Show More
@@ -1,707 +1,721 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, Union |
|
16 | from typing import List, Tuple, Union | |
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_empty_lines(lines): | |||
|
22 | """Remove leading empty lines | |||
|
23 | ||||
|
24 | If the leading lines are empty or contain only whitespace, they will be | |||
|
25 | removed. | |||
|
26 | """ | |||
|
27 | if not lines: | |||
|
28 | return lines | |||
|
29 | for i, line in enumerate(lines): | |||
|
30 | if line and not line.isspace(): | |||
|
31 | return lines[i:] | |||
|
32 | return lines | |||
|
33 | ||||
21 | def leading_indent(lines): |
|
34 | def leading_indent(lines): | |
22 | """Remove leading indentation. |
|
35 | """Remove leading indentation. | |
23 |
|
36 | |||
24 | If the first line starts with a spaces or tabs, the same whitespace will be |
|
37 | If the first line starts with a spaces or tabs, the same whitespace will be | |
25 | removed from each following line in the cell. |
|
38 | removed from each following line in the cell. | |
26 | """ |
|
39 | """ | |
27 | if not lines: |
|
40 | if not lines: | |
28 | return lines |
|
41 | return lines | |
29 | m = _indent_re.match(lines[0]) |
|
42 | m = _indent_re.match(lines[0]) | |
30 | if not m: |
|
43 | if not m: | |
31 | return lines |
|
44 | return lines | |
32 | space = m.group(0) |
|
45 | space = m.group(0) | |
33 | n = len(space) |
|
46 | n = len(space) | |
34 | return [l[n:] if l.startswith(space) else l |
|
47 | return [l[n:] if l.startswith(space) else l | |
35 | for l in lines] |
|
48 | for l in lines] | |
36 |
|
49 | |||
37 | class PromptStripper: |
|
50 | class PromptStripper: | |
38 | """Remove matching input prompts from a block of input. |
|
51 | """Remove matching input prompts from a block of input. | |
39 |
|
52 | |||
40 | Parameters |
|
53 | Parameters | |
41 | ---------- |
|
54 | ---------- | |
42 | prompt_re : regular expression |
|
55 | prompt_re : regular expression | |
43 | A regular expression matching any input prompt (including continuation, |
|
56 | A regular expression matching any input prompt (including continuation, | |
44 | e.g. ``...``) |
|
57 | e.g. ``...``) | |
45 | initial_re : regular expression, optional |
|
58 | initial_re : regular expression, optional | |
46 | A regular expression matching only the initial prompt, but not continuation. |
|
59 | A regular expression matching only the initial prompt, but not continuation. | |
47 | If no initial expression is given, prompt_re will be used everywhere. |
|
60 | If no initial expression is given, prompt_re will be used everywhere. | |
48 | Used mainly for plain Python prompts (``>>>``), where the continuation prompt |
|
61 | Used mainly for plain Python prompts (``>>>``), where the continuation prompt | |
49 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. |
|
62 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. | |
50 |
|
63 | |||
51 | If initial_re and prompt_re differ, |
|
64 | If initial_re and prompt_re differ, | |
52 | only initial_re will be tested against the first line. |
|
65 | only initial_re will be tested against the first line. | |
53 | If any prompt is found on the first two lines, |
|
66 | If any prompt is found on the first two lines, | |
54 | prompts will be stripped from the rest of the block. |
|
67 | prompts will be stripped from the rest of the block. | |
55 | """ |
|
68 | """ | |
56 | def __init__(self, prompt_re, initial_re=None): |
|
69 | def __init__(self, prompt_re, initial_re=None): | |
57 | self.prompt_re = prompt_re |
|
70 | self.prompt_re = prompt_re | |
58 | self.initial_re = initial_re or prompt_re |
|
71 | self.initial_re = initial_re or prompt_re | |
59 |
|
72 | |||
60 | def _strip(self, lines): |
|
73 | def _strip(self, lines): | |
61 | return [self.prompt_re.sub('', l, count=1) for l in lines] |
|
74 | return [self.prompt_re.sub('', l, count=1) for l in lines] | |
62 |
|
75 | |||
63 | def __call__(self, lines): |
|
76 | def __call__(self, lines): | |
64 | if not lines: |
|
77 | if not lines: | |
65 | return lines |
|
78 | return lines | |
66 | if self.initial_re.match(lines[0]) or \ |
|
79 | if self.initial_re.match(lines[0]) or \ | |
67 | (len(lines) > 1 and self.prompt_re.match(lines[1])): |
|
80 | (len(lines) > 1 and self.prompt_re.match(lines[1])): | |
68 | return self._strip(lines) |
|
81 | return self._strip(lines) | |
69 | return lines |
|
82 | return lines | |
70 |
|
83 | |||
71 | classic_prompt = PromptStripper( |
|
84 | classic_prompt = PromptStripper( | |
72 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), |
|
85 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), | |
73 | initial_re=re.compile(r'^>>>( |$)') |
|
86 | initial_re=re.compile(r'^>>>( |$)') | |
74 | ) |
|
87 | ) | |
75 |
|
88 | |||
76 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) |
|
89 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) | |
77 |
|
90 | |||
78 | def cell_magic(lines): |
|
91 | def cell_magic(lines): | |
79 | if not lines or not lines[0].startswith('%%'): |
|
92 | if not lines or not lines[0].startswith('%%'): | |
80 | return lines |
|
93 | return lines | |
81 | if re.match(r'%%\w+\?', lines[0]): |
|
94 | if re.match(r'%%\w+\?', lines[0]): | |
82 | # This case will be handled by help_end |
|
95 | # This case will be handled by help_end | |
83 | return lines |
|
96 | return lines | |
84 | magic_name, _, first_line = lines[0][2:-1].partition(' ') |
|
97 | magic_name, _, first_line = lines[0][2:-1].partition(' ') | |
85 | body = ''.join(lines[1:]) |
|
98 | body = ''.join(lines[1:]) | |
86 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' |
|
99 | return ['get_ipython().run_cell_magic(%r, %r, %r)\n' | |
87 | % (magic_name, first_line, body)] |
|
100 | % (magic_name, first_line, body)] | |
88 |
|
101 | |||
89 |
|
102 | |||
90 | def _find_assign_op(token_line) -> Union[int, None]: |
|
103 | def _find_assign_op(token_line) -> Union[int, None]: | |
91 | """Get the index of the first assignment in the line ('=' not inside brackets) |
|
104 | """Get the index of the first assignment in the line ('=' not inside brackets) | |
92 |
|
105 | |||
93 | Note: We don't try to support multiple special assignment (a = b = %foo) |
|
106 | Note: We don't try to support multiple special assignment (a = b = %foo) | |
94 | """ |
|
107 | """ | |
95 | paren_level = 0 |
|
108 | paren_level = 0 | |
96 | for i, ti in enumerate(token_line): |
|
109 | for i, ti in enumerate(token_line): | |
97 | s = ti.string |
|
110 | s = ti.string | |
98 | if s == '=' and paren_level == 0: |
|
111 | if s == '=' and paren_level == 0: | |
99 | return i |
|
112 | return i | |
100 | if s in {'(','[','{'}: |
|
113 | if s in {'(','[','{'}: | |
101 | paren_level += 1 |
|
114 | paren_level += 1 | |
102 | elif s in {')', ']', '}'}: |
|
115 | elif s in {')', ']', '}'}: | |
103 | if paren_level > 0: |
|
116 | if paren_level > 0: | |
104 | paren_level -= 1 |
|
117 | paren_level -= 1 | |
105 |
|
118 | |||
106 | def find_end_of_continued_line(lines, start_line: int): |
|
119 | def find_end_of_continued_line(lines, start_line: int): | |
107 | """Find the last line of a line explicitly extended using backslashes. |
|
120 | """Find the last line of a line explicitly extended using backslashes. | |
108 |
|
121 | |||
109 | Uses 0-indexed line numbers. |
|
122 | Uses 0-indexed line numbers. | |
110 | """ |
|
123 | """ | |
111 | end_line = start_line |
|
124 | end_line = start_line | |
112 | while lines[end_line].endswith('\\\n'): |
|
125 | while lines[end_line].endswith('\\\n'): | |
113 | end_line += 1 |
|
126 | end_line += 1 | |
