Show More
@@ -1,650 +1,652 b'' | |||||
1 | """Input transformer machinery to support IPython special syntax. |
|
1 | """Input transformer machinery to support IPython special syntax. | |
2 |
|
2 | |||
3 | This includes the machinery to recognise and transform ``%magic`` commands, |
|
3 | This includes the machinery to recognise and transform ``%magic`` commands, | |
4 | ``!system`` commands, ``help?`` querying, prompt stripping, and so forth. |
|
4 | ``!system`` commands, ``help?`` querying, prompt stripping, and so forth. | |
5 |
|
5 | |||
6 | Added: IPython 7.0. Replaces inputsplitter and inputtransformer which were |
|
6 | Added: IPython 7.0. Replaces inputsplitter and inputtransformer which were | |
7 | deprecated in 7.0. |
|
7 | deprecated in 7.0. | |
8 | """ |
|
8 | """ | |
9 |
|
9 | |||
10 | # Copyright (c) IPython Development Team. |
|
10 | # Copyright (c) IPython Development Team. | |
11 | # Distributed under the terms of the Modified BSD License. |
|
11 | # Distributed under the terms of the Modified BSD License. | |
12 |
|
12 | |||
13 | from codeop import compile_command |
|
13 | from codeop import compile_command | |
14 | import re |
|
14 | import re | |
15 | import tokenize |
|
15 | import tokenize | |
16 | from typing import List, Tuple |
|
16 | from typing import List, Tuple | |
17 | import warnings |
|
17 | import warnings | |
18 |
|
18 | |||
19 | _indent_re = re.compile(r'^[ \t]+') |
|
19 | _indent_re = re.compile(r'^[ \t]+') | |
20 |
|
20 | |||
21 | def leading_indent(lines): |
|
21 | def leading_indent(lines): | |
22 | """Remove leading indentation. |
|
22 | """Remove leading indentation. | |
23 |
|
23 | |||
24 | If the first line starts with a spaces or tabs, the same whitespace will be |
|
24 | If the first line starts with a spaces or tabs, the same whitespace will be | |
25 | removed from each following line in the cell. |
|
25 | removed from each following line in the cell. | |
26 | """ |
|
26 | """ | |
27 |
if not lines: |
|
27 | if not lines: | |
28 | return lines |
|
28 | return lines | |
29 | m = _indent_re.match(lines[0]) |
|
29 | m = _indent_re.match(lines[0]) | |
30 | if not m: |
|
30 | if not m: | |
31 | return lines |
|
31 | return lines | |
32 | space = m.group(0) |
|
32 | space = m.group(0) | |
33 | n = len(space) |
|
33 | n = len(space) | |
34 | return [l[n:] if l.startswith(space) else l |
|
34 | return [l[n:] if l.startswith(space) else l | |
35 | for l in lines] |
|
35 | for l in lines] | |
36 |
|
36 | |||
37 | class PromptStripper: |
|
37 | class PromptStripper: | |
38 | """Remove matching input prompts from a block of input. |
|
38 | """Remove matching input prompts from a block of input. | |
39 |
|
39 | |||
40 | Parameters |
|
40 | Parameters | |
41 | ---------- |
|
41 | ---------- | |
42 | prompt_re : regular expression |
|
42 | prompt_re : regular expression | |
43 | A regular expression matching any input prompt (including continuation, |
|
43 | A regular expression matching any input prompt (including continuation, | |
44 | e.g. ``...``) |
|
44 | e.g. ``...``) | |
45 | initial_re : regular expression, optional |
|
45 | initial_re : regular expression, optional | |
46 | A regular expression matching only the initial prompt, but not continuation. |
|
46 | A regular expression matching only the initial prompt, but not continuation. | |
47 | If no initial expression is given, prompt_re will be used everywhere. |
|
47 | If no initial expression is given, prompt_re will be used everywhere. | |
48 | Used mainly for plain Python prompts (``>>>``), where the continuation prompt |
|
48 | Used mainly for plain Python prompts (``>>>``), where the continuation prompt | |
49 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. |
|
49 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. | |
50 |
|
50 | |||
51 | If initial_re and prompt_re differ, |
|
51 | If initial_re and prompt_re differ, | |
52 | only initial_re will be tested against the first line. |
|
52 | only initial_re will be tested against the first line. | |
53 | If any prompt is found on the first two lines, |
|
53 | If any prompt is found on the first two lines, | |
54 | prompts will be stripped from the rest of the block. |
|
54 | prompts will be stripped from the rest of the block. | |
55 | """ |
|
55 | """ | |
56 | def __init__(self, prompt_re, initial_re=None): |
|
56 | def __init__(self, prompt_re, initial_re=None): | |
57 | self.prompt_re = prompt_re |
|
57 | self.prompt_re = prompt_re | |
58 | self.initial_re = initial_re or prompt_re |
|
58 | self.initial_re = initial_re or prompt_re | |
59 |
|
59 | |||
60 | def _strip(self, lines): |
|
60 | def _strip(self, lines): | |
61 | return [self.prompt_re.sub('', l, count=1) for l in lines] |
|
61 | return [self.prompt_re.sub('', l, count=1) for l in lines] | |
62 |
|
62 | |||
63 | def __call__(self, lines): |
|
63 | def __call__(self, lines): | |
|
64 | if not lines: | |||
|
65 | return lines | |||
64 | if self.initial_re.match(lines[0]) or \ |
|
66 | if self.initial_re.match(lines[0]) or \ | |
65 | (len(lines) > 1 and self.prompt_re.match(lines[1])): |
|
67 | (len(lines) > 1 and self.prompt_re.match(lines[1])): | |
66 | return self._strip(lines) |
|
68 | return self._strip(lines) | |
67 | return lines |
|
69 | return lines | |
68 |
|
70 | |||
69 | classic_prompt = PromptStripper( |
|
71 | classic_prompt = PromptStripper( | |
70 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), |
|
72 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), | |
71 | initial_re=re.compile(r'^>>>( |$)') |
|
73 | initial_re=re.compile(r'^>>>( |$)') | |
72 | ) |
|
74 | ) | |
73 |
|
75 | |||
74 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) |
|
76 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) | |
75 |
|
77 | |||
76 | def cell_magic(lines): |
|
78 | def cell_magic(lines): | |
77 | if not lines[0].startswith('%%'): |
|
79 | if not lines or not lines[0].startswith('%%'): | |
78 | return lines |
|
80 | return lines | |
79 | if re.match('%%\w+\?', lines[0]): |
|
81 | if re.match('%%\w+\?', lines[0]): | |
80 | # This case will be handled by help_end |
|
82 | # This case will be handled by help_end | |
81 | return lines |
|
83 | return lines | |
82 | magic_name, _, first_line = lines[0][2:-1].partition(' ') |
|
84 | magic_name, _, first_line = lines[0][2:-1].partition(' ') | |
83 | body = ''.join(lines[1:]) |
|
85 | body = ''.join(lines[1:]) | |
84 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' |
|
86 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' | |
