Show More
@@ -1,652 +1,667 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: |
|
64 | if not lines: | |
65 | return lines |
|
65 | return lines | |
66 | if self.initial_re.match(lines[0]) or \ |
|
66 | if self.initial_re.match(lines[0]) or \ | |
67 | (len(lines) > 1 and self.prompt_re.match(lines[1])): |
|
67 | (len(lines) > 1 and self.prompt_re.match(lines[1])): | |
68 | return self._strip(lines) |
|
68 | return self._strip(lines) | |
69 | return lines |
|
69 | return lines | |
70 |
|
70 | |||
71 | classic_prompt = PromptStripper( |
|
71 | classic_prompt = PromptStripper( | |
72 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), |
|
72 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), | |
73 | initial_re=re.compile(r'^>>>( |$)') |
|
73 | initial_re=re.compile(r'^>>>( |$)') | |
74 | ) |
|
74 | ) | |
75 |
|
75 | |||
76 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) |
|
76 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) | |
77 |
|
77 | |||
78 | def cell_magic(lines): |
|
78 | def cell_magic(lines): | |
79 | if not lines or not lines[0].startswith('%%'): |
|
79 | if not lines or not lines[0].startswith('%%'): | |
80 | return lines |
|
80 | return lines | |
81 | if re.match('%%\w+\?', lines[0]): |
|
81 | if re.match('%%\w+\?', lines[0]): | |
82 | # This case will be handled by help_end |
|
82 | # This case will be handled by help_end | |
83 | return lines |
|
83 | return lines | |
84 | magic_name, _, first_line = lines[0][2:-1].partition(' ') |
|
84 | magic_name, _, first_line = lines[0][2:-1].partition(' ') | |
85 | body = ''.join(lines[1:]) |
|
85 | body = ''.join(lines[1:]) | |
86 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' |
|
86 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' | |
87 | % (magic_name, first_line, body)] |
|
87 | % (magic_name, first_line, body)] | |
88 |
|
88 | |||
89 |
|
89 | |||
90 | def _find_assign_op(token_line): |
|
90 | def _find_assign_op(token_line): | |
91 | """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) | |
92 |
|
92 | |||
93 | 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) | |
94 | """ |
|
94 | """ | |
95 | paren_level = 0 |
|
95 | paren_level = 0 | |
96 | for i, ti in enumerate(token_line): |
|
96 | for i, ti in enumerate(token_line): | |
97 | s = ti.string |
|
97 | s = ti.string | |
98 | if s == '=' and paren_level == 0: |
|
98 | if s == '=' and paren_level == 0: | |
99 | return i |
|
99 | return i | |
100 | if s in '([{': |
|
100 | if s in '([{': | |
101 | paren_level += 1 |
|
101 | paren_level += 1 | |
102 | elif s in ')]}': |
|
102 | elif s in ')]}': | |
103 | if paren_level > 0: |
|
103 | if paren_level > 0: | |
104 | paren_level -= 1 |
|
104 | paren_level -= 1 | |
105 |
|
105 | |||
106 | def find_end_of_continued_line(lines, start_line: int): |
|
106 | def find_end_of_continued_line(lines, start_line: int): | |
107 | """Find the last line of a line explicitly extended using backslashes. |
|
107 | """Find the last line of a line explicitly extended using backslashes. | |
108 |
|
108 | |||
109 | Uses 0-indexed line numbers. |
|
109 | Uses 0-indexed line numbers. | |
110 | """ |
|
110 | """ | |
111 | end_line = start_line |
|
111 | end_line = start_line | |
112 | while lines[end_line].endswith('\\\n'): |
|
112 | while lines[end_line].endswith('\\\n'): | |
113 | end_line += 1 |
|
113 | end_line += 1 | |
114 | if end_line >= len(lines): |
|
114 | if end_line >= len(lines): | |
115 | break |
|
115 | break | |
116 | return end_line |
|
116 | return end_line | |
117 |
|
117 | |||
118 | 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): | |
119 | """Assemble a single line from multiple continued line pieces |
|
119 | """Assemble a single line from multiple continued line pieces | |
120 |
|
120 | |||
121 | 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 | |
122 | ``\`` in the block. |
|
122 | ``\`` in the block. | |
123 |
|
123 | |||
124 | For example, this code continues over multiple lines:: |
|
124 | For example, this code continues over multiple lines:: | |
125 |
|
125 | |||
126 | if (assign_ix is not None) \ |
|
126 | if (assign_ix is not None) \ | |
127 | and (len(line) >= assign_ix + 2) \ |
|
127 | and (len(line) >= assign_ix + 2) \ | |
128 | and (line[assign_ix+1].string == '%') \ |
|
128 | and (line[assign_ix+1].string == '%') \ | |
129 | and (line[assign_ix+2].type == tokenize.NAME): |
|
129 | and (line[assign_ix+2].type == tokenize.NAME): | |
130 |
|
130 | |||
131 | This statement contains four continued line pieces. |
|
131 | This statement contains four continued line pieces. | |
132 | Assembling these pieces into a single line would give:: |
|
132 | Assembling these pieces into a single line would give:: | |
133 |
|
133 | |||
134 | 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[... | |
135 |
|
135 | |||
136 | This uses 0-indexed line numbers. *start* is (lineno, colno). |
|
136 | This uses 0-indexed line numbers. *start* is (lineno, colno). | |
137 |
|
137 | |||
138 | Used to allow ``%magic`` and ``!system`` commands to be continued over |
|
138 | Used to allow ``%magic`` and ``!system`` commands to be continued over | |
139 | multiple lines. |
|
139 | multiple lines. | |
140 | """ |
|
140 | """ | |
141 | 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] | |
142 | 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 | |
143 | + [parts[-1][:-1]]) # Strip newline from last line |
|
143 | + [parts[-1][:-1]]) # Strip newline from last line | |
144 |
|
144 | |||
145 | class TokenTransformBase: |
|
145 | class TokenTransformBase: | |
146 | """Base class for transformations which examine tokens. |
|
146 | """Base class for transformations which examine tokens. | |