114 | if end_line >= len(lines): |
|
127 | if end_line >= len(lines): | |
115 | break |
|
128 | break | |
116 | return end_line |
|
129 | return end_line | |
117 |
|
130 | |||
118 | def assemble_continued_line(lines, start: Tuple[int, int], end_line: int): |
|
131 | def assemble_continued_line(lines, start: Tuple[int, int], end_line: int): | |
119 | r"""Assemble a single line from multiple continued line pieces |
|
132 | r"""Assemble a single line from multiple continued line pieces | |
120 |
|
133 | |||
121 | Continued lines are lines ending in ``\``, and the line following the last |
|
134 | Continued lines are lines ending in ``\``, and the line following the last | |
122 | ``\`` in the block. |
|
135 | ``\`` in the block. | |
123 |
|
136 | |||
124 | For example, this code continues over multiple lines:: |
|
137 | For example, this code continues over multiple lines:: | |
125 |
|
138 | |||
126 | if (assign_ix is not None) \ |
|
139 | if (assign_ix is not None) \ | |
127 | and (len(line) >= assign_ix + 2) \ |
|
140 | and (len(line) >= assign_ix + 2) \ | |
128 | and (line[assign_ix+1].string == '%') \ |
|
141 | and (line[assign_ix+1].string == '%') \ | |
129 | and (line[assign_ix+2].type == tokenize.NAME): |
|
142 | and (line[assign_ix+2].type == tokenize.NAME): | |
130 |
|
143 | |||
131 | This statement contains four continued line pieces. |
|
144 | This statement contains four continued line pieces. | |
132 | Assembling these pieces into a single line would give:: |
|
145 | Assembling these pieces into a single line would give:: | |
133 |
|
146 | |||
134 | if (assign_ix is not None) and (len(line) >= assign_ix + 2) and (line[... |
|
147 | if (assign_ix is not None) and (len(line) >= assign_ix + 2) and (line[... | |
135 |
|
148 | |||
136 | This uses 0-indexed line numbers. *start* is (lineno, colno). |
|
149 | This uses 0-indexed line numbers. *start* is (lineno, colno). | |
137 |
|
150 | |||
138 | Used to allow ``%magic`` and ``!system`` commands to be continued over |
|
151 | Used to allow ``%magic`` and ``!system`` commands to be continued over | |
139 | multiple lines. |
|
152 | multiple lines. | |
140 | """ |
|
153 | """ | |
141 | parts = [lines[start[0]][start[1]:]] + lines[start[0]+1:end_line+1] |
|
154 | 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 |
|
155 | return ' '.join([p[:-2] for p in parts[:-1]] # Strip backslash+newline | |
143 | + [parts[-1][:-1]]) # Strip newline from last line |
|
156 | + [parts[-1][:-1]]) # Strip newline from last line | |
144 |
|
157 | |||
145 | class TokenTransformBase: |
|
158 | class TokenTransformBase: | |
146 | """Base class for transformations which examine tokens. |
|
159 | """Base class for transformations which examine tokens. | |
147 |
|
160 | |||
148 | Special syntax should not be transformed when it occurs inside strings or |
|
161 | 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 |
|
162 | 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. |
|
163 | tokenise the code as Python, and recognise the special syntax in the tokens. | |
151 |
|
164 | |||
152 | IPython's special syntax is not valid Python syntax, so tokenising may go |
|
165 | 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 |
|
166 | wrong after the special syntax starts. These classes therefore find and | |
154 | transform *one* instance of special syntax at a time into regular Python |
|
167 | transform *one* instance of special syntax at a time into regular Python | |
155 | syntax. After each transformation, tokens are regenerated to find the next |
|
168 | syntax. After each transformation, tokens are regenerated to find the next | |
156 | piece of special syntax. |
|
169 | piece of special syntax. | |
157 |
|
170 | |||
158 | Subclasses need to implement one class method (find) |
|
171 | Subclasses need to implement one class method (find) | |
159 | and one regular method (transform). |
|
172 | and one regular method (transform). | |
160 |
|
173 | |||
161 | The priority attribute can select which transformation to apply if multiple |
|
174 | The priority attribute can select which transformation to apply if multiple | |
162 | transformers match in the same place. Lower numbers have higher priority. |
|
175 | 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. |
|
176 | This allows "%magic?" to be turned into a help call rather than a magic call. | |
164 | """ |
|
177 | """ | |
165 | # Lower numbers -> higher priority (for matches in the same location) |
|
178 | # Lower numbers -> higher priority (for matches in the same location) | |
166 | priority = 10 |
|
179 | priority = 10 | |
167 |
|
180 | |||
168 | def sortby(self): |
|
181 | def sortby(self): | |
169 | return self.start_line, self.start_col, self.priority |
|
182 | return self.start_line, self.start_col, self.priority | |
170 |
|
183 | |||
171 | def __init__(self, start): |
|
184 | def __init__(self, start): | |
172 | self.start_line = start[0] - 1 # Shift from 1-index to 0-index |
|
185 | self.start_line = start[0] - 1 # Shift from 1-index to 0-index | |
173 | self.start_col = start[1] |
|
186 | self.start_col = start[1] | |
174 |
|
187 | |||
175 | @classmethod |
|
188 | @classmethod | |
176 | def find(cls, tokens_by_line): |
|
189 | def find(cls, tokens_by_line): | |
177 | """Find one instance of special syntax in the provided tokens. |
|
190 | """Find one instance of special syntax in the provided tokens. | |
178 |
|
191 | |||
179 | Tokens are grouped into logical lines for convenience, |
|
192 | Tokens are grouped into logical lines for convenience, | |
180 | so it is easy to e.g. look at the first token of each line. |
|
193 | 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. |
|
194 | *tokens_by_line* is a list of lists of tokenize.TokenInfo objects. | |
182 |
|
195 | |||
183 | This should return an instance of its class, pointing to the start |
|
196 | This should return an instance of its class, pointing to the start | |
184 | position it has found, or None if it found no match. |
|
197 | position it has found, or None if it found no match. | |
185 | """ |
|
198 | """ | |
186 | raise NotImplementedError |
|
199 | raise NotImplementedError | |
187 |
|
200 | |||
188 | def transform(self, lines: List[str]): |
|
201 | def transform(self, lines: List[str]): | |
189 | """Transform one instance of special syntax found by ``find()`` |
|
202 | """Transform one instance of special syntax found by ``find()`` | |
190 |
|
203 | |||
191 | Takes a list of strings representing physical lines, |
|
204 | Takes a list of strings representing physical lines, | |
192 | returns a similar list of transformed lines. |
|
205 | returns a similar list of transformed lines. | |
193 | """ |
|
206 | """ | |
194 | raise NotImplementedError |
|
207 | raise NotImplementedError | |
195 |
|
208 | |||
196 | class MagicAssign(TokenTransformBase): |
|
209 | class MagicAssign(TokenTransformBase): | |
197 | """Transformer for assignments from magics (a = %foo)""" |
|
210 | """Transformer for assignments from magics (a = %foo)""" | |
198 | @classmethod |
|
211 | @classmethod | |
199 | def find(cls, tokens_by_line): |
|
212 | def find(cls, tokens_by_line): | |
200 | """Find the first magic assignment (a = %foo) in the cell. |
|
213 | """Find the first magic assignment (a = %foo) in the cell. | |
201 | """ |
|
214 | """ | |
202 | for line in tokens_by_line: |
|
215 | for line in tokens_by_line: | |
203 | assign_ix = _find_assign_op(line) |
|
216 | assign_ix = _find_assign_op(line) | |
204 | if (assign_ix is not None) \ |
|
217 | if (assign_ix is not None) \ | |
205 | and (len(line) >= assign_ix + 2) \ |
