Show More
@@ -1,203 +1,255 b'' | |||||
1 | import re |
|
1 | import re | |
2 | from typing import List, Tuple |
|
2 | from typing import List, Tuple | |
3 | from IPython.utils import tokenize2 |
|
3 | from IPython.utils import tokenize2 | |
4 | from IPython.utils.tokenutil import generate_tokens |
|
4 | from IPython.utils.tokenutil import generate_tokens | |
5 |
|
5 | |||
6 | def leading_indent(lines): |
|
6 | def leading_indent(lines): | |
7 | """Remove leading indentation. |
|
7 | """Remove leading indentation. | |
8 |
|
8 | |||
9 | If the first line starts with a spaces or tabs, the same whitespace will be |
|
9 | If the first line starts with a spaces or tabs, the same whitespace will be | |
10 | removed from each following line. |
|
10 | removed from each following line. | |
11 | """ |
|
11 | """ | |
12 | m = re.match(r'^[ \t]+', lines[0]) |
|
12 | m = re.match(r'^[ \t]+', lines[0]) | |
13 | if not m: |
|
13 | if not m: | |
14 | return lines |
|
14 | return lines | |
15 | space = m.group(0) |
|
15 | space = m.group(0) | |
16 | n = len(space) |
|
16 | n = len(space) | |
17 | return [l[n:] if l.startswith(space) else l |
|
17 | return [l[n:] if l.startswith(space) else l | |
18 | for l in lines] |
|
18 | for l in lines] | |
19 |
|
19 | |||
20 | class PromptStripper: |
|
20 | class PromptStripper: | |
21 | """Remove matching input prompts from a block of input. |
|
21 | """Remove matching input prompts from a block of input. | |
22 |
|
22 | |||
23 | Parameters |
|
23 | Parameters | |
24 | ---------- |
|
24 | ---------- | |
25 | prompt_re : regular expression |
|
25 | prompt_re : regular expression | |
26 | A regular expression matching any input prompt (including continuation) |
|
26 | A regular expression matching any input prompt (including continuation) | |
27 | initial_re : regular expression, optional |
|
27 | initial_re : regular expression, optional | |
28 | A regular expression matching only the initial prompt, but not continuation. |
|
28 | A regular expression matching only the initial prompt, but not continuation. | |
29 | If no initial expression is given, prompt_re will be used everywhere. |
|
29 | If no initial expression is given, prompt_re will be used everywhere. | |
30 | Used mainly for plain Python prompts, where the continuation prompt |
|
30 | Used mainly for plain Python prompts, where the continuation prompt | |
31 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. |
|
31 | ``...`` is a valid Python expression in Python 3, so shouldn't be stripped. | |
32 |
|
32 | |||
33 | If initial_re and prompt_re differ, |
|
33 | If initial_re and prompt_re differ, | |
34 | only initial_re will be tested against the first line. |
|
34 | only initial_re will be tested against the first line. | |
35 | If any prompt is found on the first two lines, |
|
35 | If any prompt is found on the first two lines, | |
36 | prompts will be stripped from the rest of the block. |
|
36 | prompts will be stripped from the rest of the block. | |
37 | """ |
|
37 | """ | |
38 | def __init__(self, prompt_re, initial_re=None): |
|
38 | def __init__(self, prompt_re, initial_re=None): | |
39 | self.prompt_re = prompt_re |
|
39 | self.prompt_re = prompt_re | |
40 | self.initial_re = initial_re or prompt_re |
|
40 | self.initial_re = initial_re or prompt_re | |
41 |
|
41 | |||
42 | def _strip(self, lines): |
|
42 | def _strip(self, lines): | |