147 |
|
147 | |||
148 | 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 | |
149 | 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 | |
150 | 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. | |
151 |
|
151 | |||
152 | 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 | |
153 | wrong after the special syntax starts. These classes therefore find and |
|
153 | wrong after the special syntax starts. These classes therefore find and | |
154 | 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 | |
155 | syntax. After each transformation, tokens are regenerated to find the next |
|
155 | syntax. After each transformation, tokens are regenerated to find the next | |
156 | piece of special syntax. |
|
156 | piece of special syntax. | |
157 |
|
157 | |||
158 | Subclasses need to implement one class method (find) |
|
158 | Subclasses need to implement one class method (find) | |
159 | and one regular method (transform). |
|
159 | and one regular method (transform). | |
160 |
|
160 | |||
161 | The priority attribute can select which transformation to apply if multiple |
|
161 | The priority attribute can select which transformation to apply if multiple | |
162 | transformers match in the same place. Lower numbers have higher priority. |
|
162 | transformers match in the same place. Lower numbers have higher priority. | |
163 | 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. | |
164 | """ |
|
164 | """ | |
165 | # Lower numbers -> higher priority (for matches in the same location) |
|
165 | # Lower numbers -> higher priority (for matches in the same location) | |
166 | priority = 10 |
|
166 | priority = 10 | |
167 |
|
167 | |||
168 | def sortby(self): |
|
168 | def sortby(self): | |
169 | return self.start_line, self.start_col, self.priority |
|
169 | return self.start_line, self.start_col, self.priority | |
170 |
|
170 | |||
171 | def __init__(self, start): |
|
171 | def __init__(self, start): | |
172 | 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 | |
173 | self.start_col = start[1] |
|
173 | self.start_col = start[1] | |
174 |
|
174 | |||
175 | @classmethod |
|
175 | @classmethod | |
176 | def find(cls, tokens_by_line): |
|
176 | def find(cls, tokens_by_line): | |
177 | """Find one instance of special syntax in the provided tokens. |
|
177 | """Find one instance of special syntax in the provided tokens. | |
178 |
|
178 | |||
179 | Tokens are grouped into logical lines for convenience, |
|
179 | Tokens are grouped into logical lines for convenience, | |
180 | 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. | |
181 | *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. | |
182 |
|
182 | |||
183 | 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 | |
184 | position it has found, or None if it found no match. |
|
184 | position it has found, or None if it found no match. | |
185 | """ |
|
185 | """ | |
186 | raise NotImplementedError |
|
186 | raise NotImplementedError | |
187 |
|
187 | |||
188 | def transform(self, lines: List[str]): |
|
188 | def transform(self, lines: List[str]): | |
189 | """Transform one instance of special syntax found by ``find()`` |
|
189 | """Transform one instance of special syntax found by ``find()`` | |
190 |
|
190 | |||
191 | Takes a list of strings representing physical lines, |
|
191 | Takes a list of strings representing physical lines, | |
192 | returns a similar list of transformed lines. |
|
192 | returns a similar list of transformed lines. | |
193 | """ |
|
193 | """ | |
194 | raise NotImplementedError |
|
194 | raise NotImplementedError | |
195 |
|
195 | |||
196 | class MagicAssign(TokenTransformBase): |
|
196 | class MagicAssign(TokenTransformBase): | |
197 | """Transformer for assignments from magics (a = %foo)""" |
|
197 | """Transformer for assignments from magics (a = %foo)""" | |
198 | @classmethod |
|
198 | @classmethod | |
199 | def find(cls, tokens_by_line): |
|
199 | def find(cls, tokens_by_line): | |
200 | """Find the first magic assignment (a = %foo) in the cell. |
|
200 | """Find the first magic assignment (a = %foo) in the cell. | |
201 | """ |
|
201 | """ | |
202 | for line in tokens_by_line: |
|
202 | for line in tokens_by_line: | |
203 | assign_ix = _find_assign_op(line) |
|
203 | assign_ix = _find_assign_op(line) | |
204 | if (assign_ix is not None) \ |
|
204 | if (assign_ix is not None) \ | |
205 | and (len(line) >= assign_ix + 2) \ |
|
205 | and (len(line) >= assign_ix + 2) \ | |
206 | and (line[assign_ix+1].string == '%') \ |
|
206 | and (line[assign_ix+1].string == '%') \ | |
207 | and (line[assign_ix+2].type == tokenize.NAME): |
|
207 | and (line[assign_ix+2].type == tokenize.NAME): | |
208 | return cls(line[assign_ix+1].start) |
|
208 | return cls(line[assign_ix+1].start) | |
209 |
|
209 | |||
210 | def transform(self, lines: List[str]): |
|
210 | def transform(self, lines: List[str]): | |
211 | """Transform a magic assignment found by the ``find()`` classmethod. |
|
211 | """Transform a magic assignment found by the ``find()`` classmethod. | |
212 | """ |
|
212 | """ | |
213 | start_line, start_col = self.start_line, self.start_col |
|
213 | start_line, start_col = self.start_line, self.start_col | |
214 | lhs = lines[start_line][:start_col] |
|
214 | lhs = lines[start_line][:start_col] | |
215 | end_line = find_end_of_continued_line(lines, start_line) |
|
215 | end_line = find_end_of_continued_line(lines, start_line) | |
216 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
216 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) | |
217 | assert rhs.startswith('%'), rhs |
|
217 | assert rhs.startswith('%'), rhs | |