|
218 | and (len(line) >= assign_ix + 2) \ | |
206 | and (line[assign_ix+1].string == '%') \ |
|
219 | and (line[assign_ix+1].string == '%') \ | |
207 | and (line[assign_ix+2].type == tokenize.NAME): |
|
220 | and (line[assign_ix+2].type == tokenize.NAME): | |
208 | return cls(line[assign_ix+1].start) |
|
221 | return cls(line[assign_ix+1].start) | |
209 |
|
222 | |||
210 | def transform(self, lines: List[str]): |
|
223 | def transform(self, lines: List[str]): | |
211 | """Transform a magic assignment found by the ``find()`` classmethod. |
|
224 | """Transform a magic assignment found by the ``find()`` classmethod. | |
212 | """ |
|
225 | """ | |
213 | start_line, start_col = self.start_line, self.start_col |
|
226 | start_line, start_col = self.start_line, self.start_col | |
214 | lhs = lines[start_line][:start_col] |
|
227 | lhs = lines[start_line][:start_col] | |
215 | end_line = find_end_of_continued_line(lines, start_line) |
|
228 | end_line = find_end_of_continued_line(lines, start_line) | |
216 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
229 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) | |
217 | assert rhs.startswith('%'), rhs |
|
230 | assert rhs.startswith('%'), rhs | |
218 | magic_name, _, args = rhs[1:].partition(' ') |
|
231 | magic_name, _, args = rhs[1:].partition(' ') | |
219 |
|
232 | |||
220 | lines_before = lines[:start_line] |
|
233 | lines_before = lines[:start_line] | |
221 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) |
|
234 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) | |
222 | new_line = lhs + call + '\n' |
|
235 | new_line = lhs + call + '\n' | |
223 | lines_after = lines[end_line+1:] |
|
236 | lines_after = lines[end_line+1:] | |
224 |
|
237 | |||
225 | return lines_before + [new_line] + lines_after |
|
238 | return lines_before + [new_line] + lines_after | |
226 |
|
239 | |||
227 |
|
240 | |||
228 | class SystemAssign(TokenTransformBase): |
|
241 | class SystemAssign(TokenTransformBase): | |
229 | """Transformer for assignments from system commands (a = !foo)""" |
|
242 | """Transformer for assignments from system commands (a = !foo)""" | |
230 | @classmethod |
|
243 | @classmethod | |
231 | def find(cls, tokens_by_line): |
|
244 | def find(cls, tokens_by_line): | |
232 | """Find the first system assignment (a = !foo) in the cell. |
|
245 | """Find the first system assignment (a = !foo) in the cell. | |
233 | """ |
|
246 | """ | |
234 | for line in tokens_by_line: |
|
247 | for line in tokens_by_line: | |
235 | assign_ix = _find_assign_op(line) |
|
248 | assign_ix = _find_assign_op(line) | |
236 | if (assign_ix is not None) \ |
|
249 | if (assign_ix is not None) \ | |
237 | and not line[assign_ix].line.strip().startswith('=') \ |
|
250 | and not line[assign_ix].line.strip().startswith('=') \ | |
238 | and (len(line) >= assign_ix + 2) \ |
|
251 | and (len(line) >= assign_ix + 2) \ | |
239 | and (line[assign_ix + 1].type == tokenize.ERRORTOKEN): |
|
252 | and (line[assign_ix + 1].type == tokenize.ERRORTOKEN): | |
240 | ix = assign_ix + 1 |
|
253 | ix = assign_ix + 1 | |
241 |
|
254 | |||
242 | while ix < len(line) and line[ix].type == tokenize.ERRORTOKEN: |
|
255 | while ix < len(line) and line[ix].type == tokenize.ERRORTOKEN: | |
243 | if line[ix].string == '!': |
|
256 | if line[ix].string == '!': | |
244 | return cls(line[ix].start) |
|
257 | return cls(line[ix].start) | |
245 | elif not line[ix].string.isspace(): |
|
258 | elif not line[ix].string.isspace(): | |
246 | break |
|
259 | break | |
247 | ix += 1 |
|
260 | ix += 1 | |
248 |
|
261 | |||
249 | def transform(self, lines: List[str]): |
|
262 | def transform(self, lines: List[str]): | |
250 | """Transform a system assignment found by the ``find()`` classmethod. |
|
263 | """Transform a system assignment found by the ``find()`` classmethod. | |
251 | """ |
|
264 | """ | |
252 | start_line, start_col = self.start_line, self.start_col |
|
265 | start_line, start_col = self.start_line, self.start_col | |
253 |
|
266 | |||
254 | lhs = lines[start_line][:start_col] |
|
267 | lhs = lines[start_line][:start_col] | |
255 | end_line = find_end_of_continued_line(lines, start_line) |
|
268 | end_line = find_end_of_continued_line(lines, start_line) | |
256 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
269 | rhs = assemble_continued_line(lines, (start_line, start_col), end_line) | |
257 | assert rhs.startswith('!'), rhs |
|
270 | assert rhs.startswith('!'), rhs | |
258 | cmd = rhs[1:] |
|
271 | cmd = rhs[1:] | |
259 |
|
272 | |||
260 | lines_before = lines[:start_line] |
|
273 | lines_before = lines[:start_line] | |
261 | call = "get_ipython().getoutput({!r})".format(cmd) |
|
274 | call = "get_ipython().getoutput({!r})".format(cmd) | |
262 | new_line = lhs + call + '\n' |
|
275 | new_line = lhs + call + '\n' | |
263 | lines_after = lines[end_line + 1:] |
|
276 | lines_after = lines[end_line + 1:] | |
264 |
|
277 | |||
265 | return lines_before + [new_line] + lines_after |
|
278 | return lines_before + [new_line] + lines_after | |
266 |
|
279 | |||
267 | # The escape sequences that define the syntax transformations IPython will |
|
280 | # The escape sequences that define the syntax transformations IPython will | |
268 | # apply to user input. These can NOT be just changed here: many regular |
|
281 | # apply to user input. These can NOT be just changed here: many regular | |
269 | # expressions and other parts of the code may use their hardcoded values, and |
|
282 | # expressions and other parts of the code may use their hardcoded values, and | |
270 | # for all intents and purposes they constitute the 'IPython syntax', so they |
|
283 | # for all intents and purposes they constitute the 'IPython syntax', so they | |
271 | # should be considered fixed. |
|
284 | # should be considered fixed. | |
272 |
|
285 | |||
273 | ESC_SHELL = '!' # Send line to underlying system shell |
|
286 | ESC_SHELL = '!' # Send line to underlying system shell | |
274 | ESC_SH_CAP = '!!' # Send line to system shell and capture output |
|
287 | ESC_SH_CAP = '!!' # Send line to system shell and capture output | |
275 | ESC_HELP = '?' # Find information about object |
|
288 | ESC_HELP = '?' # Find information about object | |
276 | ESC_HELP2 = '??' # Find extra-detailed information about object |
|
289 | ESC_HELP2 = '??' # Find extra-detailed information about object | |
277 | ESC_MAGIC = '%' # Call magic function |
|
290 | ESC_MAGIC = '%' # Call magic function | |
278 | ESC_MAGIC2 = '%%' # Call cell-magic function |
|
291 | ESC_MAGIC2 = '%%' # Call cell-magic function | |
279 | ESC_QUOTE = ',' # Split args on whitespace, quote each as string and call |
|
292 | ESC_QUOTE = ',' # Split args on whitespace, quote each as string and call | |
280 | ESC_QUOTE2 = ';' # Quote all args as a single string, call |
|
293 | ESC_QUOTE2 = ';' # Quote all args as a single string, call | |
281 | ESC_PAREN = '/' # Call first argument with rest of line as arguments |
|
294 | ESC_PAREN = '/' # Call first argument with rest of line as arguments | |
282 |
|
295 | |||
283 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} |
|
296 | ESCAPE_SINGLES = {'!', '?', '%', ',', ';', '/'} | |
284 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately |
|
297 | ESCAPE_DOUBLES = {'!!', '??'} # %% (cell magic) is handled separately | |
285 |
|
298 | |||
286 | def _make_help_call(target, esc, next_input=None): |
|
299 | def _make_help_call(target, esc, next_input=None): | |
287 | """Prepares a pinfo(2)/psearch call from a target name and the escape |
|
