|
1
|
|
|
import re |
|
2
|
|
|
|
|
3
|
|
|
import jinja2 |
|
4
|
|
|
import jinja2.ext |
|
5
|
|
|
import jinja2.lexer |
|
6
|
|
|
|
|
7
|
|
|
from . import StateMachine |
|
8
|
|
|
|
|
9
|
|
|
|
|
10
|
|
|
class SGMLCompressContext(StateMachine): |
|
11
|
|
|
re_whitespace = re.compile('[ \\t\\r\\n]+') |
|
12
|
|
|
block_tags = {} # block content will be treated as literal text |
|
13
|
|
|
jumps = { # state machine jumps |
|
14
|
|
|
'text': { |
|
15
|
|
|
'<': 'tag', |
|
16
|
|
|
'<!--': 'comment', |
|
17
|
|
|
'<![CDATA[': 'cdata', |
|
18
|
|
|
}, |
|
19
|
|
|
'lit1': {'"': 'tag'}, |
|
20
|
|
|
'lit2': ("'", 'tag'), |
|
21
|
|
|
'tag': { |
|
22
|
|
|
'>': 'text', |
|
23
|
|
|
'"': 'lit1', |
|
24
|
|
|
"'": 'lit2' |
|
25
|
|
|
}, |
|
26
|
|
|
'comment': {'-->': 'text'}, |
|
27
|
|
|
'cdata': {']]>': 'text'} |
|
28
|
|
|
} |
|
29
|
|
|
skip_until_text = None # inside text until this is met |
|
30
|
|
|
current = 'text' |
|
31
|
|
|
|
|
32
|
|
|
@property |
|
33
|
|
|
def nearest(self): |
|
34
|
|
|
if self.skip_until_text and self.current == 'text': |
|
35
|
|
|
mark = self.skip_until_text |
|
36
|
|
|
index = self.pending.find(mark, len(self.start)) |
|
37
|
|
|
if index == -1: |
|
38
|
|
|
return len(self.pending), '', None |
|
39
|
|
|
return index, mark, self.current |
|
40
|
|
|
return super(SGMLCompressContext, self).nearest |
|
41
|
|
|
|
|
42
|
|
|
def transform_tag(self, data, mark, next): |
|
43
|
|
|
tagstart = self.start == '<' |
|
44
|
|
|
data = self.re_whitespace.sub(' ', data[1:] if tagstart else data) |
|
45
|
|
|
if tagstart: |
|
46
|
|
|
data = data.lstrip() if next is None else data.strip() |
|
47
|
|
|
tagname = data.split(' ', 1)[0] |
|
48
|
|
|
self.skip_until_text = self.block_tags.get(tagname) |
|
49
|
|
|
return '<' + data |
|
50
|
|
|
elif next is None: |
|
51
|
|
|
return data.rstrip() |
|
52
|
|
|
return self.start if data.strip() == self.start else data |
|
53
|
|
|
|
|
54
|
|
|
def transform_text(self, data, mark, next): |
|
55
|
|
|
if not self.skip_until_text: |
|
56
|
|
|
return self.start if data.strip() == self.start else data |
|
57
|
|
|
elif next is not None: |
|
58
|
|
|
self.skip_until_text = None |
|
59
|
|
|
return data |
|
60
|
|
|
|
|
61
|
|
|
|
|
62
|
|
|
class HTMLCompressContext(SGMLCompressContext): |
|
63
|
|
|
block_tags = { |
|
64
|
|
|
'textarea': '</textarea>', |
|
65
|
|
|
'pre': '</pre>', |
|
66
|
|
|
'script': '</script>', |
|
67
|
|
|
'style': '</style>', |
|
68
|
|
|
} |
|
69
|
|
|
|
|
70
|
|
|
|
|
71
|
|
|
class HTMLCompress(jinja2.ext.Extension): |
|
72
|
|
|
context_class = HTMLCompressContext |
|
73
|
|
|
token_class = jinja2.lexer.Token |
|
74
|
|
|
block_tokens = { |
|
75
|
|
|
'variable_begin': 'variable_end', |
|
76
|
|
|
'block_begin': 'block_end' |
|
77
|
|
|
} |
|
78
|
|
|
|
|
79
|
|
|
def filter_stream(self, stream): |
|
80
|
|
|
transform = self.context_class() |
|
81
|
|
|
lineno = 0 |
|
82
|
|
|
skip_until_token = None |
|
83
|
|
|
for token in stream: |
|
84
|
|
|
if skip_until_token: |
|
85
|
|
|
yield token |
|
86
|
|
|
if token.type == skip_until_token: |
|
87
|
|
|
skip_until_token = None |
|
88
|
|
|
continue |
|
89
|
|
|
|
|
90
|
|
|
if token.type != 'data': |
|
91
|
|
|
for data in transform.finish(): |
|
92
|
|
|
yield self.token_class(lineno, 'data', data) |
|
93
|
|
|
yield token |
|
94
|
|
|
skip_until_token = self.block_tokens.get(token.type) |
|
95
|
|
|
continue |
|
96
|
|
|
|
|
97
|
|
|
if not transform.pending: |
|
98
|
|
|
lineno = token.lineno |
|
99
|
|
|
|
|
100
|
|
|
for data in transform.feed(token.value): |
|
101
|
|
|
yield self.token_class(lineno, 'data', data) |
|
102
|
|
|
lineno = token.lineno |
|
103
|
|
|
|
|
104
|
|
|
for data in transform.finish(): |
|
105
|
|
|
yield self.token_class(lineno, 'data', data) |
|
106
|
|
|
|
ergoithz /
browsepy