218 | magic_name, _, args = rhs[1:].partition(' ') |
|
218 | magic_name, _, args = rhs[1:].partition(' ') | |
219 |
|
219 | |||
220 | lines_before = lines[:start_line] |
|
220 | lines_before = lines[:start_line] | |
221 | 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) | |
222 | new_line = lhs + call + '\n' |
|
222 | new_line = lhs + call + '\n' | |
223 | lines_after = lines[end_line+1:] |
|
223 | lines_after = lines[end_line+1:] | |
224 |
|
224 | |||
225 | return lines_before + [new_line] + lines_after |
|
225 | return lines_before + [new_line] + lines_after | |
226 |
|
226 | |||
227 |
|
227 | |||
228 | class SystemAssign(TokenTransformBase): |
|
228 | class SystemAssign(TokenTransformBase): | |
229 | """Transformer for assignments from system commands (a = !foo)""" |
|
229 | """Transformer for assignments from system commands (a = !foo)""" | |
230 | @classmethod |
|
230 | @classmethod | |
231 | def find(cls, tokens_by_line): |
|
231 | def find(cls, tokens_by_line): | |
232 | """Find the first system assignment (a = !foo) in the cell. |
|
232 | """Find the first system assignment (a = !foo) in the cell. | |
233 | """ |
|
233 | """ | |
234 | for line in tokens_by_line: |
|
234 | for line in tokens_by_line: | |
235 | assign_ix = _find_assign_op(line) |
|
235 | assign_ix = _find_assign_op(line) | |
236 | if (assign_ix is not None) \ |
|
236 | if (assign_ix is not None) \ | |
237 | and (len(line) >= assign_ix + 2) \ |
|
237 | and (len(line) >= assign_ix + 2) \ | |
238 | and (line[assign_ix + 1].type == tokenize.ERRORTOKEN): |
|
238 | and (line[assign_ix + 1].type == tokenize.ERRORTOKEN): | |
239 | ix = assign_ix + 1 |
|
239 | ix = assign_ix + 1 | |
240 |
|
240 | |||
241 | while ix < len(line) and line[ix].type == tokenize.ERRORTOKEN: |
|
241 | while ix < len(line) and line[ix].type == tokenize.ERRORTOKEN: | |
242 | if line[ix].string == '!': |
|
242 | if line[ix].string == '!': | |
243 | return cls(line[ix].start) |
|
243 | return cls(line[ix].start) | |
244 | elif not line[ix].string.isspace(): |
|
244 | elif not line[ix].string.isspace(): | |
245 | break |
|
245 | break | |
246 | ix += 1 |
|
246 | ix += 1 | |
247 |
|
247 | |||
248 | def transform(self, lines: List[str]): |
|
248 | def transform(self, lines: List[str]): | |
249 | """Transform a system assignment found by the ``find()`` classmethod. |
|
249 | """Transform a system assignment found by the ``find()`` classmethod. | |
250 | """ |
|
250 | """ | |
251 | start_line, start_col = self.start_line, self.start_col |
|
251 | start_line, start_col = self.start_line, self.start_col | |
252 |
|
252 | |||
253 | lhs = lines[start_line][:start_col] |
|
253 | lhs = lines[start_line][:start_col] | |
254 | end_line = find_end_of_continued_line(lines, start_line) |
|
254 | end_line = find_end_of_continued_line(lines, start_line) | |
255 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
255 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) | |
256 | assert rhs.startswith('!'), rhs |
|
256 | # assert rhs.startswith('!'), rhs | |
257 | cmd = rhs[1:] |
|
257 | cmd = rhs[1:] | |
258 |
|
258 | |||
259 | lines_before = lines[:start_line] |
|
259 | lines_before = lines[:start_line] | |
260 | call = "get_ipython().getoutput({!r})".format(cmd) |
|
260 | call = "get_ipython().getoutput({!r})".format(cmd) | |
261 | new_line = lhs + call + '\n' |
|
261 | new_line = lhs + call + '\n' | |
262 | lines_after = lines[end_line + 1:] |
|
262 | lines_after = lines[end_line + 1:] | |
263 |
|
263 | |||
264 | return lines_before + [new_line] + lines_after |
|
264 | return lines_before + [new_line] + lines_after | |
265 |
|
265 | |||
266 | # The escape sequences that define the syntax transformations IPython will |
|
266 | # The escape sequences that define the syntax transformations IPython will | |
267 | # 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 | |
268 | # 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 | |
269 | # 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 | |
270 | # should be considered fixed. |
|
270 | # should be considered fixed. | |
271 |
|
271 | |||
272 | ESC_SHELL = '!' # Send line to underlying system shell |
|
272 | ESC_SHELL = '!' # Send line to underlying system shell | |
273 | ESC_SH_CAP = '!!' # Send line to system shell and capture output |
|
273 | ESC_SH_CAP = '!!' # Send line to system shell and capture output | |
274 | ESC_HELP = '?' # Find information about object |
|
274 | ESC_HELP = '?' # Find information about object | |
275 | ESC_HELP2 = '??' # Find extra-detailed information about object |
|
275 | ESC_HELP2 = '??' # Find extra-detailed information about object | |
276 | ESC_MAGIC = '%' # Call magic function |
|
276 | ESC_MAGIC = '%' # Call magic function | |
277 | ESC_MAGIC2 = '%%' # Call cell-magic function |
|
277 | ESC_MAGIC2 = '%%' # Call cell-magic function | |
278 | 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 | |
279 | ESC_QUOTE2 = ';' # Quote all args as a single string, call |
|
279 | ESC_QUOTE2 = ';' # Quote all args as a single string, call | |
280 | ESC_PAREN = '/' # Call first argument with rest of line as arguments |
|
280 | ESC_PAREN = '/' # Call first argument with rest of line as arguments | |
281 |
|
281 | |||
282 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} |
|
282 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} | |
283 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately |
|
283 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately | |
284 |
|
284 | |||
285 | def _make_help_call(target, esc, next_input=None): |
|
285 | def _make_help_call(target, esc, next_input=None): | |
286 | """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 | |