85 | % (magic_name, first_line, body)] |
|
87 | % (magic_name, first_line, body)] | |
86 |
|
88 | |||
87 |
|
89 | |||
88 | def _find_assign_op(token_line): |
|
90 | def _find_assign_op(token_line): | |
89 | """Get the index of the first assignment in the line ('=' not inside brackets) |
|
91 | """Get the index of the first assignment in the line ('=' not inside brackets) | |
90 |
|
92 | |||
91 | Note: We don't try to support multiple special assignment (a = b = %foo) |
|
93 | Note: We don't try to support multiple special assignment (a = b = %foo) | |
92 | """ |
|
94 | """ | |
93 | paren_level = 0 |
|
95 | paren_level = 0 | |
94 | for i, ti in enumerate(token_line): |
|
96 | for i, ti in enumerate(token_line): | |
95 | s = ti.string |
|
97 | s = ti.string | |
96 | if s == '=' and paren_level == 0: |
|
98 | if s == '=' and paren_level == 0: | |
97 |
return i |
|
99 | return i | |
98 | if s in '([{': |
|
100 | if s in '([{': | |
99 | paren_level += 1 |
|
101 | paren_level += 1 | |
100 | elif s in ')]}': |
|
102 | elif s in ')]}': | |
101 | if paren_level > 0: |
|
103 | if paren_level > 0: | |
102 | paren_level -= 1 |
|
104 | paren_level -= 1 | |
103 |
|
105 | |||
104 | def find_end_of_continued_line(lines, start_line: int): |
|
106 | def find_end_of_continued_line(lines, start_line: int): | |
105 | """Find the last line of a line explicitly extended using backslashes. |
|
107 | """Find the last line of a line explicitly extended using backslashes. | |
106 |
|
108 | |||
107 | Uses 0-indexed line numbers. |
|
109 | Uses 0-indexed line numbers. | |
108 | """ |
|
110 | """ | |
109 | end_line = start_line |
|
111 | end_line = start_line | |
110 | while lines[end_line].endswith('\\\n'): |
|
112 | while lines[end_line].endswith('\\\n'): | |
111 | end_line += 1 |
|
113 | end_line += 1 | |
112 | if end_line >= len(lines): |
|
114 | if end_line >= len(lines): | |
113 | break |
|
115 | break | |
114 | return end_line |
|
116 | return end_line | |
115 |
|
117 | |||
116 | def assemble_continued_line(lines, start: Tuple[int, int], end_line: int): |
|
118 | def assemble_continued_line(lines, start: Tuple[int, int], end_line: int): | |
117 |
"""Assemble a single line from multiple continued line pieces |
|
119 | """Assemble a single line from multiple continued line pieces | |
118 |
|
120 | |||
119 | Continued lines are lines ending in ``\``, and the line following the last |
|
121 | Continued lines are lines ending in ``\``, and the line following the last | |
120 | ``\`` in the block. |
|
122 | ``\`` in the block. | |
121 |
|
123 | |||
122 | For example, this code continues over multiple lines:: |
|
124 | For example, this code continues over multiple lines:: | |
123 |
|
125 | |||
124 | if (assign_ix is not None) \ |
|
126 | if (assign_ix is not None) \ | |
125 | and (len(line) >= assign_ix + 2) \ |
|
127 | and (len(line) >= assign_ix + 2) \ | |
126 | and (line[assign_ix+1].string == '%') \ |
|
128 | and (line[assign_ix+1].string == '%') \ | |
127 | and (line[assign_ix+2].type == tokenize.NAME): |
|
129 | and (line[assign_ix+2].type == tokenize.NAME): | |
128 |
|
130 | |||
129 | This statement contains four continued line pieces. |
|
131 | This statement contains four continued line pieces. | |
130 | Assembling these pieces into a single line would give:: |
|
132 | Assembling these pieces into a single line would give:: | |
131 |
|
133 | |||
132 | if (assign_ix is not None) and (len(line) >= assign_ix + 2) and (line[... |
|
134 | if (assign_ix is not None) and (len(line) >= assign_ix + 2) and (line[... | |
133 |
|
135 | |||
134 | This uses 0-indexed line numbers. *start* is (lineno, colno). |
|
136 | This uses 0-indexed line numbers. *start* is (lineno, colno). | |
135 |
|
137 | |||
136 | Used to allow ``%magic`` and ``!system`` commands to be continued over |
|
138 | Used to allow ``%magic`` and ``!system`` commands to be continued over | |
137 | multiple lines. |
|
139 | multiple lines. | |
138 | """ |
|
140 | """ | |
139 | parts = [lines[start[0]][start[1]:]] + lines[start[0]+1:end_line+1] |
|
141 | parts = [lines[start[0]][start[1]:]] + lines[start[0]+1:end_line+1] | |
140 | return ' '.join([p[:-2] for p in parts[:-1]] # Strip backslash+newline |
|
142 | return ' '.join([p[:-2] for p in parts[:-1]] # Strip backslash+newline | |
141 | + [parts[-1][:-1]]) # Strip newline from last line |
|
143 | + [parts[-1][:-1]]) # Strip newline from last line | |
142 |
|
144 | |||
143 | class TokenTransformBase: |
|
145 | class TokenTransformBase: | |
144 | """Base class for transformations which examine tokens. |
|
146 | """Base class for transformations which examine tokens. | |
145 |
|
147 | |||
146 | Special syntax should not be transformed when it occurs inside strings or |
|
148 | Special syntax should not be transformed when it occurs inside strings or | |
147 | comments. This is hard to reliably avoid with regexes. The solution is to |
|
149 | comments. This is hard to reliably avoid with regexes. The solution is to | |
148 | tokenise the code as Python, and recognise the special syntax in the tokens. |
|
150 | tokenise the code as Python, and recognise the special syntax in the tokens. | |
149 |
|
151 | |||
150 | IPython's special syntax is not valid Python syntax, so tokenising may go |
|
152 | IPython's special syntax is not valid Python syntax, so tokenising may go | |
151 | wrong after the special syntax starts. These classes therefore find and |
|
153 | wrong after the special syntax starts. These classes therefore find and | |
152 | transform *one* instance of special syntax at a time into regular Python |
|
154 | transform *one* instance of special syntax at a time into regular Python | |
153 | syntax. After each transformation, tokens are regenerated to find the next |
|
155 | syntax. After each transformation, tokens are regenerated to find the next | |
154 | piece of special syntax. |
|
156 | piece of special syntax. | |
155 |
|
157 | |||
156 | Subclasses need to implement one class method (find) |
|
158 | Subclasses need to implement one class method (find) | |
157 | and one regular method (transform). |
|
159 | and one regular method (transform). | |
158 |
|
160 | |||
159 | The priority attribute can select which transformation to apply if multiple |
|
161 | The priority attribute can select which transformation to apply if multiple | |
160 | transformers match in the same place. Lower numbers have higher priority. |
|
162 | transformers match in the same place. Lower numbers have higher priority. | |
161 | This allows "%magic?" to be turned into a help call rather than a magic call. |