300 | """Prepares a pinfo(2)/psearch call from a target name and the escape | |
288 | (i.e. ? or ??)""" |
|
301 | (i.e. ? or ??)""" | |
289 | method = 'pinfo2' if esc == '??' \ |
|
302 | method = 'pinfo2' if esc == '??' \ | |
290 | else 'psearch' if '*' in target \ |
|
303 | else 'psearch' if '*' in target \ | |
291 | else 'pinfo' |
|
304 | else 'pinfo' | |
292 | arg = " ".join([method, target]) |
|
305 | arg = " ".join([method, target]) | |
293 | #Prepare arguments for get_ipython().run_line_magic(magic_name, magic_args) |
|
306 | #Prepare arguments for get_ipython().run_line_magic(magic_name, magic_args) | |
294 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') |
|
307 | t_magic_name, _, t_magic_arg_s = arg.partition(' ') | |
295 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) |
|
308 | t_magic_name = t_magic_name.lstrip(ESC_MAGIC) | |
296 | if next_input is None: |
|
309 | if next_input is None: | |
297 | return 'get_ipython().run_line_magic(%r, %r)' % (t_magic_name, t_magic_arg_s) |
|
310 | return 'get_ipython().run_line_magic(%r, %r)' % (t_magic_name, t_magic_arg_s) | |
298 | else: |
|
311 | else: | |
299 | return 'get_ipython().set_next_input(%r);get_ipython().run_line_magic(%r, %r)' % \ |
|
312 | return 'get_ipython().set_next_input(%r);get_ipython().run_line_magic(%r, %r)' % \ | |
300 | (next_input, t_magic_name, t_magic_arg_s) |
|
313 | (next_input, t_magic_name, t_magic_arg_s) | |
301 |
|
314 | |||
302 | def _tr_help(content): |
|
315 | def _tr_help(content): | |
303 | """Translate lines escaped with: ? |
|
316 | """Translate lines escaped with: ? | |
304 |
|
317 | |||
305 | A naked help line should fire the intro help screen (shell.show_usage()) |
|
318 | A naked help line should fire the intro help screen (shell.show_usage()) | |
306 | """ |
|
319 | """ | |
307 | if not content: |
|
320 | if not content: | |
308 | return 'get_ipython().show_usage()' |
|
321 | return 'get_ipython().show_usage()' | |
309 |
|
322 | |||
310 | return _make_help_call(content, '?') |
|
323 | return _make_help_call(content, '?') | |
311 |
|
324 | |||
312 | def _tr_help2(content): |
|
325 | def _tr_help2(content): | |
313 | """Translate lines escaped with: ?? |
|
326 | """Translate lines escaped with: ?? | |
314 |
|
327 | |||
315 | A naked help line should fire the intro help screen (shell.show_usage()) |
|
328 | A naked help line should fire the intro help screen (shell.show_usage()) | |
316 | """ |
|
329 | """ | |
317 | if not content: |
|
330 | if not content: | |
318 | return 'get_ipython().show_usage()' |
|
331 | return 'get_ipython().show_usage()' | |
319 |
|
332 | |||
320 | return _make_help_call(content, '??') |
|
333 | return _make_help_call(content, '??') | |
321 |
|
334 | |||
322 | def _tr_magic(content): |
|
335 | def _tr_magic(content): | |
323 | "Translate lines escaped with a percent sign: %" |
|
336 | "Translate lines escaped with a percent sign: %" | |
324 | name, _, args = content.partition(' ') |
|
337 | name, _, args = content.partition(' ') | |
325 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) |
|
338 | return 'get_ipython().run_line_magic(%r, %r)' % (name, args) | |
326 |
|
339 | |||
327 | def _tr_quote(content): |
|
340 | def _tr_quote(content): | |
328 | "Translate lines escaped with a comma: ," |
|
341 | "Translate lines escaped with a comma: ," | |
329 | name, _, args = content.partition(' ') |
|
342 | name, _, args = content.partition(' ') | |
330 | return '%s("%s")' % (name, '", "'.join(args.split()) ) |
|
343 | return '%s("%s")' % (name, '", "'.join(args.split()) ) | |
331 |
|
344 | |||
332 | def _tr_quote2(content): |
|
345 | def _tr_quote2(content): | |
333 | "Translate lines escaped with a semicolon: ;" |
|
346 | "Translate lines escaped with a semicolon: ;" | |
334 | name, _, args = content.partition(' ') |
|
347 | name, _, args = content.partition(' ') | |
335 | return '%s("%s")' % (name, args) |
|
348 | return '%s("%s")' % (name, args) | |
336 |
|
349 | |||
337 | def _tr_paren(content): |
|
350 | def _tr_paren(content): | |
338 | "Translate lines escaped with a slash: /" |
|
351 | "Translate lines escaped with a slash: /" | |
339 | name, _, args = content.partition(' ') |
|
352 | name, _, args = content.partition(' ') | |
340 | return '%s(%s)' % (name, ", ".join(args.split())) |
|
353 | return '%s(%s)' % (name, ", ".join(args.split())) | |
341 |
|
354 | |||
342 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, |
|
355 | tr = { ESC_SHELL : 'get_ipython().system({!r})'.format, | |
343 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, |
|
356 | ESC_SH_CAP : 'get_ipython().getoutput({!r})'.format, | |
344 | ESC_HELP : _tr_help, |
|
357 | ESC_HELP : _tr_help, | |
345 | ESC_HELP2 : _tr_help2, |
|
358 | ESC_HELP2 : _tr_help2, | |
346 | ESC_MAGIC : _tr_magic, |
|
359 | ESC_MAGIC : _tr_magic, | |
347 | ESC_QUOTE : _tr_quote, |
|
360 | ESC_QUOTE : _tr_quote, | |
348 | ESC_QUOTE2 : _tr_quote2, |
|
361 | ESC_QUOTE2 : _tr_quote2, | |
349 | ESC_PAREN : _tr_paren } |
|
362 | ESC_PAREN : _tr_paren } | |
350 |
|
363 | |||
351 | class EscapedCommand(TokenTransformBase): |
|
364 | class EscapedCommand(TokenTransformBase): | |
352 | """Transformer for escaped commands like %foo, !foo, or /foo""" |
|
365 | """Transformer for escaped commands like %foo, !foo, or /foo""" | |
353 | @classmethod |
|
366 | @classmethod | |
354 | def find(cls, tokens_by_line): |
|
367 | def find(cls, tokens_by_line): | |
355 | """Find the first escaped command (%foo, !foo, etc.) in the cell. |
|
368 | """Find the first escaped command (%foo, !foo, etc.) in the cell. | |
356 | """ |
|
369 | """ | |
357 | for line in tokens_by_line: |
|
370 | for line in tokens_by_line: | |
358 | if not line: |
|
371 | if not line: | |
359 | continue |
|
372 | continue | |
360 | ix = 0 |
|
373 | ix = 0 | |
361 | ll = len(line) |
|
374 | ll = len(line) | |
362 | while ll > ix and line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
375 | while ll > ix and line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
363 | ix += 1 |
|
376 | ix += 1 | |
364 | if ix >= ll: |
|
377 | if ix >= ll: | |
365 | continue |
|
378 | continue | |
366 | if line[ix].string in ESCAPE_SINGLES: |
|
379 | if line[ix].string in ESCAPE_SINGLES: | |
367 | return cls(line[ix].start) |
|
380 | return cls(line[ix].start) | |
368 |
|
381 | |||
369 | def transform(self, lines): |
|
382 | def transform(self, lines): | |
370 | """Transform an escaped line found by the ``find()`` classmethod. |
|
383 | """Transform an escaped line found by the ``find()`` classmethod. | |
371 | """ |
|
384 | """ | |
372 | start_line, start_col = self.start_line, self.start_col |
|
385 | start_line, start_col = self.start_line, self.start_col | |
373 |
|
386 | |||
374 | indent = lines[start_line][:start_col] |
|
387 | indent = lines[start_line][:start_col] | |
375 | end_line = find_end_of_continued_line(lines, start_line) |
|
388 | end_line = find_end_of_continued_line(lines, start_line) | |
376 | line = assemble_continued_line(lines, (start_line, start_col), end_line) |
|
389 | line = assemble_continued_line(lines, (start_line, start_col), end_line) | |
377 |
|
390 | |||
378 | if len(line) > 1 and line[:2] in ESCAPE_DOUBLES: |
|
391 | if len(line) > 1 and line[:2] in ESCAPE_DOUBLES: | |
379 | escape, content = line[:2], line[2:] |
|
392 | escape, content = line[:2], line[2:] | |
380 | else: |
|
393 | else: | |
381 | escape, content = line[:1], line[1:] |
|
394 | escape, content = line[:1], line[1:] | |
382 |
|
395 | |||
383 | if escape in tr: |
|
396 | if escape in tr: | |
384 | call = tr[escape](content) |
|
397 | call = tr[escape](content) | |
385 | else: |