43 | return [self.prompt_re.sub('', l, count=1) for l in lines] |
|
43 | return [self.prompt_re.sub('', l, count=1) for l in lines] | |
44 |
|
44 | |||
45 | def __call__(self, lines): |
|
45 | def __call__(self, lines): | |
46 | if self.initial_re.match(lines[0]) or \ |
|
46 | if self.initial_re.match(lines[0]) or \ | |
47 | (len(lines) > 1 and self.prompt_re.match(lines[1])): |
|
47 | (len(lines) > 1 and self.prompt_re.match(lines[1])): | |
48 | return self._strip(lines) |
|
48 | return self._strip(lines) | |
49 | return lines |
|
49 | return lines | |
50 |
|
50 | |||
51 | classic_prompt = PromptStripper( |
|
51 | classic_prompt = PromptStripper( | |
52 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), |
|
52 | prompt_re=re.compile(r'^(>>>|\.\.\.)( |$)'), | |
53 | initial_re=re.compile(r'^>>>( |$)') |
|
53 | initial_re=re.compile(r'^>>>( |$)') | |
54 | ) |
|
54 | ) | |
55 |
|
55 | |||
56 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) |
|
56 | ipython_prompt = PromptStripper(re.compile(r'^(In \[\d+\]: |\s*\.{3,}: ?)')) | |
57 |
|
57 | |||
58 | def cell_magic(lines): |
|
58 | def cell_magic(lines): | |
59 | if not lines[0].startswith('%%'): |
|
59 | if not lines[0].startswith('%%'): | |
60 | return lines |
|
60 | return lines | |
61 | if re.match('%%\w+\?', lines[0]): |
|
61 | if re.match('%%\w+\?', lines[0]): | |
62 | # This case will be handled by help_end |
|
62 | # This case will be handled by help_end | |
63 | return lines |
|
63 | return lines | |
64 | magic_name, first_line = lines[0][2:].partition(' ') |
|
64 | magic_name, first_line = lines[0][2:].partition(' ') | |
65 | body = '\n'.join(lines[1:]) |
|
65 | body = '\n'.join(lines[1:]) | |
66 | return ['get_ipython().run_cell_magic(%r, %r, %r)' % (magic_name, first_line, body)] |
|
66 | return ['get_ipython().run_cell_magic(%r, %r, %r)' % (magic_name, first_line, body)] | |
67 |
|
67 | |||
68 | line_transforms = [ |
|
68 | line_transforms = [ | |
69 | leading_indent, |
|
69 | leading_indent, | |
70 | classic_prompt, |
|
70 | classic_prompt, | |
71 | ipython_prompt, |
|
71 | ipython_prompt, | |
72 | cell_magic, |
|
72 | cell_magic, | |
73 | ] |
|
73 | ] | |
74 |
|
74 | |||
75 | # ----- |
|
75 | # ----- | |
76 |
|
76 | |||
77 | def help_end(tokens_by_line): |
|
77 | def help_end(tokens_by_line): | |
78 | pass |
|
78 | pass | |
79 |
|
79 | |||
80 | def escaped_command(tokens_by_line): |
|
80 | def escaped_command(tokens_by_line): | |
81 | pass |
|
81 | pass | |
82 |
|
82 | |||
83 | def _find_assign_op(token_line): |
|
83 | def _find_assign_op(token_line): | |
84 | # Find the first assignment in the line ('=' not inside brackets) |
|
84 | # Find the first assignment in the line ('=' not inside brackets) | |
85 | # We don't try to support multiple special assignment (a = b = %foo) |
|
85 | # We don't try to support multiple special assignment (a = b = %foo) | |
86 | paren_level = 0 |
|
86 | paren_level = 0 | |
87 | for i, ti in enumerate(token_line): |
|
87 | for i, ti in enumerate(token_line): | |
88 | s = ti.string |
|
88 | s = ti.string | |
89 | if s == '=' and paren_level == 0: |
|
89 | if s == '=' and paren_level == 0: | |
90 | return i |
|
90 | return i | |
91 | if s in '([{': |
|
91 | if s in '([{': | |
92 | paren_level += 1 |
|
92 | paren_level += 1 | |
93 | elif s in ')]}': |
|
93 | elif s in ')]}': | |
94 | paren_level -= 1 |
|