|
163 | This allows "%magic?" to be turned into a help call rather than a magic call. | |
162 | """ |
|
164 | """ | |
163 | # Lower numbers -> higher priority (for matches in the same location) |
|
165 | # Lower numbers -> higher priority (for matches in the same location) | |
164 | priority = 10 |
|
166 | priority = 10 | |
165 |
|
167 | |||
166 | def sortby(self): |
|
168 | def sortby(self): | |
167 | return self.start_line, self.start_col, self.priority |
|
169 | return self.start_line, self.start_col, self.priority | |
168 |
|
170 | |||
169 | def __init__(self, start): |
|
171 | def __init__(self, start): | |
170 | self.start_line = start[0] - 1 # Shift from 1-index to 0-index |
|
172 | self.start_line = start[0] - 1 # Shift from 1-index to 0-index | |
171 | self.start_col = start[1] |
|
173 | self.start_col = start[1] | |
172 |
|
174 | |||
173 | @classmethod |
|
175 | @classmethod | |
174 | def find(cls, tokens_by_line): |
|
176 | def find(cls, tokens_by_line): | |
175 | """Find one instance of special syntax in the provided tokens. |
|
177 | """Find one instance of special syntax in the provided tokens. | |
176 |
|
178 | |||
177 | Tokens are grouped into logical lines for convenience, |
|
179 | Tokens are grouped into logical lines for convenience, | |
178 | so it is easy to e.g. look at the first token of each line. |
|
180 | so it is easy to e.g. look at the first token of each line. | |
179 | *tokens_by_line* is a list of lists of tokenize.TokenInfo objects. |
|
181 | *tokens_by_line* is a list of lists of tokenize.TokenInfo objects. | |
180 |
|
182 | |||
181 | This should return an instance of its class, pointing to the start |
|
183 | This should return an instance of its class, pointing to the start | |
182 | position it has found, or None if it found no match. |
|
184 | position it has found, or None if it found no match. | |
183 | """ |
|
185 | """ | |
184 | raise NotImplementedError |
|
186 | raise NotImplementedError | |
185 |
|
187 | |||
186 | def transform(self, lines: List[str]): |
|
188 | def transform(self, lines: List[str]): | |
187 | """Transform one instance of special syntax found by ``find()`` |
|
189 | """Transform one instance of special syntax found by ``find()`` | |
188 |
|
190 | |||
189 | Takes a list of strings representing physical lines, |
|
191 | Takes a list of strings representing physical lines, | |
190 | returns a similar list of transformed lines. |
|
192 | returns a similar list of transformed lines. | |
191 | """ |
|
193 | """ | |
192 | raise NotImplementedError |
|
194 | raise NotImplementedError | |
193 |
|
195 | |||
194 | class MagicAssign(TokenTransformBase): |
|
196 | class MagicAssign(TokenTransformBase): | |
195 | """Transformer for assignments from magics (a = %foo)""" |
|
197 | """Transformer for assignments from magics (a = %foo)""" | |
196 | @classmethod |
|
198 | @classmethod | |
197 | def find(cls, tokens_by_line): |
|
199 | def find(cls, tokens_by_line): | |
198 | """Find the first magic assignment (a = %foo) in the cell. |
|
200 | """Find the first magic assignment (a = %foo) in the cell. | |
199 | """ |
|
201 | """ | |
200 | for line in tokens_by_line: |
|
202 | for line in tokens_by_line: | |
201 | assign_ix = _find_assign_op(line) |
|
203 | assign_ix = _find_assign_op(line) | |
202 | if (assign_ix is not None) \ |
|
204 | if (assign_ix is not None) \ | |
203 | and (len(line) >= assign_ix + 2) \ |
|
205 | and (len(line) >= assign_ix + 2) \ | |
204 | and (line[assign_ix+1].string == '%') \ |
|
206 | and (line[assign_ix+1].string == '%') \ | |
205 | and (line[assign_ix+2].type == tokenize.NAME): |
|
207 | and (line[assign_ix+2].type == tokenize.NAME): | |
206 | return cls(line[assign_ix+1].start) |
|
208 | return cls(line[assign_ix+1].start) | |
207 |
|
209 | |||
208 | def transform(self, lines: List[str]): |
|
210 | def transform(self, lines: List[str]): | |
209 | """Transform a magic assignment found by the ``find()`` classmethod. |
|
211 | """Transform a magic assignment found by the ``find()`` classmethod. | |
210 | """ |
|
212 | """ | |
211 | start_line, start_col = self.start_line, self.start_col |
|
213 | start_line, start_col = self.start_line, self.start_col | |
212 | lhs = lines[start_line][:start_col] |
|
214 | lhs = lines[start_line][:start_col] | |
213 | end_line = find_end_of_continued_line(lines, start_line) |
|
215 | end_line = find_end_of_continued_line(lines, start_line) | |
214 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
216 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) | |
215 | assert rhs.startswith('%'), rhs |
|
217 | assert rhs.startswith('%'), rhs | |
216 | magic_name, _, args = rhs[1:].partition(' ') |
|
218 | magic_name, _, args = rhs[1:].partition(' ') | |
217 |
|
219 | |||
218 | lines_before = lines[:start_line] |
|
220 | lines_before = lines[:start_line] | |
219 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) |
|
221 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) | |
220 | new_line = lhs + call + '\n' |
|
222 | new_line = lhs + call + '\n' | |
221 | lines_after = lines[end_line+1:] |
|
223 | lines_after = lines[end_line+1:] | |
222 |
|
224 | |||
223 | return lines_before + [new_line] + lines_after |
|
225 | return lines_before + [new_line] + lines_after | |
224 |
|
226 | |||
225 |
|
227 | |||
226 | class SystemAssign(TokenTransformBase): |
|
228 | class SystemAssign(TokenTransformBase): | |
227 | """Transformer for assignments from system commands (a = !foo)""" |
|
229 | """Transformer for assignments from system commands (a = !foo)""" | |
228 | @classmethod |
|
230 | @classmethod | |
229 | def find(cls, tokens_by_line): |
|
231 | def find(cls, tokens_by_line): | |
230 | """Find the first system assignment (a = !foo) in the cell. |
|
232 | """Find the first system assignment (a = !foo) in the cell. | |
231 | """ |
|
233 | """ | |
232 | for line in tokens_by_line: |
|
234 | for line in tokens_by_line: | |
233 | assign_ix = _find_assign_op(line) |
|
235 | assign_ix = _find_assign_op(line) | |
234 | if (assign_ix is not None) \ |
|
236 | if (assign_ix is not None) \ | |
235 | and (len(line) >= assign_ix + 2) \ |
|
237 | and (len(line) >= assign_ix + 2) \ | |
236 | and (line[assign_ix + 1].type == tokenize.ERRORTOKEN): |
|
238 | and (line[assign_ix + 1].type == tokenize.ERRORTOKEN): | |
237 | ix = assign_ix + 1 |
|
239 | ix = assign_ix + 1 | |
238 |
|
240 | |||
239 | while ix < len(line) and line[ix].type == tokenize.ERRORTOKEN: |
|
241 | while ix < len(line) and line[ix].type == tokenize.ERRORTOKEN: | |