|
398 | else: | |
386 | call = '' |
|
399 | call = '' | |
387 |
|
400 | |||
388 | lines_before = lines[:start_line] |
|
401 | lines_before = lines[:start_line] | |
389 | new_line = indent + call + '\n' |
|
402 | new_line = indent + call + '\n' | |
390 | lines_after = lines[end_line + 1:] |
|
403 | lines_after = lines[end_line + 1:] | |
391 |
|
404 | |||
392 | return lines_before + [new_line] + lines_after |
|
405 | return lines_before + [new_line] + lines_after | |
393 |
|
406 | |||
394 | _help_end_re = re.compile(r"""(%{0,2} |
|
407 | _help_end_re = re.compile(r"""(%{0,2} | |
395 | [a-zA-Z_*][\w*]* # Variable name |
|
408 | [a-zA-Z_*][\w*]* # Variable name | |
396 | (\.[a-zA-Z_*][\w*]*)* # .etc.etc |
|
409 | (\.[a-zA-Z_*][\w*]*)* # .etc.etc | |
397 | ) |
|
410 | ) | |
398 | (\?\??)$ # ? or ?? |
|
411 | (\?\??)$ # ? or ?? | |
399 | """, |
|
412 | """, | |
400 | re.VERBOSE) |
|
413 | re.VERBOSE) | |
401 |
|
414 | |||
402 | class HelpEnd(TokenTransformBase): |
|
415 | class HelpEnd(TokenTransformBase): | |
403 | """Transformer for help syntax: obj? and obj??""" |
|
416 | """Transformer for help syntax: obj? and obj??""" | |
404 | # This needs to be higher priority (lower number) than EscapedCommand so |
|
417 | # This needs to be higher priority (lower number) than EscapedCommand so | |
405 | # that inspecting magics (%foo?) works. |
|
418 | # that inspecting magics (%foo?) works. | |
406 | priority = 5 |
|
419 | priority = 5 | |
407 |
|
420 | |||
408 | def __init__(self, start, q_locn): |
|
421 | def __init__(self, start, q_locn): | |
409 | super().__init__(start) |
|
422 | super().__init__(start) | |
410 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed |
|
423 | self.q_line = q_locn[0] - 1 # Shift from 1-indexed to 0-indexed | |
411 | self.q_col = q_locn[1] |
|
424 | self.q_col = q_locn[1] | |
412 |
|
425 | |||
413 | @classmethod |
|
426 | @classmethod | |
414 | def find(cls, tokens_by_line): |
|
427 | def find(cls, tokens_by_line): | |
415 | """Find the first help command (foo?) in the cell. |
|
428 | """Find the first help command (foo?) in the cell. | |
416 | """ |
|
429 | """ | |
417 | for line in tokens_by_line: |
|
430 | for line in tokens_by_line: | |
418 | # Last token is NEWLINE; look at last but one |
|
431 | # Last token is NEWLINE; look at last but one | |
419 | if len(line) > 2 and line[-2].string == '?': |
|
432 | if len(line) > 2 and line[-2].string == '?': | |
420 | # Find the first token that's not INDENT/DEDENT |
|
433 | # Find the first token that's not INDENT/DEDENT | |
421 | ix = 0 |
|
434 | ix = 0 | |
422 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
435 | while line[ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
423 | ix += 1 |
|
436 | ix += 1 | |
424 | return cls(line[ix].start, line[-2].start) |
|
437 | return cls(line[ix].start, line[-2].start) | |
425 |
|
438 | |||
426 | def transform(self, lines): |
|
439 | def transform(self, lines): | |
427 | """Transform a help command found by the ``find()`` classmethod. |
|
440 | """Transform a help command found by the ``find()`` classmethod. | |
428 | """ |
|
441 | """ | |
429 | piece = ''.join(lines[self.start_line:self.q_line+1]) |
|
442 | piece = ''.join(lines[self.start_line:self.q_line+1]) | |
430 | indent, content = piece[:self.start_col], piece[self.start_col:] |
|
443 | indent, content = piece[:self.start_col], piece[self.start_col:] | |
431 | lines_before = lines[:self.start_line] |
|
444 | lines_before = lines[:self.start_line] | |
432 | lines_after = lines[self.q_line + 1:] |
|
445 | lines_after = lines[self.q_line + 1:] | |
433 |
|
446 | |||
434 | m = _help_end_re.search(content) |
|
447 | m = _help_end_re.search(content) | |
435 | if not m: |
|
448 | if not m: | |
436 | raise SyntaxError(content) |
|
449 | raise SyntaxError(content) | |
437 | assert m is not None, content |
|
450 | assert m is not None, content | |
438 | target = m.group(1) |
|
451 | target = m.group(1) | |
439 | esc = m.group(3) |
|
452 | esc = m.group(3) | |
440 |
|
453 | |||
441 | # If we're mid-command, put it back on the next prompt for the user. |
|
454 | # If we're mid-command, put it back on the next prompt for the user. | |
442 | next_input = None |
|
455 | next_input = None | |
443 | if (not lines_before) and (not lines_after) \ |
|
456 | if (not lines_before) and (not lines_after) \ | |
444 | and content.strip() != m.group(0): |
|
457 | and content.strip() != m.group(0): | |
445 | next_input = content.rstrip('?\n') |
|
458 | next_input = content.rstrip('?\n') | |
446 |
|
459 | |||
447 | call = _make_help_call(target, esc, next_input=next_input) |
|
460 | call = _make_help_call(target, esc, next_input=next_input) | |
448 | new_line = indent + call + '\n' |
|
461 | new_line = indent + call + '\n' | |
449 |
|
462 | |||
450 | return lines_before + [new_line] + lines_after |
|
463 | return lines_before + [new_line] + lines_after | |
451 |
|
464 | |||
452 | def make_tokens_by_line(lines:List[str]): |
|
465 | def make_tokens_by_line(lines:List[str]): | |
453 | """Tokenize a series of lines and group tokens by line. |
|
466 | """Tokenize a series of lines and group tokens by line. | |
454 |
|
467 | |||
455 | The tokens for a multiline Python string or expression are grouped as one |
|
468 | The tokens for a multiline Python string or expression are grouped as one | |
456 | line. All lines except the last lines should keep their line ending ('\\n', |
|
469 | line. All lines except the last lines should keep their line ending ('\\n', | |
457 | '\\r\\n') for this to properly work. Use `.splitlines(keeplineending=True)` |
|
470 | '\\r\\n') for this to properly work. Use `.splitlines(keeplineending=True)` | |
458 | for example when passing block of text to this function. |
|
471 | for example when passing block of text to this function. | |
459 |
|
472 | |||
460 | """ |
|
473 | """ | |
461 | # NL tokens are used inside multiline expressions, but also after blank |
|
474 | # NL tokens are used inside multiline expressions, but also after blank | |
462 | # lines or comments. This is intentional - see https://bugs.python.org/issue17061 |
|
475 | # lines or comments. This is intentional - see https://bugs.python.org/issue17061 | |
463 | # We want to group the former case together but split the latter, so we |
|
476 | # We want to group the former case together but split the latter, so we | |
464 | # track parentheses level, similar to the internals of tokenize. |
|
477 | # track parentheses level, similar to the internals of tokenize. | |
465 | NEWLINE, NL = tokenize.NEWLINE, tokenize.NL |
|
478 | NEWLINE, NL = tokenize.NEWLINE, tokenize.NL | |
466 | tokens_by_line = [[]] |
|
479 | tokens_by_line = [[]] | |
467 | if len(lines) > 1 and not lines[0].endswith(('\n', '\r', '\r\n', '\x0b', '\x0c')): |
|
480 | if len(lines) > 1 and not lines[0].endswith(('\n', '\r', '\r\n', '\x0b', '\x0c')): | |
468 | warnings.warn("`make_tokens_by_line` received a list of lines which do not have lineending markers ('\\n', '\\r', '\\r\\n', '\\x0b', '\\x0c'), behavior will be unspecified") |
|
481 | warnings.warn("`make_tokens_by_line` received a list of lines which do not have lineending markers ('\\n', '\\r', '\\r\\n', '\\x0b', '\\x0c'), behavior will be unspecified") | |
469 | parenlev = 0 |
|
482 | parenlev = 0 | |
470 | try: |
|
483 | try: | |
471 | for token in tokenize.generate_tokens(iter(lines).__next__): |
|
484 | for token in tokenize.generate_tokens(iter(lines).__next__): | |
472 | tokens_by_line[-1].append(token) |
|
485 | tokens_by_line[-1].append(token) | |
473 | if (token.type == NEWLINE) \ |
|