94 | paren_level -= 1 | |
95 |
|
95 | |||
96 | class MagicAssign: |
|
96 | class MagicAssign: | |
97 | @staticmethod |
|
97 | @staticmethod | |
98 | def find(tokens_by_line): |
|
98 | def find(tokens_by_line): | |
99 | """Find the first magic assignment (a = %foo) in the cell. |
|
99 | """Find the first magic assignment (a = %foo) in the cell. | |
100 |
|
100 | |||
101 | Returns (line, column) of the % if found, or None. |
|
101 | Returns (line, column) of the % if found, or None. | |
102 | """ |
|
102 | """ | |
103 | for line in tokens_by_line: |
|
103 | for line in tokens_by_line: | |
104 | assign_ix = _find_assign_op(line) |
|
104 | assign_ix = _find_assign_op(line) | |
105 | if (assign_ix is not None) \ |
|
105 | if (assign_ix is not None) \ | |
106 | and (len(line) >= assign_ix + 2) \ |
|
106 | and (len(line) >= assign_ix + 2) \ | |
107 | and (line[assign_ix+1].string == '%') \ |
|
107 | and (line[assign_ix+1].string == '%') \ | |
108 | and (line[assign_ix+2].type == tokenize2.NAME): |
|
108 | and (line[assign_ix+2].type == tokenize2.NAME): | |
109 | return line[assign_ix+1].start |
|
109 | return line[assign_ix+1].start | |
110 |
|
110 | |||
111 | @staticmethod |
|
111 | @staticmethod | |
112 | def transform(lines: List[str], start: Tuple[int, int]): |
|
112 | def transform(lines: List[str], start: Tuple[int, int]): | |
113 | """Transform a magic assignment found by find |
|
113 | """Transform a magic assignment found by find | |
114 | """ |
|
114 | """ | |
115 | start_line = start[0] - 1 # Shift from 1-index to 0-index |
|
115 | start_line = start[0] - 1 # Shift from 1-index to 0-index | |
116 | start_col = start[1] |
|
116 | start_col = start[1] | |
117 |
|
117 | |||
118 | print("Start at", start_line, start_col) |
|
118 | print("Start at", start_line, start_col) | |
119 | print("Line", lines[start_line]) |
|
119 | print("Line", lines[start_line]) | |
120 |
|
120 | |||
121 | lhs, rhs = lines[start_line][:start_col], lines[start_line][start_col:-1] |
|
121 | lhs, rhs = lines[start_line][:start_col], lines[start_line][start_col:-1] | |
122 | assert rhs.startswith('%'), rhs |
|
122 | assert rhs.startswith('%'), rhs | |
123 | magic_name, _, args = rhs[1:].partition(' ') |
|
123 | magic_name, _, args = rhs[1:].partition(' ') | |
124 | args_parts = [args] |
|
124 | args_parts = [args] | |
125 | end_line = start_line |
|
125 | end_line = start_line | |
126 | # Follow explicit (backslash) line continuations |
|
126 | # Follow explicit (backslash) line continuations | |
127 | while end_line < len(lines) and args_parts[-1].endswith('\\'): |
|
127 | while end_line < len(lines) and args_parts[-1].endswith('\\'): | |
128 | end_line += 1 |
|
128 | end_line += 1 | |
129 | args_parts[-1] = args_parts[-1][:-1] # Trim backslash |
|
129 | args_parts[-1] = args_parts[-1][:-1] # Trim backslash | |
130 | args_parts.append(lines[end_line][:-1]) # Trim newline |
|
130 | args_parts.append(lines[end_line][:-1]) # Trim newline | |
131 | args = ' '.join(args_parts) |
|
131 | args = ' '.join(args_parts) | |
132 |
|
132 | |||
133 | lines_before = lines[:start_line] |
|
133 | lines_before = lines[:start_line] | |
134 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) |
|
134 | call = "get_ipython().run_line_magic({!r}, {!r})".format(magic_name, args) | |
135 | new_line = lhs + call + '\n' |
|
135 | new_line = lhs + call + '\n' | |