240 | if line[ix].string == '!': |
|
242 | if line[ix].string == '!': | |
241 | return cls(line[ix].start) |
|
243 | return cls(line[ix].start) | |
242 | elif not line[ix].string.isspace(): |
|
244 | elif not line[ix].string.isspace(): | |
243 | break |
|
245 | break | |
244 | ix += 1 |
|
246 | ix += 1 | |
245 |
|
247 | |||
246 | def transform(self, lines: List[str]): |
|
248 | def transform(self, lines: List[str]): | |
247 | """Transform a system assignment found by the ``find()`` classmethod. |
|
249 | """Transform a system assignment found by the ``find()`` classmethod. | |
248 | """ |
|
250 | """ | |
249 | start_line, start_col = self.start_line, self.start_col |
|
251 | start_line, start_col = self.start_line, self.start_col | |
250 |
|
252 | |||
251 | lhs = lines[start_line][:start_col] |
|
253 | lhs = lines[start_line][:start_col] | |
252 | end_line = find_end_of_continued_line(lines, start_line) |
|
254 | end_line = find_end_of_continued_line(lines, start_line) | |
253 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
255 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) | |
254 | assert rhs.startswith('!'), rhs |
|
256 | assert rhs.startswith('!'), rhs | |
255 | cmd = rhs[1:] |
|
257 | cmd = rhs[1:] | |
256 |
|
258 | |||
257 | lines_before = lines[:start_line] |
|
259 | lines_before = lines[:start_line] | |
258 | call = "get_ipython().getoutput({!r})".format(cmd) |
|
260 | call = "get_ipython().getoutput({!r})".format(cmd) | |
259 | new_line = lhs + call + '\n' |
|
261 | new_line = lhs + call + '\n' | |
260 | lines_after = lines[end_line + 1:] |
|
262 | lines_after = lines[end_line + 1:] | |
261 |
|
263 | |||
262 | return lines_before + [new_line] + lines_after |
|
264 | return lines_before + [new_line] + lines_after | |
263 |
|
265 | |||
264 | # The escape sequences that define the syntax transformations IPython will |
|
266 | # The escape sequences that define the syntax transformations IPython will | |
265 | # apply to user input. These can NOT be just changed here: many regular |
|
267 | # apply to user input. These can NOT be just changed here: many regular | |
266 | # expressions and other parts of the code may use their hardcoded values, and |
|
268 | # expressions and other parts of the code may use their hardcoded values, and | |
267 | # for all intents and purposes they constitute the 'IPython syntax', so they |
|
269 | # for all intents and purposes they constitute the 'IPython syntax', so they | |
268 | # should be considered fixed. |
|
270 | # should be considered fixed. | |
269 |
|
271 | |||
270 | ESC_SHELL = '!' # Send line to underlying system shell |
|
272 | ESC_SHELL = '!' # Send line to underlying system shell | |
271 | ESC_SH_CAP = '!!' # Send line to system shell and capture output |
|
273 | ESC_SH_CAP = '!!' # Send line to system shell and capture output | |
272 | ESC_HELP = '?' # Find information about object |
|
274 | ESC_HELP = '?' # Find information about object | |
273 | ESC_HELP2 = '??' # Find extra-detailed information about object |
|
275 | ESC_HELP2 = '??' # Find extra-detailed information about object | |
274 | ESC_MAGIC = '%' # Call magic function |
|
276 | ESC_MAGIC = '%' # Call magic function | |
275 | ESC_MAGIC2 = '%%' # Call cell-magic function |
|
277 | ESC_MAGIC2 = '%%' # Call cell-magic function | |
276 | ESC_QUOTE = ',' # Split args on whitespace, quote each as string and call |
|
278 | ESC_QUOTE = ',' # Split args on whitespace, quote each as string and call | |
277 | ESC_QUOTE2 = ';' # Quote all args as a single string, call |
|
279 | ESC_QUOTE2 = ';' # Quote all args as a single string, call | |
278 | ESC_PAREN = '/' # Call first argument with rest of line as arguments |
|
280 | ESC_PAREN = '/' # Call first argument with rest of line as arguments | |
279 |
|
281 | |||
280 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} |
|
282 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} | |
281 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately |
|
283 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately | |
282 |
|
284 | |||
283 | def _make_help_call(target, esc, next_input=None): |
|
285 | def _make_help_call(target, esc, next_input=None): | |
284 | """Prepares a pinfo(2)/psearch call from a target name and the escape |
|
286 | """Prepares a pinfo(2)/psearch call from a target name and the escape | |
285 | (i.e. ? or ??)""" |
|
287 | (i.e. ? or ??)""" | |
286 | method = 'pinfo2' if esc == '??' \ |
|
288 | method = 'pinfo2' if esc == '??' \ | |
287 | else 'psearch' if '*' in target \ |
|
289 | else 'psearch' if '*' in target \ | |
288 | else 'pinfo' |
|
290 | else 'pinfo' | |
289 | arg = " ".join([method, target]) |
|
291 | arg = " ".join([method, target]) | |
290 | #Prepare arguments for get_ipython().run_line_magic(magic_name, magic_args) |
|
292 | #Prepare arguments for get_ipython().run_line_magic(magic_name, magic_args) | |
291 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') |
|
293 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') | |
292 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) |
|
294 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) | |
293 | if next_input is None: |
|
295 | if next_input is None: | |
294 | return 'get_ipython().run_line_magic(%r, %r)' % (t_magic_name, t_magic_arg_s) |
|
296 | return 'get_ipython().run_line_magic(%r, %r)' % (t_magic_name, t_magic_arg_s) | |
295 | else: |
|
297 | else: | |
296 | return 'get_ipython().set_next_input(%r);get_ipython().run_line_magic(%r, %r)' % \ |
|
298 | return 'get_ipython().set_next_input(%r);get_ipython().run_line_magic(%r, %r)' % \ | |
297 | (next_input, t_magic_name, t_magic_arg_s) |
|
299 | (next_input, t_magic_name, t_magic_arg_s) | |
298 |
|
300 | |||
299 | def _tr_help(content): |
|
301 | def _tr_help(content): | |
300 | """Translate lines escaped with: ? |
|
302 | """Translate lines escaped with: ? | |
301 |
|
303 | |||
302 | A naked help line should fire the intro help screen (shell.show_usage()) |
|
304 | A naked help line should fire the intro help screen (shell.show_usage()) | |
303 | """ |
|
305 | """ | |
304 | if not content: |
|
306 | if not content: | |
305 | return 'get_ipython().show_usage()' |
|
307 | return 'get_ipython().show_usage()' | |
306 |
|
308 | |||
307 | return _make_help_call(content, '?') |
|
309 | return _make_help_call(content, '?') | |
308 |
|
310 | |||
309 | def _tr_help2(content): |
|
311 | def _tr_help2(content): | |
310 | """Translate lines escaped with: ?? |
|
312 | """Translate lines escaped with: ?? | |