287 | (i.e. ? or ??)""" |
|
287 | (i.e. ? or ??)""" | |
288 | method = 'pinfo2' if esc == '??' \ |
|
288 | method = 'pinfo2' if esc == '??' \ | |
289 | else 'psearch' if '*' in target \ |
|
289 | else 'psearch' if '*' in target \ | |
290 | else 'pinfo' |
|
290 | else 'pinfo' | |
291 | arg = " ".join([method, target]) |
|
291 | arg = " ".join([method, target]) | |
292 | #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) | |
293 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') |
|
293 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') | |
294 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) |
|
294 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) | |
295 | if next_input is None: |
|
295 | if next_input is None: | |
296 | 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) | |
297 | else: |
|
297 | else: | |
298 | 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)' % \ | |
299 | (next_input, t_magic_name, t_magic_arg_s) |
|
299 | (next_input, t_magic_name, t_magic_arg_s) | |
300 |
|
300 | |||
301 | def _tr_help(content): |
|
301 | def _tr_help(content): | |
302 | """Translate lines escaped with: ? |
|
302 | """Translate lines escaped with: ? | |
303 |
|
303 | |||
304 | 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()) | |
305 | """ |
|
305 | """ | |
306 | if not content: |
|
306 | if not content: | |
307 | return 'get_ipython().show_usage()' |
|
307 | return 'get_ipython().show_usage()' | |
308 |
|
308 | |||
309 | return _make_help_call(content, '?') |
|
309 | return _make_help_call(content, '?') | |
310 |
|
310 | |||
311 | def _tr_help2(content): |
|
311 | def _tr_help2(content): | |
312 | """Translate lines escaped with: ?? |
|
312 | """Translate lines escaped with: ?? | |
313 |
|
313 | |||
314 | 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()) | |
315 | """ |
|
315 | """ | |
316 | if not content: |
|
316 | if not content: | |
317 | return 'get_ipython().show_usage()' |
|
317 | return 'get_ipython().show_usage()' | |
318 |
|
318 | |||
319 | return _make_help_call(content, '??') |
|
319 | return _make_help_call(content, '??') | |
320 |
|
320 | |||
321 | def _tr_magic(content): |
|
321 | def _tr_magic(content): | |
322 | "Translate lines escaped with a percent sign: %" |
|
322 | "Translate lines escaped with a percent sign: %" | |
323 | name, _, args = content.partition(' ') |
|
323 | name, _, args = content.partition(' ') | |
324 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) |
|
324 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) | |
325 |
|
325 | |||
326 | def _tr_quote(content): |
|
326 | def _tr_quote(content): | |
327 | "Translate lines escaped with a comma: ," |
|
327 | "Translate lines escaped with a comma: ," | |
328 | name, _, args = content.partition(' ') |
|
328 | name, _, args = content.partition(' ') | |
329 | return '%s("%s")' % (name, '", "'.join(args.split()) ) |
|
329 | return '%s("%s")' % (name, '", "'.join(args.split()) ) | |
330 |
|
330 | |||
331 | def _tr_quote2(content): |
|
331 | def _tr_quote2(content): | |
332 | "Translate lines escaped with a semicolon: ;" |
|
332 | "Translate lines escaped with a semicolon: ;" | |
333 | name, _, args = content.partition(' ') |
|
333 | name, _, args = content.partition(' ') | |
334 | return '%s("%s")' % (name, args) |
|
334 | return '%s("%s")' % (name, args) | |
335 |
|
335 | |||
336 | def _tr_paren(content): |
|
336 | def _tr_paren(content): | |
337 | "Translate lines escaped with a slash: /" |
|
337 | "Translate lines escaped with a slash: /" | |
338 | name, _, args = content.partition(' ') |
|
338 | name, _, args = content.partition(' ') | |
339 | return '%s(%s)' % (name, ", ".join(args.split())) |
|
339 | return '%s(%s)' % (name, ", ".join(args.split())) | |
340 |
|
340 | |||
341 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, |
|
341 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, | |
342 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, |
|
342 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, | |
343 | ESC_HELP : _tr_help, |
|
343 | ESC_HELP : _tr_help, | |
344 | ESC_HELP2 : _tr_help2, |
|
344 | ESC_HELP2 : _tr_help2, | |
345 | ESC_MAGIC : _tr_magic, |
|
345 | ESC_MAGIC : _tr_magic, | |
346 | ESC_QUOTE : _tr_quote, |
|
346 | ESC_QUOTE : _tr_quote, | |
347 | ESC_QUOTE2 : _tr_quote2, |
|
347 | ESC_QUOTE2 : _tr_quote2, | |
348 | ESC_PAREN : _tr_paren } |
|
348 | ESC_PAREN : _tr_paren } | |
349 |
|
349 | |||
350 | class EscapedCommand(TokenTransformBase): |
|
350 | class EscapedCommand(TokenTransformBase): | |
351 | """Transformer for escaped commands like %foo, !foo, or /foo""" |
|
351 | """Transformer for escaped commands like %foo, !foo, or /foo""" | |
352 | @classmethod |
|
352 | @classmethod | |
353 | def find(cls, tokens_by_line): |
|
353 | def find(cls, tokens_by_line): | |
354 | """Find the first escaped command (%foo, !foo, etc.) in the cell. |
|
354 | """Find the first escaped command (%foo, !foo, etc.) in the cell. | |
355 | """ |
|
355 | """ | |
356 | for line in tokens_by_line: |
|
356 | for line in tokens_by_line: | |
357 | ix = 0 |
|
357 | ix = 0 | |
358 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
358 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
359 | ix += 1 |
|
359 | ix += 1 | |
360 | if line[ix].string in ESCAPE_SINGLES: |
|
360 | if line[ix].string in ESCAPE_SINGLES: | |
361 | return cls(line[ix].start) |
|
361 | return cls(line[ix].start) | |
362 |
|
362 | |||
363 | def transform(self, lines): |
|
363 | def transform(self, lines): | |
364 | """Transform an escaped line found by the ``find()`` classmethod. |
|
364 | """Transform an escaped line found by the ``find()`` classmethod. | |