136 | lines_after = lines[end_line+1:] |
|
136 | lines_after = lines[end_line+1:] | |
137 |
|
137 | |||
138 | return lines_before + [new_line] + lines_after |
|
138 | return lines_before + [new_line] + lines_after | |
139 |
|
139 | |||
|
140 | ||||
|
141 | class SystemAssign: | |||
|
142 | @staticmethod | |||
|
143 | def find(tokens_by_line): | |||
|
144 | """Find the first system assignment (a = !foo) in the cell. | |||
|
145 | ||||
|
146 | Returns (line, column) of the ! if found, or None. | |||
|
147 | """ | |||
|
148 | for line in tokens_by_line: | |||
|
149 | assign_ix = _find_assign_op(line) | |||
|
150 | if (assign_ix is not None) \ | |||
|
151 | and (len(line) >= assign_ix + 2) \ | |||
|
152 | and (line[assign_ix + 1].type == tokenize2.ERRORTOKEN): | |||
|
153 | ix = assign_ix + 1 | |||
|
154 | ||||
|
155 | while ix < len(line) and line[ix].type == tokenize2.ERRORTOKEN: | |||
|
156 | if line[ix].string == '!': | |||
|
157 | return line[ix].start | |||
|
158 | elif not line[ix].string.isspace(): | |||
|
159 | break | |||
|
160 | ix += 1 | |||
|
161 | ||||
|
162 | @staticmethod | |||
|
163 | def transform(lines: List[str], start: Tuple[int, int]): | |||
|
164 | """Transform a system assignment found by find | |||
|
165 | """ | |||
|
166 | start_line = start[0] - 1 # Shift from 1-index to 0-index | |||
|
167 | start_col = start[1] | |||
|
168 | ||||
|
169 | print("Start at", start_line, start_col) | |||
|
170 | print("Line", lines[start_line]) | |||
|
171 | ||||
|
172 | lhs, rhs = lines[start_line][:start_col], lines[start_line][ | |||
|
173 | start_col:-1] | |||
|
174 | assert rhs.startswith('!'), rhs | |||
|
175 | cmd_parts = [rhs[1:]] | |||
|
176 | end_line = start_line | |||
|
177 | # Follow explicit (backslash) line continuations | |||
|
178 | while end_line < len(lines) and cmd_parts[-1].endswith('\\'): | |||
|
179 | end_line += 1 | |||
|
180 | cmd_parts[-1] = cmd_parts[-1][:-1] # Trim backslash | |||
|
181 | cmd_parts.append(lines[end_line][:-1]) # Trim newline | |||
|
182 | cmd = ' '.join(cmd_parts) | |||
|
183 | ||||
|
184 | lines_before = lines[:start_line] | |||
|
185 | call = "get_ipython().getoutput({!r})".format(cmd) | |||
|
186 | new_line = lhs + call + '\n' | |||
|
187 | lines_after = lines[end_line + 1:] | |||
|
188 | ||||
|
189 | return lines_before + [new_line] + lines_after | |||
|
190 | ||||
140 | def make_tokens_by_line(lines): |
|
191 | def make_tokens_by_line(lines): | |
141 | tokens_by_line = [[]] |
|
192 | tokens_by_line = [[]] | |
142 | for token in generate_tokens(iter(lines).__next__): |
|
193 | for token in generate_tokens(iter(lines).__next__): | |
143 | tokens_by_line[-1].append(token) |
|
194 | tokens_by_line[-1].append(token) | |
144 | if token.type == tokenize2.NEWLINE: |
|
195 | if token.type == tokenize2.NEWLINE: | |
145 | tokens_by_line.append([]) |
|
196 | tokens_by_line.append([]) | |
146 |
|
197 | |||
147 | return tokens_by_line |
|
198 | return tokens_by_line | |
148 |
|
199 | |||
149 | class TokenTransformers: |
|
200 | class TokenTransformers: | |
150 | def __init__(self): |
|
201 | def __init__(self): | |
151 | self.transformers = [ |
|
202 | self.transformers = [ | |
152 | MagicAssign |
|
203 | MagicAssign, | |
|
204 | SystemAssign, | |||
153 | ] |
|
205 | ] | |
154 |
|
206 | |||
155 | def do_one_transform(self, lines): |
|
207 | def do_one_transform(self, lines): | |
