rendercomment.py 4.37 KB
Newer Older
1
#!/usr/bin/env python3
2

3
import re
4
import pygit2
5
6
import sys
import bleach
7
import markdown
8

9
import aurweb.config
10
11
import aurweb.db

12
13
14
repo_path = aurweb.config.get('serve', 'repo-path')
commit_uri = aurweb.config.get('options', 'commit_uri')

15

16
17
18
19
20
class LinkifyExtension(markdown.extensions.Extension):
    """
    Turn URLs into links, even without explicit markdown.
    Do not linkify URLs in code blocks.
    """
21

22
23
24
25
    # Captures http(s) and ftp URLs until the first non URL-ish character.
    # Excludes trailing punctuation.
    _urlre = (r'(\b(?:https?|ftp):\/\/[\w\/\#~:.?+=&%@!\-;,]+?'
              r'(?=[.:?\-;,]*(?:[^\w\/\#~:.?+=&%@!\-;,]|$)))')
26
27

    def extendMarkdown(self, md, md_globals):
28
29
        processor = markdown.inlinepatterns.AutolinkInlineProcessor(self._urlre, md)
        md.inlinePatterns.add('linkify', processor, '_end')
30
31


32
33
34
35
36
37
38
39
40
41
42
43
44
45
class FlysprayLinksPreprocessor(markdown.preprocessors.Preprocessor):
    _fsre = re.compile(r'\b(FS#(\d+))\b')
    _sub = r'[\1](https://bugs.archlinux.org/task/\2)'

    def run(self, lines):
        return [self._fsre.sub(self._sub, line) for line in lines]


class FlysprayLinksExtension(markdown.extensions.Extension):
    def extendMarkdown(self, md, md_globals):
        preprocessor = FlysprayLinksPreprocessor(md)
        md.preprocessors.add('flyspray-links', preprocessor, '_end')


46
47
48
49
50
51
52
53
54
class GitCommitsInlineProcessor(markdown.inlinepatterns.InlineProcessor):
    """
    Turn Git hashes like f7f5152be5ab into links to AUR's cgit.

    Only commit references that do exist are linkified. Hashes are shortened to
    shorter non-ambiguous prefixes. Only hashes with at least 7 digits are
    considered.
    """

55
56
57
58
    _repo = pygit2.Repository(repo_path)

    def __init__(self, md, head):
        self._head = head
59
        super().__init__(r'\b([0-9a-f]{7,40})\b', md)
60

61
62
    def handleMatch(self, m, data):
        oid = m.group(1)
63
        if oid not in self._repo:
64
65
            # Unkwown OID; preserve the orginal text.
            return None, None, None
66
67
68
69
70
71
72

        prefixlen = 12
        while prefixlen < 40:
            if oid[:prefixlen] in self._repo:
                break
            prefixlen += 1

73
74
75
76
        el = markdown.util.etree.Element('a')
        el.set('href', commit_uri % (self._head, oid[:prefixlen]))
        el.text = markdown.util.AtomicString(oid[:prefixlen])
        return el, m.start(0), m.end(0)
77
78
79
80
81
82
83
84
85
86


class GitCommitsExtension(markdown.extensions.Extension):
    _head = None

    def __init__(self, head):
        self._head = head
        super(markdown.extensions.Extension, self).__init__()

    def extendMarkdown(self, md, md_globals):
87
88
        processor = GitCommitsInlineProcessor(md, self._head)
        md.inlinePatterns.add('git-commits', processor, '_end')
89
90


91
92
93
94
95
96
97
98
99
100
101
102
103
104
class HeadingTreeprocessor(markdown.treeprocessors.Treeprocessor):
    def run(self, doc):
        for elem in doc:
            if elem.tag == 'h1':
                elem.tag = 'h5'
            elif elem.tag in ['h2', 'h3', 'h4', 'h5']:
                elem.tag = 'h6'


class HeadingExtension(markdown.extensions.Extension):
    def extendMarkdown(self, md, md_globals):
        md.treeprocessors.add('heading', HeadingTreeprocessor(md), '_end')


105
def get_comment(conn, commentid):
106
107
108
109
110
    cur = conn.execute('SELECT PackageComments.Comments, PackageBases.Name '
                       'FROM PackageComments INNER JOIN PackageBases '
                       'ON PackageBases.ID = PackageComments.PackageBaseID '
                       'WHERE PackageComments.ID = ?', [commentid])
    return cur.fetchone()
111
112
113
114
115
116
117
118
119
120
121
122


def save_rendered_comment(conn, commentid, html):
    conn.execute('UPDATE PackageComments SET RenderedComment = ? WHERE ID = ?',
                 [html, commentid])


def main():
    commentid = int(sys.argv[1])

    conn = aurweb.db.Connection()

123
    text, pkgbase = get_comment(conn, commentid)
124
125
    html = markdown.markdown(text, extensions=['fenced_code',
                                               LinkifyExtension(),
126
                                               FlysprayLinksExtension(),
127
128
                                               GitCommitsExtension(pkgbase),
                                               HeadingExtension()])
129
130
    allowed_tags = (bleach.sanitizer.ALLOWED_TAGS +
                    ['p', 'pre', 'h4', 'h5', 'h6', 'br', 'hr'])
131
    html = bleach.clean(html, tags=allowed_tags)
132
133
134
135
136
137
138
139
    save_rendered_comment(conn, commentid, html)

    conn.commit()
    conn.close()


if __name__ == '__main__':
    main()