|
1
|
|
|
import re |
|
2
|
|
|
|
|
3
|
|
|
import jinja2 |
|
4
|
|
|
import jinja2.ext |
|
5
|
|
|
import jinja2.lexer |
|
6
|
|
|
|
|
7
|
|
|
|
|
8
|
|
|
class HTMLCompressFeed(object): |
|
9
|
|
|
re_whitespace = re.compile('(\"[^\"\\n]*\"|\'[^\'\\n]*\')|([ \\t\\r\\n]+)') |
|
10
|
|
|
token_class = jinja2.lexer.Token |
|
11
|
|
|
block_tokens = { |
|
12
|
|
|
'variable_begin': 'variable_end', |
|
13
|
|
|
'block_begin': 'block_end' |
|
14
|
|
|
} |
|
15
|
|
|
ignore_elements = ['textarea', 'pre', 'script', 'style'] |
|
16
|
|
|
|
|
17
|
|
|
def __init__(self): |
|
18
|
|
|
self.intag = False |
|
19
|
|
|
self.pending = '' |
|
20
|
|
|
self.lineno = 0 |
|
21
|
|
|
self.skip_until = None |
|
22
|
|
|
self.ignore_until = None |
|
23
|
|
|
|
|
24
|
|
|
def finalize(self, strip=False): |
|
25
|
|
|
if self.ignore_until: |
|
26
|
|
|
data = self.pending |
|
27
|
|
|
elif self.intag: |
|
28
|
|
|
data = self._collapse(self.pending) |
|
29
|
|
|
else: |
|
30
|
|
|
data = self.pending.rstrip() if strip else self.pending |
|
31
|
|
|
if data.strip(): |
|
32
|
|
|
yield self.token_class(self.lineno, 'data', data) |
|
33
|
|
|
self.pending = '' |
|
34
|
|
|
|
|
35
|
|
|
def _collapse(self, data): |
|
36
|
|
|
last = self.re_whitespace.groups |
|
37
|
|
|
return self.re_whitespace.sub( |
|
38
|
|
|
lambda m: ' ' if m.group(last) else m.group(0), |
|
39
|
|
|
data |
|
40
|
|
|
) |
|
41
|
|
|
|
|
42
|
|
|
def _process(self, value, lineno): |
|
43
|
|
|
if self.ignore_until: |
|
44
|
|
|
s, p, value = value.partition(self.ignore_until) |
|
45
|
|
|
if p: |
|
46
|
|
|
self.intag = False |
|
47
|
|
|
self.ignore_until = None |
|
48
|
|
|
s = s + p |
|
49
|
|
|
yield self.token_class(lineno, 'data', s), value |
|
50
|
|
|
elif self.intag: |
|
51
|
|
|
s, p, value = value.partition('>') |
|
52
|
|
|
s = self._collapse(s) |
|
53
|
|
|
if p: |
|
54
|
|
|
self.intag = False |
|
55
|
|
|
s = s.rstrip() + p |
|
56
|
|
|
yield self.token_class(lineno, 'data', s), value |
|
57
|
|
|
else: |
|
58
|
|
|
s, p, value = value.partition('<') |
|
59
|
|
|
if p: |
|
60
|
|
|
self.intag = True |
|
61
|
|
|
s = s + p if s.strip() else p |
|
62
|
|
|
yield self.token_class(lineno, 'data', s), value |
|
63
|
|
|
for elm in self.ignore_elements: |
|
64
|
|
|
if value.startswith(elm): |
|
65
|
|
|
self.ignore_until = '</{}>'.format(elm) |
|
66
|
|
|
break |
|
67
|
|
|
|
|
68
|
|
|
def feed(self, token): |
|
69
|
|
|
if self.skip_until: |
|
70
|
|
|
yield token |
|
71
|
|
|
if token.type == self.skip_until: |
|
72
|
|
|
self.skip_until = None |
|
73
|
|
|
return |
|
74
|
|
|
|
|
75
|
|
|
if token.type in self.block_tokens: |
|
76
|
|
|
for data in self.finalize(token.type == 'block_begin'): |
|
77
|
|
|
yield data |
|
78
|
|
|
yield token |
|
79
|
|
|
self.skip_until = self.block_tokens[token.type] |
|
80
|
|
|
return |
|
81
|
|
|
|
|
82
|
|
|
lineno = self.lineno |
|
83
|
|
|
size = len(token.value) |
|
84
|
|
|
value = self.pending + token.value |
|
85
|
|
|
loop = True |
|
86
|
|
|
while loop: |
|
87
|
|
|
loop = False |
|
88
|
|
|
lineno = self.lineno if len(value) > size else token.lineno |
|
89
|
|
|
for data, value in self._process(value, lineno): |
|
90
|
|
|
loop = value |
|
91
|
|
|
yield data |
|
92
|
|
|
self.lineno = lineno |
|
93
|
|
|
self.pending = value |
|
94
|
|
|
|
|
95
|
|
|
|
|
96
|
|
|
class HTMLCompress(jinja2.ext.Extension): |
|
97
|
|
|
def filter_stream(self, stream): |
|
98
|
|
|
feed = HTMLCompressFeed() |
|
99
|
|
|
for token in stream: |
|
100
|
|
|
for data in feed.feed(token): |
|
101
|
|
|
yield data |
|
102
|
|
|
for data in feed.finalize(): |
|
103
|
|
|
yield data |
|
104
|
|
|
|
ergoithz /
browsepy