156 | """Find and run the transform earliest in the code. |
|
208 | """Find and run the transform earliest in the code. | |
157 |
|
209 | |||
158 | Returns (changed, lines). |
|
210 | Returns (changed, lines). | |
159 |
|
211 | |||
160 | This method is called repeatedly until changed is False, indicating |
|
212 | This method is called repeatedly until changed is False, indicating | |
161 | that all available transformations are complete. |
|
213 | that all available transformations are complete. | |
162 |
|
214 | |||
163 | The tokens following IPython special syntax might not be valid, so |
|
215 | The tokens following IPython special syntax might not be valid, so | |
164 | the transformed code is retokenised every time to identify the next |
|
216 | the transformed code is retokenised every time to identify the next | |
165 | piece of special syntax. Hopefully long code cells are mostly valid |
|
217 | piece of special syntax. Hopefully long code cells are mostly valid | |
166 | Python, not using lots of IPython special syntax, so this shouldn't be |
|
218 | Python, not using lots of IPython special syntax, so this shouldn't be | |
167 | a performance issue. |
|
219 | a performance issue. | |
168 | """ |
|
220 | """ | |
169 | tokens_by_line = make_tokens_by_line(lines) |
|
221 | tokens_by_line = make_tokens_by_line(lines) | |
170 | candidates = [] |
|
222 | candidates = [] | |
171 | for transformer in self.transformers: |
|
223 | for transformer in self.transformers: | |
172 | locn = transformer.find(tokens_by_line) |
|
224 | locn = transformer.find(tokens_by_line) | |
173 | if locn: |
|
225 | if locn: | |
174 | candidates.append((locn, transformer)) |
|
226 | candidates.append((locn, transformer)) | |
175 |
|
227 | |||
176 | if not candidates: |
|
228 | if not candidates: | |
177 | # Nothing to transform |
|
229 | # Nothing to transform | |
178 | return False, lines |
|
230 | return False, lines | |
179 |
|
231 | |||
180 | first_locn, transformer = min(candidates) |
|
232 | first_locn, transformer = min(candidates) | |
181 | return True, transformer.transform(lines, first_locn) |
|
233 | return True, transformer.transform(lines, first_locn) | |
182 |
|
234 | |||
183 | def __call__(self, lines): |
|
235 | def __call__(self, lines): | |
184 | while True: |
|
236 | while True: | |
185 | changed, lines = self.do_one_transform(lines) |
|
237 | changed, lines = self.do_one_transform(lines) | |
186 | if not changed: |
|
238 | if not changed: | |
187 | return lines |
|
239 | return lines | |
188 |
|
240 | |||
189 | def assign_from_system(tokens_by_line, lines): |
|
241 | def assign_from_system(tokens_by_line, lines): | |
190 | pass |
|
242 | pass | |
191 |
|
243 | |||
192 |
|
244 | |||
193 | def transform_cell(cell): |
|
245 | def transform_cell(cell): | |
194 | if not cell.endswith('\n'): |
|
246 | if not cell.endswith('\n'): | |
195 | cell += '\n' # Ensure every line has a newline |
|
247 | cell += '\n' # Ensure every line has a newline | |
196 | lines = cell.splitlines(keepends=True) |
|
248 | lines = cell.splitlines(keepends=True) | |
197 | for transform in line_transforms: |
|
249 | for transform in line_transforms: | |
198 | #print(transform, lines) |
|
250 | #print(transform, lines) | |
199 | lines = transform(lines) |
|
251 | lines = transform(lines) | |
200 |
|
252 | |||
201 | lines = TokenTransformers()(lines) |
|