365 | """ |
|
365 | """ | |
366 | start_line, start_col = self.start_line, self.start_col |
|
366 | start_line, start_col = self.start_line, self.start_col | |
367 |
|
367 | |||
368 | indent = lines[start_line][:start_col] |
|
368 | indent = lines[start_line][:start_col] | |
369 | end_line = find_end_of_continued_line(lines, start_line) |
|
369 | end_line = find_end_of_continued_line(lines, start_line) | |
370 | line = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
370 | line = assemble_continued_line(lines, (start_line, start_col), end_line) | |
371 |
|
371 | |||
372 | if line[:2] in ESCAPE_DOUBLES: |
|
372 | if len(line) > 1 and line[:2] in ESCAPE_DOUBLES: | |
373 | escape, content = line[:2], line[2:] |
|
373 | escape, content = line[:2], line[2:] | |
374 | else: |
|
374 | else: | |
375 | escape, content = line[:1], line[1:] |
|
375 | escape, content = line[:1], line[1:] | |
|
376 | ||||
|
377 | if escape in tr: | |||
376 | call = tr[escape](content) |
|
378 | call = tr[escape](content) | |
|
379 | else: | |||
|
380 | call = '' | |||
377 |
|
381 | |||
378 | lines_before = lines[:start_line] |
|
382 | lines_before = lines[:start_line] | |
379 | new_line = indent + call + '\n' |
|
383 | new_line = indent + call + '\n' | |
380 | lines_after = lines[end_line + 1:] |
|
384 | lines_after = lines[end_line + 1:] | |
381 |
|
385 | |||
382 | return lines_before + [new_line] + lines_after |
|
386 | return lines_before + [new_line] + lines_after | |
383 |
|
387 | |||
384 | _help_end_re = re.compile(r"""(%{0,2} |
|
388 | _help_end_re = re.compile(r"""(%{0,2} | |
385 | [a-zA-Z_*][\w*]* # Variable name |
|
389 | [a-zA-Z_*][\w*]* # Variable name | |
386 | (\.[a-zA-Z_*][\w*]*)* # .etc.etc |
|
390 | (\.[a-zA-Z_*][\w*]*)* # .etc.etc | |
387 | ) |
|
391 | ) | |
388 | (\?\??)$ # ? or ?? |
|
392 | (\?\??)$ # ? or ?? | |
389 | """, |
|
393 | """, | |
390 | re.VERBOSE) |
|
394 | re.VERBOSE) | |
391 |
|
395 | |||
392 | class HelpEnd(TokenTransformBase): |
|
396 | class HelpEnd(TokenTransformBase): | |
393 | """Transformer for help syntax: obj? and obj??""" |
|
397 | """Transformer for help syntax: obj? and obj??""" | |
394 | # This needs to be higher priority (lower number) than EscapedCommand so |
|
398 | # This needs to be higher priority (lower number) than EscapedCommand so | |
395 | # that inspecting magics (%foo?) works. |
|
399 | # that inspecting magics (%foo?) works. | |
396 | priority = 5 |
|
400 | priority = 5 | |
397 |
|
401 | |||
398 | def __init__(self, start, q_locn): |
|
402 | def __init__(self, start, q_locn): | |
399 | super().__init__(start) |
|
403 | super().__init__(start) | |
400 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed |
|
404 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed | |
401 | self.q_col = q_locn[1] |
|
405 | self.q_col = q_locn[1] | |
402 |
|
406 | |||
403 | @classmethod |
|
407 | @classmethod | |
404 | def find(cls, tokens_by_line): |
|
408 | def find(cls, tokens_by_line): | |
405 | """Find the first help command (foo?) in the cell. |
|
409 | """Find the first help command (foo?) in the cell. | |
406 | """ |
|
410 | """ | |
407 | for line in tokens_by_line: |
|
411 | for line in tokens_by_line: | |
408 | # Last token is NEWLINE; look at last but one |
|
412 | # Last token is NEWLINE; look at last but one | |
409 | if len(line) > 2 and line[-2].string == '?': |
|
413 | if len(line) > 2 and line[-2].string == '?': | |
410 | # Find the first token that's not INDENT/DEDENT |
|
414 | # Find the first token that's not INDENT/DEDENT | |
411 | ix = 0 |
|
415 | ix = 0 | |
412 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
416 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
413 | ix += 1 |
|
417 | ix += 1 | |
414 | return cls(line[ix].start, line[-2].start) |
|
418 | return cls(line[ix].start, line[-2].start) | |
415 |
|
419 | |||
416 | def transform(self, lines): |
|
420 | def transform(self, lines): | |
417 | """Transform a help command found by the ``find()`` classmethod. |
|
421 | """Transform a help command found by the ``find()`` classmethod. | |
418 | """ |
|
422 | """ | |
419 | piece = ''.join(lines[self.start_line:self.q_line+1]) |
|
423 | piece = ''.join(lines[self.start_line:self.q_line+1]) | |
420 | indent, content = piece[:self.start_col], piece[self.start_col:] |
|
424 | indent, content = piece[:self.start_col], piece[self.start_col:] | |
421 | lines_before = lines[:self.start_line] |
|
425 | lines_before = lines[:self.start_line] | |
422 | lines_after = lines[self.q_line + 1:] |
|
426 | lines_after = lines[self.q_line + 1:] | |
423 |
|
427 | |||
424 | m = _help_end_re.search(content) |
|
428 | m = _help_end_re.search(content) | |
425 | if not m: |
|
429 | if not m: | |
426 | raise SyntaxError(content) |
|
430 | raise SyntaxError(content) | |
427 | assert m is not None, content |
|
431 | assert m is not None, content | |
428 | target = m.group(1) |
|
432 | target = m.group(1) | |
429 | esc = m.group(3) |
|
433 | esc = m.group(3) | |
430 |
|
434 | |||
431 | # If we're mid-command, put it back on the next prompt for the user. |
|
435 | # If we're mid-command, put it back on the next prompt for the user. | |
432 | next_input = None |
|
436 | next_input = None | |
433 | if (not lines_before) and (not lines_after) \ |
|
437 | if (not lines_before) and (not lines_after) \ | |
434 | and content.strip() != m.group(0): |
|
438 | and content.strip() != m.group(0): | |
435 | next_input = content.rstrip('?\n') |
|
439 | next_input = content.rstrip('?\n') | |
436 |
|
440 | |||
437 | call = _make_help_call(target, esc, next_input=next_input) |
|
441 | call = _make_help_call(target, esc, next_input=next_input) | |