311 |
|
313 | |||
312 | A naked help line should fire the intro help screen (shell.show_usage()) |
|
314 | A naked help line should fire the intro help screen (shell.show_usage()) | |
313 | """ |
|
315 | """ | |
314 | if not content: |
|
316 | if not content: | |
315 | return 'get_ipython().show_usage()' |
|
317 | return 'get_ipython().show_usage()' | |
316 |
|
318 | |||
317 | return _make_help_call(content, '??') |
|
319 | return _make_help_call(content, '??') | |
318 |
|
320 | |||
319 | def _tr_magic(content): |
|
321 | def _tr_magic(content): | |
320 | "Translate lines escaped with a percent sign: %" |
|
322 | "Translate lines escaped with a percent sign: %" | |
321 | name, _, args = content.partition(' ') |
|
323 | name, _, args = content.partition(' ') | |
322 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) |
|
324 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) | |
323 |
|
325 | |||
324 | def _tr_quote(content): |
|
326 | def _tr_quote(content): | |
325 | "Translate lines escaped with a comma: ," |
|
327 | "Translate lines escaped with a comma: ," | |
326 | name, _, args = content.partition(' ') |
|
328 | name, _, args = content.partition(' ') | |
327 | return '%s("%s")' % (name, '", "'.join(args.split()) ) |
|
329 | return '%s("%s")' % (name, '", "'.join(args.split()) ) | |
328 |
|
330 | |||
329 | def _tr_quote2(content): |
|
331 | def _tr_quote2(content): | |
330 | "Translate lines escaped with a semicolon: ;" |
|
332 | "Translate lines escaped with a semicolon: ;" | |
331 | name, _, args = content.partition(' ') |
|
333 | name, _, args = content.partition(' ') | |
332 | return '%s("%s")' % (name, args) |
|
334 | return '%s("%s")' % (name, args) | |
333 |
|
335 | |||
334 | def _tr_paren(content): |
|
336 | def _tr_paren(content): | |
335 | "Translate lines escaped with a slash: /" |
|
337 | "Translate lines escaped with a slash: /" | |
336 | name, _, args = content.partition(' ') |
|
338 | name, _, args = content.partition(' ') | |
337 | return '%s(%s)' % (name, ", ".join(args.split())) |
|
339 | return '%s(%s)' % (name, ", ".join(args.split())) | |
338 |
|
340 | |||
339 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, |
|
341 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, | |
340 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, |
|
342 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, | |
341 | ESC_HELP : _tr_help, |
|
343 | ESC_HELP : _tr_help, | |
342 | ESC_HELP2 : _tr_help2, |
|
344 | ESC_HELP2 : _tr_help2, | |
343 | ESC_MAGIC : _tr_magic, |
|
345 | ESC_MAGIC : _tr_magic, | |
344 | ESC_QUOTE : _tr_quote, |
|
346 | ESC_QUOTE : _tr_quote, | |
345 | ESC_QUOTE2 : _tr_quote2, |
|
347 | ESC_QUOTE2 : _tr_quote2, | |
346 | ESC_PAREN : _tr_paren } |
|
348 | ESC_PAREN : _tr_paren } | |
347 |
|
349 | |||
348 | class EscapedCommand(TokenTransformBase): |
|
350 | class EscapedCommand(TokenTransformBase): | |
349 | """Transformer for escaped commands like %foo, !foo, or /foo""" |
|
351 | """Transformer for escaped commands like %foo, !foo, or /foo""" | |
350 | @classmethod |
|
352 | @classmethod | |
351 | def find(cls, tokens_by_line): |
|
353 | def find(cls, tokens_by_line): | |
352 | """Find the first escaped command (%foo, !foo, etc.) in the cell. |
|
354 | """Find the first escaped command (%foo, !foo, etc.) in the cell. | |
353 | """ |
|
355 | """ | |
354 | for line in tokens_by_line: |
|
356 | for line in tokens_by_line: | |
355 | ix = 0 |
|
357 | ix = 0 | |
356 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
358 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
357 | ix += 1 |
|
359 | ix += 1 | |
358 | if line[ix].string in ESCAPE_SINGLES: |
|
360 | if line[ix].string in ESCAPE_SINGLES: | |
359 | return cls(line[ix].start) |
|
361 | return cls(line[ix].start) | |
360 |
|
362 | |||
361 | def transform(self, lines): |
|
363 | def transform(self, lines): | |
362 | """Transform an escaped line found by the ``find()`` classmethod. |
|
364 | """Transform an escaped line found by the ``find()`` classmethod. | |
363 | """ |
|
365 | """ | |
364 | start_line, start_col = self.start_line, self.start_col |
|
366 | start_line, start_col = self.start_line, self.start_col | |
365 |
|
367 | |||
366 | indent = lines[start_line][:start_col] |
|
368 | indent = lines[start_line][:start_col] | |
367 | end_line = find_end_of_continued_line(lines, start_line) |
|
369 | end_line = find_end_of_continued_line(lines, start_line) | |
368 | line = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
370 | line = assemble_continued_line(lines, (start_line, start_col), end_line) | |
369 |
|
371 | |||
370 | if line[:2] in ESCAPE_DOUBLES: |
|
372 | if line[:2] in ESCAPE_DOUBLES: | |
371 | escape, content = line[:2], line[2:] |
|
373 | escape, content = line[:2], line[2:] | |
372 | else: |
|
374 | else: | |
373 | escape, content = line[:1], line[1:] |
|
375 | escape, content = line[:1], line[1:] | |
374 | call = tr[escape](content) |
|
376 | call = tr[escape](content) | |
375 |
|
377 | |||
376 | lines_before = lines[:start_line] |
|
378 | lines_before = lines[:start_line] | |
377 | new_line = indent + call + '\n' |
|
379 | new_line = indent + call + '\n' | |
378 | lines_after = lines[end_line + 1:] |
|
380 | lines_after = lines[end_line + 1:] | |
379 |
|
381 | |||
380 | return lines_before + [new_line] + lines_after |
|
382 | return lines_before + [new_line] + lines_after | |
381 |
|
383 | |||
382 | _help_end_re = re.compile(r"""(%{0,2} |
|
384 | _help_end_re = re.compile(r"""(%{0,2} | |
383 | [a-zA-Z_*][\w*]* # Variable name |
|
385 | [a-zA-Z_*][\w*]* # Variable name | |
384 | (\.[a-zA-Z_*][\w*]*)* # .etc.etc |
|
386 | (\.[a-zA-Z_*][\w*]*)* # .etc.etc | |
385 | ) |
|
387 | ) | |
386 | (\?\??)$ # ? or ?? |
|
388 | (\?\??)$ # ? or ?? | |
387 | """, |
|
389 | """, | |
388 | re.VERBOSE) |
|
390 | re.VERBOSE) | |
389 |
|
391 | |||
390 | class HelpEnd(TokenTransformBase): |
|
392 | class HelpEnd(TokenTransformBase): | |
391 | """Transformer for help syntax: obj? and obj??""" |
|
393 | """Transformer for help syntax: obj? and obj??""" | |
392 | # This needs to be higher priority (lower number) than EscapedCommand so |
|
394 | # This needs to be higher priority (lower number) than EscapedCommand so | |
393 | # that inspecting magics (%foo?) works. |
|
395 | # that inspecting magics (%foo?) works. | |
394 | priority = 5 |
|
396 | priority = 5 | |
395 |
|
397 | |||
396 | def __init__(self, start, q_locn): |
|
398 | def __init__(self, start, q_locn): | |
397 | super().__init__(start) |
|