486 | if (token.type == NEWLINE) \ | |
474 | or ((token.type == NL) and (parenlev <= 0)): |
|
487 | or ((token.type == NL) and (parenlev <= 0)): | |
475 | tokens_by_line.append([]) |
|
488 | tokens_by_line.append([]) | |
476 | elif token.string in {'(', '[', '{'}: |
|
489 | elif token.string in {'(', '[', '{'}: | |
477 | parenlev += 1 |
|
490 | parenlev += 1 | |
478 | elif token.string in {')', ']', '}'}: |
|
491 | elif token.string in {')', ']', '}'}: | |
479 | if parenlev > 0: |
|
492 | if parenlev > 0: | |
480 | parenlev -= 1 |
|
493 | parenlev -= 1 | |
481 | except tokenize.TokenError: |
|
494 | except tokenize.TokenError: | |
482 | # Input ended in a multiline string or expression. That's OK for us. |
|
495 | # Input ended in a multiline string or expression. That's OK for us. | |
483 | pass |
|
496 | pass | |
484 |
|
497 | |||
485 |
|
498 | |||
486 | if not tokens_by_line[-1]: |
|
499 | if not tokens_by_line[-1]: | |
487 | tokens_by_line.pop() |
|
500 | tokens_by_line.pop() | |
488 |
|
501 | |||
489 |
|
502 | |||
490 | return tokens_by_line |
|
503 | return tokens_by_line | |
491 |
|
504 | |||
492 | def show_linewise_tokens(s: str): |
|
505 | def show_linewise_tokens(s: str): | |
493 | """For investigation and debugging""" |
|
506 | """For investigation and debugging""" | |
494 | if not s.endswith('\n'): |
|
507 | if not s.endswith('\n'): | |
495 | s += '\n' |
|
508 | s += '\n' | |
496 | lines = s.splitlines(keepends=True) |
|
509 | lines = s.splitlines(keepends=True) | |
497 | for line in make_tokens_by_line(lines): |
|
510 | for line in make_tokens_by_line(lines): | |
498 | print("Line -------") |
|
511 | print("Line -------") | |
499 | for tokinfo in line: |
|
512 | for tokinfo in line: | |
500 | print(" ", tokinfo) |
|
513 | print(" ", tokinfo) | |
501 |
|
514 | |||
502 | # Arbitrary limit to prevent getting stuck in infinite loops |
|
515 | # Arbitrary limit to prevent getting stuck in infinite loops | |
503 | TRANSFORM_LOOP_LIMIT = 500 |
|
516 | TRANSFORM_LOOP_LIMIT = 500 | |
504 |
|
517 | |||
505 | class TransformerManager: |
|
518 | class TransformerManager: | |
506 | """Applies various transformations to a cell or code block. |
|
519 | """Applies various transformations to a cell or code block. | |
507 |
|
520 | |||
508 | The key methods for external use are ``transform_cell()`` |
|
521 | The key methods for external use are ``transform_cell()`` | |
509 | and ``check_complete()``. |
|
522 | and ``check_complete()``. | |
510 | """ |
|
523 | """ | |
511 | def __init__(self): |
|
524 | def __init__(self): | |
512 | self.cleanup_transforms = [ |
|
525 | self.cleanup_transforms = [ | |
|
526 | leading_empty_lines, | |||
513 | leading_indent, |
|
527 | leading_indent, | |
514 | classic_prompt, |
|
528 | classic_prompt, | |
515 | ipython_prompt, |
|
529 | ipython_prompt, | |
516 | ] |
|
530 | ] | |
517 | self.line_transforms = [ |
|
531 | self.line_transforms = [ | |
518 | cell_magic, |
|
532 | cell_magic, | |
519 | ] |
|
533 | ] | |
520 | self.token_transformers = [ |
|
534 | self.token_transformers = [ | |
521 | MagicAssign, |
|
535 | MagicAssign, | |
522 | SystemAssign, |
|
536 | SystemAssign, | |
523 | EscapedCommand, |
|
537 | EscapedCommand, | |
524 | HelpEnd, |
|
538 | HelpEnd, | |
525 | ] |
|
539 | ] | |
526 |
|
540 | |||
527 | def do_one_token_transform(self, lines): |
|
541 | def do_one_token_transform(self, lines): | |
528 | """Find and run the transform earliest in the code. |
|
542 | """Find and run the transform earliest in the code. | |
529 |
|
543 | |||
530 | Returns (changed, lines). |
|
544 | Returns (changed, lines). | |
531 |
|
545 | |||
532 | This method is called repeatedly until changed is False, indicating |
|
546 | This method is called repeatedly until changed is False, indicating | |
533 | that all available transformations are complete. |
|
547 | that all available transformations are complete. | |
534 |
|
548 | |||
535 | The tokens following IPython special syntax might not be valid, so |
|
549 | The tokens following IPython special syntax might not be valid, so | |
536 | the transformed code is retokenised every time to identify the next |
|
550 | the transformed code is retokenised every time to identify the next | |
537 | piece of special syntax. Hopefully long code cells are mostly valid |
|
551 | piece of special syntax. Hopefully long code cells are mostly valid | |
538 | Python, not using lots of IPython special syntax, so this shouldn't be |
|
552 | Python, not using lots of IPython special syntax, so this shouldn't be | |
539 | a performance issue. |
|
553 | a performance issue. | |
540 | """ |
|
554 | """ | |
541 | tokens_by_line = make_tokens_by_line(lines) |
|
555 | tokens_by_line = make_tokens_by_line(lines) | |
542 | candidates = [] |
|
556 | candidates = [] | |
543 | for transformer_cls in self.token_transformers: |
|
557 | for transformer_cls in self.token_transformers: | |
544 | transformer = transformer_cls.find(tokens_by_line) |
|
558 | transformer = transformer_cls.find(tokens_by_line) | |
545 | if transformer: |
|
559 | if transformer: | |
546 | candidates.append(transformer) |
|
560 | candidates.append(transformer) | |
547 |
|
561 | |||
548 | if not candidates: |
|
562 | if not candidates: | |
549 | # Nothing to transform |
|
563 | # Nothing to transform | |
550 | return False, lines |
|
564 | return False, lines | |
551 | ordered_transformers = sorted(candidates, key=TokenTransformBase.sortby) |
|
565 | ordered_transformers = sorted(candidates, key=TokenTransformBase.sortby) | |
552 | for transformer in ordered_transformers: |
|
566 | for transformer in ordered_transformers: | |
553 | try: |
|
567 | try: | |
554 | return True, transformer.transform(lines) |
|
568 | return True, transformer.transform(lines) | |
555 | except SyntaxError: |
|
569 | except SyntaxError: | |
556 | pass |
|
570 | pass | |
557 | return False, lines |
|
571 | return False, lines | |
558 |
|
572 | |||
559 | def do_token_transforms(self, lines): |
|
573 | def do_token_transforms(self, lines): | |
560 | for _ in range(TRANSFORM_LOOP_LIMIT): |
|
574 | for _ in range(TRANSFORM_LOOP_LIMIT): | |
561 | changed, lines = self.do_one_token_transform(lines) |
|
575 | changed, lines = self.do_one_token_transform(lines) | |
562 | if not changed: |
|
576 | if not changed: | |
563 | return lines |
|
577 | return lines | |
564 |
|
578 | |||
565 | raise RuntimeError("Input transformation still changing after " |
|
579 | raise RuntimeError("Input transformation still changing after " | |
566 | "%d iterations. Aborting." % TRANSFORM_LOOP_LIMIT) |
|
580 | "%d iterations. Aborting." % TRANSFORM_LOOP_LIMIT) | |
567 |
|
581 | |||
568 | def transform_cell(self, cell: str) -> str: |
|
582 | def transform_cell(self, cell: str) -> str: | |
569 | """Transforms a cell of input code""" |
|
583 | """Transforms a cell of input code""" | |
570 | if not cell.endswith('\n'): |
|
584 | if not cell.endswith('\n'): | |
571 | cell += '\n' # Ensure the cell has a trailing newline |
|
585 | cell += '\n' # Ensure the cell has a trailing newline | |
572 | lines = cell.splitlines(keepends=True) |
|
586 | lines = cell.splitlines(keepends=True) | |
573 | for transform in self.cleanup_transforms + self.line_transforms: |
|
587 | for transform in self.cleanup_transforms + self.line_transforms: | |
574 | lines = transform(lines) |
|
588 | lines = transform(lines) | |
575 |
|
589 | |||
576 | lines = self.do_token_transforms(lines) |
|
590 | lines = self.do_token_transforms(lines) | |
577 | return ''.join(lines) |
|
591 | return ''.join(lines) | |
578 |
|
592 | |||
579 | def check_complete(self, cell: str): |