253 | lines = TokenTransformers()(lines) | |
202 | for line in lines: |
|
254 | for line in lines: | |
203 | print('~~', line) |
|
255 | print('~~', line) |
@@ -1,37 +1,54 b'' | |||||
1 | import nose.tools as nt |
|
1 | import nose.tools as nt | |
2 |
|
2 | |||
3 | from IPython.core import inputtransformer2 |
|
3 | from IPython.core import inputtransformer2 as ipt2 | |
4 | from IPython.core.inputtransformer2 import make_tokens_by_line |
|
4 | from IPython.core.inputtransformer2 import make_tokens_by_line | |
5 |
|
5 | |||
6 | MULTILINE_MAGIC_ASSIGN = ("""\ |
|
6 | MULTILINE_MAGIC_ASSIGN = ("""\ | |
7 | a = f() |
|
7 | a = f() | |
8 | b = %foo \\ |
|
8 | b = %foo \\ | |
9 | bar |
|
9 | bar | |
10 | g() |
|
10 | g() | |
11 | """.splitlines(keepends=True), """\ |
|
11 | """.splitlines(keepends=True), """\ | |
12 | a = f() |
|
12 | a = f() | |
13 | b = get_ipython().run_line_magic('foo', ' bar') |
|
13 | b = get_ipython().run_line_magic('foo', ' bar') | |
14 | g() |
|
14 | g() | |
15 | """.splitlines(keepends=True)) |
|
15 | """.splitlines(keepends=True)) | |
16 |
|
16 | |||
17 | MULTILINE_SYSTEM_ASSIGN = ("""\ |
|
17 | MULTILINE_SYSTEM_ASSIGN = ("""\ | |
18 | a = f() |
|
18 | a = f() | |
19 | b = !foo \\ |
|
19 | b = !foo \\ | |
20 | bar |
|
20 | bar | |
21 | g() |
|
21 | g() | |
22 | """.splitlines(keepends=True), """\ |
|
22 | """.splitlines(keepends=True), """\ | |
23 | a = f() |
|
23 | a = f() | |
24 | b = get_ipython().getoutput('foo bar') |
|
24 | b = get_ipython().getoutput('foo bar') | |
25 | g() |
|
25 | g() | |
26 | """.splitlines(keepends=True)) |
|
26 | """.splitlines(keepends=True)) | |
27 |
|
27 | |||
28 | def test_find_assign_magic(): |
|
28 | def test_find_assign_magic(): | |
29 | tbl = make_tokens_by_line(MULTILINE_MAGIC_ASSIGN[0]) |
|
29 | tbl = make_tokens_by_line(MULTILINE_MAGIC_ASSIGN[0]) | |
30 |
nt.assert_equal(i |
|
30 | nt.assert_equal(ipt2.MagicAssign.find(tbl), (2, 4)) | |
31 |
|
31 | |||
32 | tbl = make_tokens_by_line(MULTILINE_SYSTEM_ASSIGN[0]) # Nothing to find |
|
32 | tbl = make_tokens_by_line(MULTILINE_SYSTEM_ASSIGN[0]) # Nothing to find | |
33 |
nt.assert_equal(i |
|
33 | nt.assert_equal(ipt2.MagicAssign.find(tbl), None) | |
34 |
|
34 | |||
35 | def test_transform_assign_magic(): |
|
35 | def test_transform_assign_magic(): | |
36 |
res = i |
|
36 | res = ipt2.MagicAssign.transform(MULTILINE_MAGIC_ASSIGN[0], (2, 4)) | |
37 | nt.assert_equal(res, MULTILINE_MAGIC_ASSIGN[1]) |
|
37 | nt.assert_equal(res, MULTILINE_MAGIC_ASSIGN[1]) | |
|
38 | ||||
|
39 | def test_find_assign_system(): | |||
|
40 | tbl = make_tokens_by_line(MULTILINE_SYSTEM_ASSIGN[0]) | |||
|
41 | nt.assert_equal(ipt2.SystemAssign.find(tbl), (2, 4)) | |||
|
42 | ||||
|
43 | tbl = make_tokens_by_line(["a = !ls\n"]) | |||
|
44 | nt.assert_equal(ipt2.SystemAssign.find(tbl), (1, 5)) | |||
|
45 | ||||
|
46 | tbl = make_tokens_by_line(["a=!ls\n"]) | |||
|
47 | nt.assert_equal(ipt2.SystemAssign.find(tbl), (1, 2)) | |||
|
48 | ||||
|
49 | tbl = make_tokens_by_line(MULTILINE_MAGIC_ASSIGN[0]) # Nothing to find | |||
|
50 | nt.assert_equal(ipt2.SystemAssign.find(tbl), None) | |||
|
51 | ||||
|
52 | def test_transform_assign_system(): | |||
|
53 | res = ipt2.SystemAssign.transform(MULTILINE_SYSTEM_ASSIGN[0], (2, 4)) | |||
|
54 | nt.assert_equal(res, MULTILINE_SYSTEM_ASSIGN[1]) |
General Comments 0
You need to be logged in to leave comments.
Login now