399 | super().__init__(start) | |
398 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed |
|
400 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed | |
399 | self.q_col = q_locn[1] |
|
401 | self.q_col = q_locn[1] | |
400 |
|
402 | |||
401 | @classmethod |
|
403 | @classmethod | |
402 | def find(cls, tokens_by_line): |
|
404 | def find(cls, tokens_by_line): | |
403 | """Find the first help command (foo?) in the cell. |
|
405 | """Find the first help command (foo?) in the cell. | |
404 | """ |
|
406 | """ | |
405 | for line in tokens_by_line: |
|
407 | for line in tokens_by_line: | |
406 | # Last token is NEWLINE; look at last but one |
|
408 | # Last token is NEWLINE; look at last but one | |
407 | if len(line) > 2 and line[-2].string == '?': |
|
409 | if len(line) > 2 and line[-2].string == '?': | |
408 | # Find the first token that's not INDENT/DEDENT |
|
410 | # Find the first token that's not INDENT/DEDENT | |
409 | ix = 0 |
|
411 | ix = 0 | |
410 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
412 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
411 | ix += 1 |
|
413 | ix += 1 | |
412 | return cls(line[ix].start, line[-2].start) |
|
414 | return cls(line[ix].start, line[-2].start) | |
413 |
|
415 | |||
414 | def transform(self, lines): |
|
416 | def transform(self, lines): | |
415 | """Transform a help command found by the ``find()`` classmethod. |
|
417 | """Transform a help command found by the ``find()`` classmethod. | |
416 | """ |
|
418 | """ | |
417 | piece = ''.join(lines[self.start_line:self.q_line+1]) |
|
419 | piece = ''.join(lines[self.start_line:self.q_line+1]) | |
418 | indent, content = piece[:self.start_col], piece[self.start_col:] |
|
420 | indent, content = piece[:self.start_col], piece[self.start_col:] | |
419 | lines_before = lines[:self.start_line] |
|
421 | lines_before = lines[:self.start_line] | |
420 | lines_after = lines[self.q_line + 1:] |
|
422 | lines_after = lines[self.q_line + 1:] | |
421 |
|
423 | |||
422 | m = _help_end_re.search(content) |
|
424 | m = _help_end_re.search(content) | |
423 | if not m: |
|
425 | if not m: | |
424 | raise SyntaxError(content) |
|
426 | raise SyntaxError(content) | |
425 | assert m is not None, content |
|
427 | assert m is not None, content | |
426 | target = m.group(1) |
|
428 | target = m.group(1) | |
427 | esc = m.group(3) |
|
429 | esc = m.group(3) | |
428 |
|
430 | |||
429 | # If we're mid-command, put it back on the next prompt for the user. |
|
431 | # If we're mid-command, put it back on the next prompt for the user. | |
430 | next_input = None |
|
432 | next_input = None | |
431 | if (not lines_before) and (not lines_after) \ |
|
433 | if (not lines_before) and (not lines_after) \ | |
432 | and content.strip() != m.group(0): |
|
434 | and content.strip() != m.group(0): | |
433 | next_input = content.rstrip('?\n') |
|
435 | next_input = content.rstrip('?\n') | |
434 |
|
436 | |||
435 | call = _make_help_call(target, esc, next_input=next_input) |
|
437 | call = _make_help_call(target, esc, next_input=next_input) | |
436 | new_line = indent + call + '\n' |
|
438 | new_line = indent + call + '\n' | |
437 |
|
439 | |||
438 | return lines_before + [new_line] + lines_after |
|
440 | return lines_before + [new_line] + lines_after | |
439 |
|
441 | |||
440 | def make_tokens_by_line(lines): |
|
442 | def make_tokens_by_line(lines): | |
441 | """Tokenize a series of lines and group tokens by line. |
|
443 | """Tokenize a series of lines and group tokens by line. | |
442 |
|
444 | |||
443 | The tokens for a multiline Python string or expression are |
|
445 | The tokens for a multiline Python string or expression are | |
444 | grouped as one line. |
|
446 | grouped as one line. | |
445 | """ |
|
447 | """ | |
446 | # NL tokens are used inside multiline expressions, but also after blank |
|
448 | # NL tokens are used inside multiline expressions, but also after blank | |
447 | # lines or comments. This is intentional - see https://bugs.python.org/issue17061 |
|
449 | # lines or comments. This is intentional - see https://bugs.python.org/issue17061 | |
448 | # We want to group the former case together but split the latter, so we |
|
450 | # We want to group the former case together but split the latter, so we | |
449 | # track parentheses level, similar to the internals of tokenize. |
|
451 | # track parentheses level, similar to the internals of tokenize. | |
450 | NEWLINE, NL = tokenize.NEWLINE, tokenize.NL |
|
452 | NEWLINE, NL = tokenize.NEWLINE, tokenize.NL | |
451 | tokens_by_line = [[]] |
|
453 | tokens_by_line = [[]] | |
452 | parenlev = 0 |
|
454 | parenlev = 0 | |
453 | try: |
|
455 | try: | |
454 | for token in tokenize.generate_tokens(iter(lines).__next__): |
|
456 | for token in tokenize.generate_tokens(iter(lines).__next__): | |
455 | tokens_by_line[-1].append(token) |
|
457 | tokens_by_line[-1].append(token) | |
456 | if (token.type == NEWLINE) \ |
|
458 | if (token.type == NEWLINE) \ | |
457 | or ((token.type == NL) and (parenlev <= 0)): |
|
459 | or ((token.type == NL) and (parenlev <= 0)): | |
458 | tokens_by_line.append([]) |
|
460 | tokens_by_line.append([]) | |
459 | elif token.string in {'(', '[', '{'}: |
|
461 | elif token.string in {'(', '[', '{'}: | |
460 | parenlev += 1 |
|
462 | parenlev += 1 | |
461 | elif token.string in {')', ']', '}'}: |
|
463 | elif token.string in {')', ']', '}'}: | |
462 | if parenlev > 0: |
|
464 | if parenlev > 0: | |
463 | parenlev -= 1 |
|
465 | parenlev -= 1 | |
464 | except tokenize.TokenError: |
|
466 | except tokenize.TokenError: | |
465 | # Input ended in a multiline string or expression. That's OK for us. |
|
467 | # Input ended in a multiline string or expression. That's OK for us. | |
466 | pass |
|
468 | pass | |
467 | if not tokens_by_line[-1]: |
|
469 | if not tokens_by_line[-1]: | |
468 | tokens_by_line.pop() |
|
470 | tokens_by_line.pop() | |
469 |
|
471 | |||
470 | return tokens_by_line |
|
472 | return tokens_by_line | |
471 |
|
473 | |||
472 | def show_linewise_tokens(s: str): |
|
474 | def show_linewise_tokens(s: str): | |
473 | """For investigation and debugging""" |
|
475 | """For investigation and debugging""" | |
474 | if not s.endswith('\n'): |
|
476 | if not s.endswith('\n'): | |
475 | s += '\n' |
|
477 | s += '\n' | |
476 | lines = s.splitlines(keepends=True) |
|
478 | lines = s.splitlines(keepends=True) | |
477 | for line in make_tokens_by_line(lines): |
|
479 | for line in make_tokens_by_line(lines): | |
478 | print("Line -------") |
|
480 | print("Line -------") | |