|
593 | def check_complete(self, cell: str): | |
580 | """Return whether a block of code is ready to execute, or should be continued |
|
594 | """Return whether a block of code is ready to execute, or should be continued | |
581 |
|
595 | |||
582 | Parameters |
|
596 | Parameters | |
583 | ---------- |
|
597 | ---------- | |
584 | source : string |
|
598 | source : string | |
585 | Python input code, which can be multiline. |
|
599 | Python input code, which can be multiline. | |
586 |
|
600 | |||
587 | Returns |
|
601 | Returns | |
588 | ------- |
|
602 | ------- | |
589 | status : str |
|
603 | status : str | |
590 | One of 'complete', 'incomplete', or 'invalid' if source is not a |
|
604 | One of 'complete', 'incomplete', or 'invalid' if source is not a | |
591 | prefix of valid code. |
|
605 | prefix of valid code. | |
592 | indent_spaces : int or None |
|
606 | indent_spaces : int or None | |
593 | The number of spaces by which to indent the next line of code. If |
|
607 | The number of spaces by which to indent the next line of code. If | |
594 | status is not 'incomplete', this is None. |
|
608 | status is not 'incomplete', this is None. | |
595 | """ |
|
609 | """ | |
596 | # Remember if the lines ends in a new line. |
|
610 | # Remember if the lines ends in a new line. | |
597 | ends_with_newline = False |
|
611 | ends_with_newline = False | |
598 | for character in reversed(cell): |
|
612 | for character in reversed(cell): | |
599 | if character == '\n': |
|
613 | if character == '\n': | |
600 | ends_with_newline = True |
|
614 | ends_with_newline = True | |
601 | break |
|
615 | break | |
602 | elif character.strip(): |
|
616 | elif character.strip(): | |
603 | break |
|
617 | break | |
604 | else: |
|
618 | else: | |
605 | continue |
|
619 | continue | |
606 |
|
620 | |||
607 | if not ends_with_newline: |
|
621 | if not ends_with_newline: | |
608 | # Append an newline for consistent tokenization |
|
622 | # Append an newline for consistent tokenization | |
609 | # See https://bugs.python.org/issue33899 |
|
623 | # See https://bugs.python.org/issue33899 | |
610 | cell += '\n' |
|
624 | cell += '\n' | |
611 |
|
625 | |||
612 | lines = cell.splitlines(keepends=True) |
|
626 | lines = cell.splitlines(keepends=True) | |
613 |
|
627 | |||
614 | if not lines: |
|
628 | if not lines: | |
615 | return 'complete', None |
|
629 | return 'complete', None | |
616 |
|
630 | |||
617 | if lines[-1].endswith('\\'): |
|
631 | if lines[-1].endswith('\\'): | |
618 | # Explicit backslash continuation |
|
632 | # Explicit backslash continuation | |
619 | return 'incomplete', find_last_indent(lines) |
|
633 | return 'incomplete', find_last_indent(lines) | |
620 |
|
634 | |||
621 | try: |
|
635 | try: | |
622 | for transform in self.cleanup_transforms: |
|
636 | for transform in self.cleanup_transforms: | |
623 | lines = transform(lines) |
|
637 | lines = transform(lines) | |
624 | except SyntaxError: |
|
638 | except SyntaxError: | |
625 | return 'invalid', None |
|
639 | return 'invalid', None | |
626 |
|
640 | |||
627 | if lines[0].startswith('%%'): |
|
641 | if lines[0].startswith('%%'): | |
628 | # Special case for cell magics - completion marked by blank line |
|
642 | # Special case for cell magics - completion marked by blank line | |
629 | if lines[-1].strip(): |
|
643 | if lines[-1].strip(): | |
630 | return 'incomplete', find_last_indent(lines) |
|
644 | return 'incomplete', find_last_indent(lines) | |
631 | else: |
|
645 | else: | |
632 | return 'complete', None |
|
646 | return 'complete', None | |
633 |
|
647 | |||
634 | try: |
|
648 | try: | |
635 | for transform in self.line_transforms: |
|
649 | for transform in self.line_transforms: | |
636 | lines = transform(lines) |
|
650 | lines = transform(lines) | |
637 | lines = self.do_token_transforms(lines) |
|
651 | lines = self.do_token_transforms(lines) | |
638 | except SyntaxError: |
|
652 | except SyntaxError: | |
639 | return 'invalid', None |
|
653 | return 'invalid', None | |
640 |
|
654 | |||
641 | tokens_by_line = make_tokens_by_line(lines) |
|
655 | tokens_by_line = make_tokens_by_line(lines) | |
642 |
|
656 | |||
643 | if not tokens_by_line: |
|
657 | if not tokens_by_line: | |
644 | return 'incomplete', find_last_indent(lines) |
|
658 | return 'incomplete', find_last_indent(lines) | |
645 |
|
659 | |||
646 | if tokens_by_line[-1][-1].type != tokenize.ENDMARKER: |
|
660 | if tokens_by_line[-1][-1].type != tokenize.ENDMARKER: | |
647 | # We're in a multiline string or expression |
|
661 | # We're in a multiline string or expression | |
648 | return 'incomplete', find_last_indent(lines) |
|
662 | return 'incomplete', find_last_indent(lines) | |
649 |
|
663 | |||
650 | newline_types = {tokenize.NEWLINE, tokenize.COMMENT, tokenize.ENDMARKER} |
|
664 | newline_types = {tokenize.NEWLINE, tokenize.COMMENT, tokenize.ENDMARKER} | |
651 |
|
665 | |||
652 | # Pop the last line which only contains DEDENTs and ENDMARKER |
|
666 | # Pop the last line which only contains DEDENTs and ENDMARKER | |
653 | last_token_line = None |
|
667 | last_token_line = None | |
654 | if {t.type for t in tokens_by_line[-1]} in [ |
|
668 | if {t.type for t in tokens_by_line[-1]} in [ | |
655 | {tokenize.DEDENT, tokenize.ENDMARKER}, |
|
669 | {tokenize.DEDENT, tokenize.ENDMARKER}, | |
656 | {tokenize.ENDMARKER} |
|
670 | {tokenize.ENDMARKER} | |
657 | ] and len(tokens_by_line) > 1: |
|
671 | ] and len(tokens_by_line) > 1: | |
658 | last_token_line = tokens_by_line.pop() |
|
672 | last_token_line = tokens_by_line.pop() | |
659 |
|
673 | |||
660 | while tokens_by_line[-1] and tokens_by_line[-1][-1].type in newline_types: |
|
674 | while tokens_by_line[-1] and tokens_by_line[-1][-1].type in newline_types: | |
661 | tokens_by_line[-1].pop() |
|
675 | tokens_by_line[-1].pop() | |
662 |
|
676 | |||
663 | if len(tokens_by_line) == 1 and not tokens_by_line[-1]: |
|
677 | if len(tokens_by_line) == 1 and not tokens_by_line[-1]: | |
664 | return 'incomplete', 0 |
|
678 | return 'incomplete', 0 | |
665 |
|
679 | |||
666 | if tokens_by_line[-1][-1].string == ':': |
|
680 | if tokens_by_line[-1][-1].string == ':': | |
667 | # The last line starts a block (e.g. 'if foo:') |
|
681 | # The last line starts a block (e.g. 'if foo:') | |
668 | ix = 0 |
|
682 | ix = 0 | |
669 | while tokens_by_line[-1][ix].type in {tokenize.INDENT, tokenize.DEDENT}: |
|
683 | while tokens_by_line[-1][ix].type in {tokenize.INDENT, tokenize.DEDENT}: | |
670 | ix += 1 |
|
684 | ix += 1 | |
671 |
|
685 | |||
672 | indent = tokens_by_line[-1][ix].start[1] |
|
686 | indent = tokens_by_line[-1][ix].start[1] | |
673 | return 'incomplete', indent + 4 |
|
687 | return 'incomplete', indent + 4 | |
674 |
|
688 | |||
675 | if tokens_by_line[-1][0].line.endswith('\\'): |
|
689 | if tokens_by_line[-1][0].line.endswith('\\'): | |
676 | return 'incomplete', None |
|
690 | return 'incomplete', None | |
677 |
|
691 | |||
678 | # At this point, our checks think the code is complete (or invalid). |
|
692 | # At this point, our checks think the code is complete (or invalid). | |
679 | # We'll use codeop.compile_command to check this with the real parser |
|
693 | # We'll use codeop.compile_command to check this with the real parser | |
680 | try: |
|
694 | try: | |
681 | with warnings.catch_warnings(): |
|
695 | with warnings.catch_warnings(): | |
682 | warnings.simplefilter('error', SyntaxWarning) |
|
696 | warnings.simplefilter('error', SyntaxWarning) | |
683 | res = compile_command(''.join(lines), symbol='exec') |
|
697 | res = compile_command(''.join(lines), symbol='exec') | |