438 | new_line = indent + call + '\n' |
|
442 | new_line = indent + call + '\n' | |
439 |
|
443 | |||
440 | return lines_before + [new_line] + lines_after |
|
444 | return lines_before + [new_line] + lines_after | |
441 |
|
445 | |||
442 | def make_tokens_by_line(lines): |
|
446 | def make_tokens_by_line(lines): | |
443 | """Tokenize a series of lines and group tokens by line. |
|
447 | """Tokenize a series of lines and group tokens by line. | |
444 |
|
448 | |||
445 | The tokens for a multiline Python string or expression are |
|
449 | The tokens for a multiline Python string or expression are | |
446 | grouped as one line. |
|
450 | grouped as one line. | |
447 | """ |
|
451 | """ | |
448 | # NL tokens are used inside multiline expressions, but also after blank |
|
452 | # NL tokens are used inside multiline expressions, but also after blank | |
449 | # lines or comments. This is intentional - see https://bugs.python.org/issue17061 |
|
453 | # lines or comments. This is intentional - see https://bugs.python.org/issue17061 | |
450 | # We want to group the former case together but split the latter, so we |
|
454 | # We want to group the former case together but split the latter, so we | |
451 | # track parentheses level, similar to the internals of tokenize. |
|
455 | # track parentheses level, similar to the internals of tokenize. | |
452 | NEWLINE, NL = tokenize.NEWLINE, tokenize.NL |
|
456 | NEWLINE, NL = tokenize.NEWLINE, tokenize.NL | |
453 | tokens_by_line = [[]] |
|
457 | tokens_by_line = [[]] | |
454 | parenlev = 0 |
|
458 | parenlev = 0 | |
455 | try: |
|
459 | try: | |
456 | for token in tokenize.generate_tokens(iter(lines).__next__): |
|
460 | for token in tokenize.generate_tokens(iter(lines).__next__): | |
457 | tokens_by_line[-1].append(token) |
|
461 | tokens_by_line[-1].append(token) | |
458 | if (token.type == NEWLINE) \ |
|
462 | if (token.type == NEWLINE) \ | |
459 | or ((token.type == NL) and (parenlev <= 0)): |
|
463 | or ((token.type == NL) and (parenlev <= 0)): | |
460 | tokens_by_line.append([]) |
|
464 | tokens_by_line.append([]) | |
461 | elif token.string in {'(', '[', '{'}: |
|
465 | elif token.string in {'(', '[', '{'}: | |
462 | parenlev += 1 |
|
466 | parenlev += 1 | |
463 | elif token.string in {')', ']', '}'}: |
|
467 | elif token.string in {')', ']', '}'}: | |
464 | if parenlev > 0: |
|
468 | if parenlev > 0: | |
465 | parenlev -= 1 |
|
469 | parenlev -= 1 | |
466 | except tokenize.TokenError: |
|
470 | except tokenize.TokenError: | |
467 | # Input ended in a multiline string or expression. That's OK for us. |
|
471 | # Input ended in a multiline string or expression. That's OK for us. | |
468 | pass |
|
472 | pass | |
469 | if not tokens_by_line[-1]: |
|
473 | if not tokens_by_line[-1]: | |
470 | tokens_by_line.pop() |
|
474 | tokens_by_line.pop() | |
471 |
|
475 | |||
472 | return tokens_by_line |
|
476 | return tokens_by_line | |
473 |
|
477 | |||
474 | def show_linewise_tokens(s: str): |
|
478 | def show_linewise_tokens(s: str): | |
475 | """For investigation and debugging""" |
|
479 | """For investigation and debugging""" | |
476 | if not s.endswith('\n'): |
|
480 | if not s.endswith('\n'): | |
477 | s += '\n' |
|
481 | s += '\n' | |
478 | lines = s.splitlines(keepends=True) |
|
482 | lines = s.splitlines(keepends=True) | |
479 | for line in make_tokens_by_line(lines): |
|
483 | for line in make_tokens_by_line(lines): | |
480 | print("Line -------") |
|
484 | print("Line -------") | |
481 | for tokinfo in line: |
|
485 | for tokinfo in line: | |
482 | print(" ", tokinfo) |
|
486 | print(" ", tokinfo) | |
483 |
|
487 | |||
484 | # Arbitrary limit to prevent getting stuck in infinite loops |
|
488 | # Arbitrary limit to prevent getting stuck in infinite loops | |
485 | TRANSFORM_LOOP_LIMIT = 500 |
|
489 | TRANSFORM_LOOP_LIMIT = 500 | |
486 |
|
490 | |||
487 | class TransformerManager: |
|
491 | class TransformerManager: | |
488 | """Applies various transformations to a cell or code block. |
|
492 | """Applies various transformations to a cell or code block. | |
489 |
|
493 | |||
490 | The key methods for external use are ``transform_cell()`` |
|
494 | The key methods for external use are ``transform_cell()`` | |
491 | and ``check_complete()``. |
|
495 | and ``check_complete()``. | |
492 | """ |
|
496 | """ | |
493 | def __init__(self): |
|
497 | def __init__(self): | |
494 | self.cleanup_transforms = [ |
|
498 | self.cleanup_transforms = [ | |
495 | leading_indent, |
|
499 | leading_indent, | |
496 | classic_prompt, |
|
500 | classic_prompt, | |
497 | ipython_prompt, |
|
501 | ipython_prompt, | |
498 | ] |
|
502 | ] | |
499 | self.line_transforms = [ |
|
503 | self.line_transforms = [ | |
500 | cell_magic, |
|
504 | cell_magic, | |
501 | ] |
|
505 | ] | |
502 | self.token_transformers = [ |
|
506 | self.token_transformers = [ | |
503 | MagicAssign, |
|
507 | MagicAssign, | |
504 | SystemAssign, |
|
508 | SystemAssign, | |
505 | EscapedCommand, |
|
509 | EscapedCommand, | |
506 | HelpEnd, |
|
510 | HelpEnd, | |
507 | ] |
|
511 | ] | |
508 |
|
512 | |||
509 | def do_one_token_transform(self, lines): |
|
513 | def do_one_token_transform(self, lines): | |
510 | """Find and run the transform earliest in the code. |
|
514 | """Find and run the transform earliest in the code. | |
511 |
|
515 | |||
512 | Returns (changed, lines). |
|
516 | Returns (changed, lines). | |
513 |
|
517 | |||
514 | This method is called repeatedly until changed is False, indicating |
|
518 | This method is called repeatedly until changed is False, indicating | |
515 | that all available transformations are complete. |
|