479 | for tokinfo in line: |
|
481 | for tokinfo in line: | |
480 | print(" ", tokinfo) |
|
482 | print(" ", tokinfo) | |
481 |
|
483 | |||
482 | # Arbitrary limit to prevent getting stuck in infinite loops |
|
484 | # Arbitrary limit to prevent getting stuck in infinite loops | |
483 | TRANSFORM_LOOP_LIMIT = 500 |
|
485 | TRANSFORM_LOOP_LIMIT = 500 | |
484 |
|
486 | |||
485 | class TransformerManager: |
|
487 | class TransformerManager: | |
486 | """Applies various transformations to a cell or code block. |
|
488 | """Applies various transformations to a cell or code block. | |
487 |
|
489 | |||
488 | The key methods for external use are ``transform_cell()`` |
|
490 | The key methods for external use are ``transform_cell()`` | |
489 | and ``check_complete()``. |
|
491 | and ``check_complete()``. | |
490 | """ |
|
492 | """ | |
491 | def __init__(self): |
|
493 | def __init__(self): | |
492 | self.cleanup_transforms = [ |
|
494 | self.cleanup_transforms = [ | |
493 | leading_indent, |
|
495 | leading_indent, | |
494 | classic_prompt, |
|
496 | classic_prompt, | |
495 | ipython_prompt, |
|
497 | ipython_prompt, | |
496 | ] |
|
498 | ] | |
497 | self.line_transforms = [ |
|
499 | self.line_transforms = [ | |
498 | cell_magic, |
|
500 | cell_magic, | |
499 | ] |
|
501 | ] | |
500 | self.token_transformers = [ |
|
502 | self.token_transformers = [ | |
501 | MagicAssign, |
|
503 | MagicAssign, | |
502 | SystemAssign, |
|
504 | SystemAssign, | |
503 | EscapedCommand, |
|
505 | EscapedCommand, | |
504 | HelpEnd, |
|
506 | HelpEnd, | |
505 | ] |
|
507 | ] | |
506 |
|
508 | |||
507 | def do_one_token_transform(self, lines): |
|
509 | def do_one_token_transform(self, lines): | |
508 | """Find and run the transform earliest in the code. |
|
510 | """Find and run the transform earliest in the code. | |
509 |
|
511 | |||
510 | Returns (changed, lines). |
|
512 | Returns (changed, lines). | |
511 |
|
513 | |||
512 | This method is called repeatedly until changed is False, indicating |
|
514 | This method is called repeatedly until changed is False, indicating | |
513 | that all available transformations are complete. |
|
515 | that all available transformations are complete. | |
514 |
|
516 | |||
515 | The tokens following IPython special syntax might not be valid, so |
|
517 | The tokens following IPython special syntax might not be valid, so | |
516 | the transformed code is retokenised every time to identify the next |
|
518 | the transformed code is retokenised every time to identify the next | |
517 | piece of special syntax. Hopefully long code cells are mostly valid |
|
519 | piece of special syntax. Hopefully long code cells are mostly valid | |
518 | Python, not using lots of IPython special syntax, so this shouldn't be |
|
520 | Python, not using lots of IPython special syntax, so this shouldn't be | |
519 | a performance issue. |
|
521 | a performance issue. | |
520 | """ |
|
522 | """ | |
521 | tokens_by_line = make_tokens_by_line(lines) |
|
523 | tokens_by_line = make_tokens_by_line(lines) | |
522 | candidates = [] |
|
524 | candidates = [] | |
523 | for transformer_cls in self.token_transformers: |
|
525 | for transformer_cls in self.token_transformers: | |
524 | transformer = transformer_cls.find(tokens_by_line) |
|
526 | transformer = transformer_cls.find(tokens_by_line) | |
525 | if transformer: |
|
527 | if transformer: | |
526 | candidates.append(transformer) |
|
528 | candidates.append(transformer) | |
527 |
|
529 | |||
528 | if not candidates: |
|
530 | if not candidates: | |
529 | # Nothing to transform |
|
531 | # Nothing to transform | |
530 | return False, lines |
|
532 | return False, lines | |
531 | ordered_transformers = sorted(candidates, key=TokenTransformBase.sortby) |
|
533 | ordered_transformers = sorted(candidates, key=TokenTransformBase.sortby) | |
532 | for transformer in ordered_transformers: |
|
534 | for transformer in ordered_transformers: | |
533 | try: |
|
535 | try: | |
534 | return True, transformer.transform(lines) |
|
536 | return True, transformer.transform(lines) | |
535 | except SyntaxError: |
|
537 | except SyntaxError: | |
536 | pass |
|
538 | pass | |
537 | return False, lines |
|
539 | return False, lines | |
538 |
|
540 | |||
539 | def do_token_transforms(self, lines): |
|
541 | def do_token_transforms(self, lines): | |
540 | for _ in range(TRANSFORM_LOOP_LIMIT): |
|
542 | for _ in range(TRANSFORM_LOOP_LIMIT): | |
541 | changed, lines = self.do_one_token_transform(lines) |
|
543 | changed, lines = self.do_one_token_transform(lines) | |
542 | if not changed: |
|
544 | if not changed: | |
543 | return lines |
|
545 | return lines | |
544 |
|
546 | |||
545 | raise RuntimeError("Input transformation still changing after " |
|
547 | raise RuntimeError("Input transformation still changing after " | |
546 | "%d iterations. Aborting." % TRANSFORM_LOOP_LIMIT) |
|
548 | "%d iterations. Aborting." % TRANSFORM_LOOP_LIMIT) | |
547 |
|
549 | |||
548 | def transform_cell(self, cell: str) -> str: |
|
550 | def transform_cell(self, cell: str) -> str: | |
549 | """Transforms a cell of input code""" |
|
551 | """Transforms a cell of input code""" | |
550 | if not cell.endswith('\n'): |
|
552 | if not cell.endswith('\n'): | |
551 | cell += '\n' # Ensure the cell has a trailing newline |
|
553 | cell += '\n' # Ensure the cell has a trailing newline | |
552 | lines = cell.splitlines(keepends=True) |
|
554 | lines = cell.splitlines(keepends=True) | |
553 | for transform in self.cleanup_transforms + self.line_transforms: |
|
555 | for transform in self.cleanup_transforms + self.line_transforms: | |
554 | lines = transform(lines) |
|
556 | lines = transform(lines) | |
555 |
|
557 | |||
556 | lines = self.do_token_transforms(lines) |
|
558 | lines = self.do_token_transforms(lines) | |
557 | return ''.join(lines) |
|
559 | return ''.join(lines) | |
558 |
|
560 | |||
559 | def check_complete(self, cell: str): |
|
561 | def check_complete(self, cell: str): | |
560 | """Return whether a block of code is ready to execute, or should be continued |
|
562 | """Return whether a block of code is ready to execute, or should be continued | |
561 |
|
563 | |||
562 | Parameters |
|
564 | Parameters | |
563 | ---------- |
|
565 | ---------- | |
564 | source : string |
|
566 | source : string | |
565 | Python input code, which can be multiline. |
|
567 | Python input code, which can be multiline. | |
566 |
|
568 | |||
567 | Returns |
|
569 | Returns | |
568 | ------- |
|
570 | ------- | |
569 | status : str |
|
571 | status : str | |