684 | except (SyntaxError, OverflowError, ValueError, TypeError, |
|
698 | except (SyntaxError, OverflowError, ValueError, TypeError, | |
685 | MemoryError, SyntaxWarning): |
|
699 | MemoryError, SyntaxWarning): | |
686 | return 'invalid', None |
|
700 | return 'invalid', None | |
687 | else: |
|
701 | else: | |
688 | if res is None: |
|
702 | if res is None: | |
689 | return 'incomplete', find_last_indent(lines) |
|
703 | return 'incomplete', find_last_indent(lines) | |
690 |
|
704 | |||
691 | if last_token_line and last_token_line[0].type == tokenize.DEDENT: |
|
705 | if last_token_line and last_token_line[0].type == tokenize.DEDENT: | |
692 | if ends_with_newline: |
|
706 | if ends_with_newline: | |
693 | return 'complete', None |
|
707 | return 'complete', None | |
694 | return 'incomplete', find_last_indent(lines) |
|
708 | return 'incomplete', find_last_indent(lines) | |
695 |
|
709 | |||
696 | # If there's a blank line at the end, assume we're ready to execute |
|
710 | # If there's a blank line at the end, assume we're ready to execute | |
697 | if not lines[-1].strip(): |
|
711 | if not lines[-1].strip(): | |
698 | return 'complete', None |
|
712 | return 'complete', None | |
699 |
|
713 | |||
700 | return 'complete', None |
|
714 | return 'complete', None | |
701 |
|
715 | |||
702 |
|
716 | |||
703 | def find_last_indent(lines): |
|
717 | def find_last_indent(lines): | |
704 | m = _indent_re.match(lines[-1]) |
|
718 | m = _indent_re.match(lines[-1]) | |
705 | if not m: |
|
719 | if not m: | |
706 | return 0 |
|
720 | return 0 | |
707 | return len(m.group(0).replace('\t', ' '*4)) |
|
721 | return len(m.group(0).replace('\t', ' '*4)) |
@@ -1,88 +1,116 b'' | |||||
1 | """Tests for the line-based transformers in IPython.core.inputtransformer2 |
|
1 | """Tests for the line-based transformers in IPython.core.inputtransformer2 | |
2 |
|
2 | |||
3 | Line-based transformers are the simpler ones; token-based transformers are |
|
3 | Line-based transformers are the simpler ones; token-based transformers are | |
4 | more complex. See test_inputtransformer2 for tests for token-based transformers. |
|
4 | more complex. See test_inputtransformer2 for tests for token-based transformers. | |
5 | """ |
|
5 | """ | |
6 | import nose.tools as nt |
|
6 | import nose.tools as nt | |
7 |
|
7 | |||
8 | from IPython.core import inputtransformer2 as ipt2 |
|
8 | from IPython.core import inputtransformer2 as ipt2 | |
9 |
|
9 | |||
10 | CELL_MAGIC = ("""\ |
|
10 | CELL_MAGIC = ("""\ | |
11 | %%foo arg |
|
11 | %%foo arg | |
12 | body 1 |
|
12 | body 1 | |
13 | body 2 |
|
13 | body 2 | |
14 | """, """\ |
|
14 | """, """\ | |
15 | get_ipython().run_cell_magic('foo', 'arg', 'body 1\\nbody 2\\n') |
|
15 | get_ipython().run_cell_magic('foo', 'arg', 'body 1\\nbody 2\\n') | |
16 | """) |
|
16 | """) | |
17 |
|
17 | |||
18 | def test_cell_magic(): |
|
18 | def test_cell_magic(): | |
19 | for sample, expected in [CELL_MAGIC]: |
|
19 | for sample, expected in [CELL_MAGIC]: | |
20 | nt.assert_equal(ipt2.cell_magic(sample.splitlines(keepends=True)), |
|
20 | nt.assert_equal(ipt2.cell_magic(sample.splitlines(keepends=True)), | |
21 | expected.splitlines(keepends=True)) |
|
21 | expected.splitlines(keepends=True)) | |
22 |
|
22 | |||
23 | CLASSIC_PROMPT = ("""\ |
|
23 | CLASSIC_PROMPT = ("""\ | |
24 | >>> for a in range(5): |
|
24 | >>> for a in range(5): | |
25 | ... print(a) |
|
25 | ... print(a) | |
26 | """, """\ |
|
26 | """, """\ | |
27 | for a in range(5): |
|
27 | for a in range(5): | |
28 | print(a) |
|
28 | print(a) | |
29 | """) |
|
29 | """) | |
30 |
|
30 | |||
31 | CLASSIC_PROMPT_L2 = ("""\ |
|
31 | CLASSIC_PROMPT_L2 = ("""\ | |
32 | for a in range(5): |
|
32 | for a in range(5): | |
33 | ... print(a) |
|
33 | ... print(a) | |
34 | ... print(a ** 2) |
|
34 | ... print(a ** 2) | |
35 | """, """\ |
|
35 | """, """\ | |
36 | for a in range(5): |
|
36 | for a in range(5): | |
37 | print(a) |
|
37 | print(a) | |
38 | print(a ** 2) |
|
38 | print(a ** 2) | |
39 | """) |
|
39 | """) | |
40 |
|
40 | |||
41 | def test_classic_prompt(): |
|
41 | def test_classic_prompt(): | |
42 | for sample, expected in [CLASSIC_PROMPT, CLASSIC_PROMPT_L2]: |
|
42 | for sample, expected in [CLASSIC_PROMPT, CLASSIC_PROMPT_L2]: | |
43 | nt.assert_equal(ipt2.classic_prompt(sample.splitlines(keepends=True)), |
|
43 | nt.assert_equal(ipt2.classic_prompt(sample.splitlines(keepends=True)), | |
44 | expected.splitlines(keepends=True)) |
|
44 | expected.splitlines(keepends=True)) | |
45 |
|
45 | |||
46 | IPYTHON_PROMPT = ("""\ |
|
46 | IPYTHON_PROMPT = ("""\ | |
47 | In [1]: for a in range(5): |
|
47 | In [1]: for a in range(5): | |
48 | ...: print(a) |
|
48 | ...: print(a) | |
49 | """, """\ |
|
49 | """, """\ | |
50 | for a in range(5): |
|
50 | for a in range(5): | |
51 | print(a) |
|
51 | print(a) | |
52 | """) |
|
52 | """) | |
53 |
|
53 | |||
54 | IPYTHON_PROMPT_L2 = ("""\ |
|
54 | IPYTHON_PROMPT_L2 = ("""\ | |
55 | for a in range(5): |
|
55 | for a in range(5): | |
56 | ...: print(a) |
|
56 | ...: print(a) | |
57 | ...: print(a ** 2) |
|
57 | ...: print(a ** 2) | |
58 | """, """\ |
|
58 | """, """\ | |
59 | for a in range(5): |
|
59 | for a in range(5): | |
60 | print(a) |
|
60 | print(a) | |
61 | print(a ** 2) |
|
61 | print(a ** 2) | |
62 | """) |
|
62 | """) | |
63 |
|
63 | |||
64 | def test_ipython_prompt(): |
|
64 | def test_ipython_prompt(): | |
65 | for sample, expected in [IPYTHON_PROMPT, IPYTHON_PROMPT_L2]: |
|
65 | for sample, expected in [IPYTHON_PROMPT, IPYTHON_PROMPT_L2]: | |
66 | nt.assert_equal(ipt2.ipython_prompt(sample.splitlines(keepends=True)), |
|
66 | nt.assert_equal(ipt2.ipython_prompt(sample.splitlines(keepends=True)), | |
67 | expected.splitlines(keepends=True)) |
|
67 | expected.splitlines(keepends=True)) | |
68 |
|
68 | |||
69 | INDENT_SPACES = ("""\ |
|
69 | INDENT_SPACES = ("""\ | |
70 | if True: |
|
70 | if True: | |
71 | a = 3 |
|
71 | a = 3 | |
72 | """, """\ |
|
72 | """, """\ | |
73 | if True: |
|
73 | if True: | |
74 | a = 3 |
|
74 | a = 3 | |
75 | """) |
|
75 | """) | |
76 |
|
76 | |||
77 | INDENT_TABS = ("""\ |
|
77 | INDENT_TABS = ("""\ | |
78 | \tif True: |
|
78 | \tif True: | |
79 | \t\tb = 4 |
|
79 | \t\tb = 4 | |
80 | """, """\ |
|
80 | """, """\ | |
81 | if True: |
|
81 | if True: | |
82 | \tb = 4 |
|
82 | \tb = 4 | |
83 | """) |
|
83 | """) | |
84 |
|
84 | |||
85 | def test_leading_indent(): |
|
85 | def test_leading_indent(): | |
86 | for sample, expected in [INDENT_SPACES, INDENT_TABS]: |
|
86 | for sample, expected in [INDENT_SPACES, INDENT_TABS]: | |
87 | nt.assert_equal(ipt2.leading_indent(sample.splitlines(keepends=True)), |
|
87 | nt.assert_equal(ipt2.leading_indent(sample.splitlines(keepends=True)), | |
88 | expected.splitlines(keepends=True)) |
|
88 | expected.splitlines(keepends=True)) | |
|
89 | ||||
|
90 | LEADING_EMPTY_LINES = ("""\ | |||
|
91 | \t | |||
|
92 | ||||
|
93 | if True: | |||
|
94 | a = 3 | |||
|
95 | ||||
|
96 | b = 4 | |||
|
97 | """, """\ | |||
|
98 | if True: | |||
|
99 | a = 3 | |||
|
100 | ||||
|
101 | b = 4 | |||
|
102 | """) | |||
|
103 | ||||
|
104 | ONLY_EMPTY_LINES = ("""\ | |||
|
105 | \t | |||
|
106 | ||||
|
107 | """, """\ | |||
|
108 | \t | |||
|
109 | ||||
|
110 | """) | |||
|
111 | ||||
|
112 | def leading_empty_lines(): | |||
|
113 | for sample, expected in [LEADING_EMPTY_LINES, ONLY_EMPTY_LINES]: | |||
|
114 | nt.assert_equal( | |||
|
115 | ipt2.leading_empty_lines(sample.splitlines(keepends=True)), | |||
|
116 | expected.splitlines(keepends=True)) |
General Comments 0
You need to be logged in to leave comments.
Login now