name__ == '__main__':
85 # without args print a diff between html.entities.html5 and new_html5
86 # with --create print the new html5 dict
87 # with --patch patch the Lib/html/entities.py file
88 new_html5 = create_dict(get_json(ENTITIES_URL))
89 if '--create' in sys.argv:
90 write_items(new_html5)
91 elif '--patch' in sys.argv:
92 fname = 'Lib/html/entities.py'
93 temp_fname = fname + '.temp'
94 with open(fname) as f1, open(temp_fname, 'w') as f2:
95 skip = False
96 for line in f1:
97 if line.startswith(HTML5_SECTION_START):
98 write_items(new_html5, file=f2)
99 skip = True
100 continue
101 if skip:
102 # skip the old items until the }
103 if line.startswith('}'):
104 skip = False
105 continue
106 f2.write(line)
107 os.remove(fname)
108 os.rename(temp_fname, fname)
109 else:
110 if html5 == new_html5:
111 print('The current dictionary is updated.')
112 else:
113 compare_dicts(html5, new_html5)
114 print('Run "./python {0} --patch" to update Lib/html/entities.html '
115 'or "./python {0} --create" to see the generated ' 'dictionary.'.format(__file__))