570 | One of 'complete', 'incomplete', or 'invalid' if source is not a |
|
572 | One of 'complete', 'incomplete', or 'invalid' if source is not a | |
571 | prefix of valid code. |
|
573 | prefix of valid code. | |
572 | indent_spaces : int or None |
|
574 | indent_spaces : int or None | |
573 | The number of spaces by which to indent the next line of code. If |
|
575 | The number of spaces by which to indent the next line of code. If | |
574 | status is not 'incomplete', this is None. |
|
576 | status is not 'incomplete', this is None. | |
575 | """ |
|
577 | """ | |
576 | if not cell.endswith('\n'): |
|
578 | if not cell.endswith('\n'): | |
577 | cell += '\n' # Ensure the cell has a trailing newline |
|
579 | cell += '\n' # Ensure the cell has a trailing newline | |
578 | lines = cell.splitlines(keepends=True) |
|
580 | lines = cell.splitlines(keepends=True) | |
579 | if lines[-1][:-1].endswith('\\'): |
|
581 | if lines[-1][:-1].endswith('\\'): | |
580 | # Explicit backslash continuation |
|
582 | # Explicit backslash continuation | |
581 | return 'incomplete', find_last_indent(lines) |
|
583 | return 'incomplete', find_last_indent(lines) | |
582 |
|
584 | |||
583 | try: |
|
585 | try: | |
584 | for transform in self.cleanup_transforms: |
|
586 | for transform in self.cleanup_transforms: | |
585 | lines = transform(lines) |
|
587 | lines = transform(lines) | |
586 | except SyntaxError: |
|
588 | except SyntaxError: | |
587 | return 'invalid', None |
|
589 | return 'invalid', None | |
588 |
|
590 | |||
589 | if lines[0].startswith('%%'): |
|
591 | if lines[0].startswith('%%'): | |
590 | # Special case for cell magics - completion marked by blank line |
|
592 | # Special case for cell magics - completion marked by blank line | |
591 | if lines[-1].strip(): |
|
593 | if lines[-1].strip(): | |
592 | return 'incomplete', find_last_indent(lines) |
|
594 | return 'incomplete', find_last_indent(lines) | |
593 | else: |
|
595 | else: | |
594 | return 'complete', None |
|
596 | return 'complete', None | |
595 |
|
597 | |||
596 | try: |
|
598 | try: | |
597 | for transform in self.line_transforms: |
|
599 | for transform in self.line_transforms: | |
598 | lines = transform(lines) |
|
600 | lines = transform(lines) | |
599 | lines = self.do_token_transforms(lines) |
|
601 | lines = self.do_token_transforms(lines) | |
600 | except SyntaxError: |
|
602 | except SyntaxError: | |
601 | return 'invalid', None |
|
603 | return 'invalid', None | |
602 |
|
604 | |||
603 | tokens_by_line = make_tokens_by_line(lines) |
|
605 | tokens_by_line = make_tokens_by_line(lines) | |
604 | if not tokens_by_line: |
|
606 | if not tokens_by_line: | |
605 | return 'incomplete', find_last_indent(lines) |
|
607 | return 'incomplete', find_last_indent(lines) | |
606 | if tokens_by_line[-1][-1].type != tokenize.ENDMARKER: |
|
608 | if tokens_by_line[-1][-1].type != tokenize.ENDMARKER: | |
607 | # We're in a multiline string or expression |
|
609 | # We're in a multiline string or expression | |
608 | return 'incomplete', find_last_indent(lines) |
|
610 | return 'incomplete', find_last_indent(lines) | |
609 | if len(tokens_by_line) == 1: |
|
611 | if len(tokens_by_line) == 1: | |
610 | return 'incomplete', find_last_indent(lines) |
|
612 | return 'incomplete', find_last_indent(lines) | |
611 | # Find the last token on the previous line that's not NEWLINE or COMMENT |
|
613 | # Find the last token on the previous line that's not NEWLINE or COMMENT | |
612 | toks_last_line = tokens_by_line[-2] |
|
614 | toks_last_line = tokens_by_line[-2] | |
613 | ix = len(toks_last_line) - 1 |
|
615 | ix = len(toks_last_line) - 1 | |
614 | while ix >= 0 and toks_last_line[ix].type in {tokenize.NEWLINE, |
|
616 | while ix >= 0 and toks_last_line[ix].type in {tokenize.NEWLINE, | |
615 | tokenize.COMMENT}: |
|
617 | tokenize.COMMENT}: | |
616 | ix -= 1 |
|
618 | ix -= 1 | |
617 |
|
619 | |||
618 | if toks_last_line[ix].string == ':': |
|
620 | if toks_last_line[ix].string == ':': | |
619 | # The last line starts a block (e.g. 'if foo:') |
|
621 | # The last line starts a block (e.g. 'if foo:') | |
620 | ix = 0 |
|
622 | ix = 0 | |
621 | while toks_last_line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
623 | while toks_last_line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
622 | ix += 1 |
|
624 | ix += 1 | |
623 | indent = toks_last_line[ix].start[1] |
|
625 | indent = toks_last_line[ix].start[1] | |
624 | return 'incomplete', indent + 4 |
|
626 | return 'incomplete', indent + 4 | |
625 |
|
627 | |||
626 | # If there's a blank line at the end, assume we're ready to execute. |
|
628 | # If there's a blank line at the end, assume we're ready to execute. | |
627 | if not lines[-1].strip(): |
|
629 | if not lines[-1].strip(): | |
628 | return 'complete', None |
|
630 | return 'complete', None | |
629 |
|
631 | |||
630 | # At this point, our checks think the code is complete (or invalid). |
|
632 | # At this point, our checks think the code is complete (or invalid). | |
631 | # We'll use codeop.compile_command to check this with the real parser. |
|
633 | # We'll use codeop.compile_command to check this with the real parser. | |
632 |
|
634 | |||
633 | try: |
|
635 | try: | |
634 | with warnings.catch_warnings(): |
|
636 | with warnings.catch_warnings(): | |
635 | warnings.simplefilter('error', SyntaxWarning) |
|
637 | warnings.simplefilter('error', SyntaxWarning) | |
636 | res = compile_command(''.join(lines), symbol='exec') |
|
638 | res = compile_command(''.join(lines), symbol='exec') | |
637 | except (SyntaxError, OverflowError, ValueError, TypeError, |
|
639 | except (SyntaxError, OverflowError, ValueError, TypeError, | |
638 | MemoryError, SyntaxWarning): |
|
640 | MemoryError, SyntaxWarning): | |
639 | return 'invalid', None |
|
641 | return 'invalid', None | |
640 | else: |
|
642 | else: | |
641 | if res is None: |
|
643 | if res is None: | |
642 | return 'incomplete', find_last_indent(lines) |
|
644 | return 'incomplete', find_last_indent(lines) | |
643 | return 'complete', None |
|
645 | return 'complete', None | |
644 |
|
646 | |||
645 |
|
647 | |||
646 | def find_last_indent(lines): |
|
648 | def find_last_indent(lines): | |
647 | m = _indent_re.match(lines[-1]) |
|
649 | m = _indent_re.match(lines[-1]) | |
648 | if not m: |
|
650 | if not m: | |
649 | return 0 |
|
651 | return 0 | |
650 | return len(m.group(0).replace('\t', ' '*4)) |
|
652 | return len(m.group(0).replace('\t', ' '*4)) |
General Comments 0
You need to be logged in to leave comments.
Login now