Show More
@@ -1,74 +1,79 b'' | |||
|
1 | 1 | # System library imports |
|
2 | 2 | from pygments.token import Token, is_token_subtype |
|
3 | 3 | |
|
4 | 4 | |
|
5 | 5 | class CompletionLexer(object): |
|
6 | 6 | """ Uses Pygments and some auxillary information to lex code snippets for |
|
7 | 7 | symbol contexts. |
|
8 | 8 | """ |
|
9 | 9 | |
|
10 | 10 | # Maps Lexer names to a list of possible name separators |
|
11 | 11 | separator_map = { 'C' : [ '.', '->' ], |
|
12 | 12 | 'C++' : [ '.', '->', '::' ], |
|
13 | 13 | 'Python' : [ '.' ] } |
|
14 | 14 | |
|
15 | 15 | def __init__(self, lexer): |
|
16 | 16 | """ Create a CompletionLexer using the specified Pygments lexer. |
|
17 | 17 | """ |
|
18 | 18 | self.lexer = lexer |
|
19 | 19 | |
|
20 | 20 | def get_context(self, string): |
|
21 | 21 | """ Assuming the cursor is at the end of the specified string, get the |
|
22 | 22 | context (a list of names) for the symbol at cursor position. |
|
23 | 23 | """ |
|
24 | 24 | context = [] |
|
25 | 25 | reversed_tokens = list(self._lexer.get_tokens(string)) |
|
26 | 26 | reversed_tokens.reverse() |
|
27 | 27 | |
|
28 | 28 | # Pygments often tacks on a newline when none is specified in the input. |
|
29 | 29 | # Remove this newline. |
|
30 | 30 | if reversed_tokens and reversed_tokens[0][1].endswith('\n') and \ |
|
31 | 31 | not string.endswith('\n'): |
|
32 | 32 | reversed_tokens.pop(0) |
|
33 | 33 | |
|
34 | 34 | current_op = '' |
|
35 | 35 | for token, text in reversed_tokens: |
|
36 | 36 | |
|
37 | 37 | if is_token_subtype(token, Token.Name): |
|
38 | 38 | |
|
39 | 39 | # Handle a trailing separator, e.g 'foo.bar.' |
|
40 | 40 | if current_op in self._name_separators: |
|
41 | 41 | if not context: |
|
42 | 42 | context.insert(0, '') |
|
43 | 43 | |
|
44 | 44 | # Handle non-separator operators and punction. |
|
45 | 45 | elif current_op: |
|
46 | 46 | break |
|
47 | 47 | |
|
48 | 48 | context.insert(0, text) |
|
49 | 49 | current_op = '' |
|
50 | 50 | |
|
51 | 51 | # Pygments doesn't understand that, e.g., '->' is a single operator |
|
52 | 52 | # in C++. This is why we have to build up an operator from |
|
53 | 53 | # potentially several tokens. |
|
54 | 54 | elif token is Token.Operator or token is Token.Punctuation: |
|
55 | current_op = text + current_op | |
|
55 | # Handle a trailing separator, e.g 'foo.bar.' | |
|
56 | if current_op in self._name_separators: | |
|
57 | if not context: | |
|
58 | context.insert(0, '') | |
|
59 | else: | |
|
60 | current_op = text + current_op | |
|
56 | 61 | |
|
57 | 62 | # Break on anything that is not a Operator, Punctuation, or Name. |
|
58 | 63 | else: |
|
59 | 64 | break |
|
60 | 65 | |
|
61 | 66 | return context |
|
62 | 67 | |
|
63 | 68 | def get_lexer(self, lexer): |
|
64 | 69 | return self._lexer |
|
65 | 70 | |
|
66 | 71 | def set_lexer(self, lexer, name_separators=None): |
|
67 | 72 | self._lexer = lexer |
|
68 | 73 | if name_separators is None: |
|
69 | 74 | self._name_separators = self.separator_map.get(lexer.name, ['.']) |
|
70 | 75 | else: |
|
71 | 76 | self._name_separators = list(name_separators) |
|
72 | 77 | |
|
73 | 78 | lexer = property(get_lexer, set_lexer) |
|
74 | 79 |
General Comments 0
You need to be logged in to leave comments.
Login now