519 | that all available transformations are complete. | |
516 |
|
520 | |||
517 | The tokens following IPython special syntax might not be valid, so |
|
521 | The tokens following IPython special syntax might not be valid, so | |
518 | the transformed code is retokenised every time to identify the next |
|
522 | the transformed code is retokenised every time to identify the next | |
519 | piece of special syntax. Hopefully long code cells are mostly valid |
|
523 | piece of special syntax. Hopefully long code cells are mostly valid | |
520 | Python, not using lots of IPython special syntax, so this shouldn't be |
|
524 | Python, not using lots of IPython special syntax, so this shouldn't be | |
521 | a performance issue. |
|
525 | a performance issue. | |
522 | """ |
|
526 | """ | |
523 | tokens_by_line = make_tokens_by_line(lines) |
|
527 | tokens_by_line = make_tokens_by_line(lines) | |
524 | candidates = [] |
|
528 | candidates = [] | |
525 | for transformer_cls in self.token_transformers: |
|
529 | for transformer_cls in self.token_transformers: | |
526 | transformer = transformer_cls.find(tokens_by_line) |
|
530 | transformer = transformer_cls.find(tokens_by_line) | |
527 | if transformer: |
|
531 | if transformer: | |
528 | candidates.append(transformer) |
|
532 | candidates.append(transformer) | |
529 |
|
533 | |||
530 | if not candidates: |
|
534 | if not candidates: | |
531 | # Nothing to transform |
|
535 | # Nothing to transform | |
532 | return False, lines |
|
536 | return False, lines | |
533 | ordered_transformers = sorted(candidates, key=TokenTransformBase.sortby) |
|
537 | ordered_transformers = sorted(candidates, key=TokenTransformBase.sortby) | |
534 | for transformer in ordered_transformers: |
|
538 | for transformer in ordered_transformers: | |
535 | try: |
|
539 | try: | |
536 | return True, transformer.transform(lines) |
|
540 | return True, transformer.transform(lines) | |
537 | except SyntaxError: |
|
541 | except SyntaxError: | |
538 | pass |
|
542 | pass | |
539 | return False, lines |
|
543 | return False, lines | |
540 |
|
544 | |||
541 | def do_token_transforms(self, lines): |
|
545 | def do_token_transforms(self, lines): | |
542 | for _ in range(TRANSFORM_LOOP_LIMIT): |
|
546 | for _ in range(TRANSFORM_LOOP_LIMIT): | |
543 | changed, lines = self.do_one_token_transform(lines) |
|
547 | changed, lines = self.do_one_token_transform(lines) | |
544 | if not changed: |
|
548 | if not changed: | |
545 | return lines |
|
549 | return lines | |
546 |
|
550 | |||
547 | raise RuntimeError("Input transformation still changing after " |
|
551 | raise RuntimeError("Input transformation still changing after " | |
548 | "%d iterations. Aborting." % TRANSFORM_LOOP_LIMIT) |
|
552 | "%d iterations. Aborting." % TRANSFORM_LOOP_LIMIT) | |
549 |
|
553 | |||
550 | def transform_cell(self, cell: str) -> str: |
|
554 | def transform_cell(self, cell: str) -> str: | |
551 | """Transforms a cell of input code""" |
|
555 | """Transforms a cell of input code""" | |
552 | if not cell.endswith('\n'): |
|
556 | if not cell.endswith('\n'): | |
553 | cell += '\n' # Ensure the cell has a trailing newline |
|
557 | cell += '\n' # Ensure the cell has a trailing newline | |
554 | lines = cell.splitlines(keepends=True) |
|
558 | lines = cell.splitlines(keepends=True) | |
555 | for transform in self.cleanup_transforms + self.line_transforms: |
|
559 | for transform in self.cleanup_transforms + self.line_transforms: | |
556 | lines = transform(lines) |
|
560 | lines = transform(lines) | |
557 |
|
561 | |||
558 | lines = self.do_token_transforms(lines) |
|
562 | lines = self.do_token_transforms(lines) | |
559 | return ''.join(lines) |
|
563 | return ''.join(lines) | |
560 |
|
564 | |||
561 | def check_complete(self, cell: str): |
|
565 | def check_complete(self, cell: str): | |
562 | """Return whether a block of code is ready to execute, or should be continued |
|
566 | """Return whether a block of code is ready to execute, or should be continued | |
563 |
|
567 | |||
564 | Parameters |
|
568 | Parameters | |
565 | ---------- |
|
569 | ---------- | |
566 | source : string |
|
570 | source : string | |
567 | Python input code, which can be multiline. |
|
571 | Python input code, which can be multiline. | |
568 |
|
572 | |||
569 | Returns |
|
573 | Returns | |
570 | ------- |
|
574 | ------- | |
571 | status : str |
|
575 | status : str | |
572 | One of 'complete', 'incomplete', or 'invalid' if source is not a |
|
576 | One of 'complete', 'incomplete', or 'invalid' if source is not a | |
573 | prefix of valid code. |
|
577 | prefix of valid code. | |
574 | indent_spaces : int or None |
|
578 | indent_spaces : int or None | |
575 | The number of spaces by which to indent the next line of code. If |
|
579 | The number of spaces by which to indent the next line of code. If | |
576 | status is not 'incomplete', this is None. |
|
580 | status is not 'incomplete', this is None. | |
577 | """ |
|
581 | """ | |
578 | cell += '\n' # Ensure the cell has a trailing newline |
|
|||
579 | lines = cell.splitlines(keepends=True) |
|
582 | lines = cell.splitlines(keepends=True) | |
580 | if lines[-1][:-1].endswith('\\'): |
|
583 | if not lines: | |
|
584 | return 'complete', None | |||
|
585 | ||||
|
586 | if lines[-1].endswith('\\'): | |||
581 | # Explicit backslash continuation |
|
587 | # Explicit backslash continuation | |
582 | return 'incomplete', find_last_indent(lines) |
|
588 | return 'incomplete', find_last_indent(lines) | |
583 |
|
589 | |||
584 | try: |
|
590 | try: | |
585 | for transform in self.cleanup_transforms: |
|
591 | for transform in self.cleanup_transforms: | |
586 | lines = transform(lines) |
|
592 | lines = transform(lines) | |
587 | except SyntaxError: |
|
593 | except SyntaxError: | |
588 | return 'invalid', None |
|
594 | return 'invalid', None | |
589 |
|
595 | |||
590 | if lines[0].startswith('%%'): |
|
596 | if lines[0].startswith('%%'): | |
591 | # Special case for cell magics - completion marked by blank line |
|
597 | # Special case for cell magics - completion marked by blank line | |
592 | if lines[-1].strip(): |
|
598 | if lines[-1].strip(): | |
593 | return 'incomplete', find_last_indent(lines) |
|
599 | return 'incomplete', find_last_indent(lines) | |
594 | else: |
|
600 | else: | |
595 | return 'complete', None |
|
601 | return 'complete', None | |
596 |
|
602 | |||
597 | try: |
|
603 | try: | |
598 | for transform in self.line_transforms: |
|
604 | for transform in self.line_transforms: | |
599 | lines = transform(lines) |
|
605 | lines = transform(lines) | |
600 | lines = self.do_token_transforms(lines) |
|
606 | lines = self.do_token_transforms(lines) | |
601 | except SyntaxError: |
|
607 | except SyntaxError: | |
602 | return 'invalid', None |
|
608 | return 'invalid', None | |
603 |
|
609 | |||
604 | tokens_by_line = make_tokens_by_line(lines) |
|
610 | tokens_by_line = make_tokens_by_line(lines) | |
605 | if not tokens_by_line: |
|
611 | if not tokens_by_line: | |
606 | return 'incomplete', find_last_indent(lines) |
|
612 | return 'incomplete', find_last_indent(lines) | |
|
613 | ||||
607 | if tokens_by_line[-1][-1].type != tokenize.ENDMARKER: |
|
614 | if tokens_by_line[-1][-1].type != tokenize.ENDMARKER: | |
608 | # We're in a multiline string or expression |
|
615 | # We're in a multiline string or expression | |
609 | return 'incomplete', find_last_indent(lines) |
|
616 | return 'incomplete', find_last_indent(lines) | |
610 | if len(tokens_by_line) == 1: |
|
617 | ||
|
618 | if len(tokens_by_line[-1]) == 1: | |||
611 | return 'incomplete', find_last_indent(lines) |
|
619 | return 'incomplete', find_last_indent(lines) | |
612 | # Find the last token on the previous line that's not NEWLINE or COMMENT |
|
620 | # Find the last token on the previous line that's not NEWLINE or COMMENT | |
613 |
toks_last_line = tokens_by_line[- |
|
621 | toks_last_line = tokens_by_line[-1] | |
614 |
ix = len(tok |
|
622 | ix = len(tokens_by_line) - 1 | |
615 | while ix >= 0 and toks_last_line[ix].type in {tokenize.NEWLINE, |
|
623 | ||
|
624 | ||||
|
625 | while ix >= 0 and toks_last_line[-1].type in {tokenize.NEWLINE, | |||
616 | tokenize.COMMENT}: |
|
626 | tokenize.COMMENT}: | |
617 | ix -= 1 |
|
627 | ix -= 1 | |
618 |
|
628 | if tokens_by_line[ix][-2].string == ':': | ||
619 | if toks_last_line[ix].string == ':': |
|
|||
620 | # The last line starts a block (e.g. 'if foo:') |
|
629 | # The last line starts a block (e.g. 'if foo:') | |
621 | ix = 0 |
|
630 | ix = 0 | |
622 | while toks_last_line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
631 | while toks_last_line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
623 | ix += 1 |
|
632 | ix += 1 | |
624 | indent = toks_last_line[ix].start[1] |
|
633 | indent = toks_last_line[ix].start[1] | |
625 | return 'incomplete', indent + 4 |
|
634 | return 'incomplete', indent + 4 | |
|
635 | if tokens_by_line[ix][-2].string == '\\': | |||
|
636 | if not tokens_by_line[ix][-2].line.endswith('\\'): | |||
|
637 | return 'invalid', None | |||
626 |
|
638 | |||
627 |
# If there's a blank line at the end, assume we're ready to execute |
|
639 | # If there's a blank line at the end, assume we're ready to execute | |
628 | if not lines[-1].strip(): |
|
640 | if not lines[-1].strip(): | |
629 | return 'complete', None |
|
641 | return 'complete', None | |
630 |
|
642 | |||
631 | # At this point, our checks think the code is complete (or invalid). |
|
643 | # At this point, our checks think the code is complete (or invalid). | |
632 |
# We'll use codeop.compile_command to check this with the real parser |
|
644 | # We'll use codeop.compile_command to check this with the real parser | |
633 |
|
||||
634 | try: |
|
645 | try: | |
635 | with warnings.catch_warnings(): |
|
646 | with warnings.catch_warnings(): | |
636 | warnings.simplefilter('error', SyntaxWarning) |
|
647 | warnings.simplefilter('error', SyntaxWarning) | |
637 | compile_command(''.join(lines), symbol='exec') |
|
648 | res = compile_command(''.join(lines), symbol='exec') | |
638 | except (SyntaxError, OverflowError, ValueError, TypeError, |
|
649 | except (SyntaxError, OverflowError, ValueError, TypeError, | |
639 | MemoryError, SyntaxWarning): |
|
650 | MemoryError, SyntaxWarning): | |
640 | return 'invalid', None |
|
651 | return 'invalid', None | |
641 | else: |
|
652 | else: | |
642 | if len(lines) > 1 and not lines[-1].strip().endswith(':') \ |
|
653 | if res is None: | |
643 | and not lines[-2][:-1].endswith('\\'): |
|
654 | return 'incomplete', find_last_indent(lines) | |
|
655 | ||||
|
656 | if toks_last_line[-2].type == tokenize.DEDENT: | |||
|
657 | if not lines[-1].endswith('\n'): | |||
644 | return 'incomplete', find_last_indent(lines) |
|
658 | return 'incomplete', find_last_indent(lines) | |
|
659 | ||||
645 | return 'complete', None |
|
660 | return 'complete', None | |
646 |
|
661 | |||
647 |
|
662 | |||
648 | def find_last_indent(lines): |
|
663 | def find_last_indent(lines): | |
649 | m = _indent_re.match(lines[-1]) |
|
664 | m = _indent_re.match(lines[-1]) | |
650 | if not m: |
|
665 | if not m: | |
651 | return 0 |
|
666 | return 0 | |
652 | return len(m.group(0).replace('\t', ' '*4)) |
|
667 | return len(m.group(0).replace('\t', ' '*4)) |
General Comments 0
You need to be logged in to leave comments.
Login now