| 1 | # SPDX-License-Identifier: MIT |
| 2 | |
| 3 | # Emitter expects events obeying the following grammar: |
| 4 | # stream ::= STREAM-START document* STREAM-END |
| 5 | # document ::= DOCUMENT-START node DOCUMENT-END |
| 6 | # node ::= SCALAR | sequence | mapping |
| 7 | # sequence ::= SEQUENCE-START node* SEQUENCE-END |
| 8 | # mapping ::= MAPPING-START (node node)* MAPPING-END |
| 9 | |
| 10 | __all__ = ['Emitter', 'EmitterError'] |
| 11 | |
| 12 | from .error import YAMLError |
| 13 | from .events import * |
| 14 | |
| 15 | class EmitterError(YAMLError): |
| 16 | pass |
| 17 | |
| 18 | class ScalarAnalysis: |
| 19 | def __init__(self, scalar, empty, multiline, |
| 20 | allow_flow_plain, allow_block_plain, |
| 21 | allow_single_quoted, allow_double_quoted, |
| 22 | allow_block): |
| 23 | self.scalar = scalar |
| 24 | self.empty = empty |
| 25 | self.multiline = multiline |
| 26 | self.allow_flow_plain = allow_flow_plain |
| 27 | self.allow_block_plain = allow_block_plain |
| 28 | self.allow_single_quoted = allow_single_quoted |
| 29 | self.allow_double_quoted = allow_double_quoted |
| 30 | self.allow_block = allow_block |
| 31 | |
| 32 | class Emitter: |
| 33 | |
| 34 | DEFAULT_TAG_PREFIXES = { |
| 35 | '!' : '!', |
| 36 | 'tag:yaml.org,2002:' : '!!', |
| 37 | } |
| 38 | |
| 39 | def __init__(self, stream, canonical=None, indent=None, width=None, |
| 40 | allow_unicode=None, line_break=None): |
| 41 | |
| 42 | # The stream should have the methods `write` and possibly `flush`. |
| 43 | self.stream = stream |
| 44 | |
| 45 | # Encoding can be overriden by STREAM-START. |
| 46 | self.encoding = None |
| 47 | |
| 48 | # Emitter is a state machine with a stack of states to handle nested |
| 49 | # structures. |
| 50 | self.states = [] |
| 51 | self.state = self.expect_stream_start |
| 52 | |
| 53 | # Current event and the event queue. |
| 54 | self.events = [] |
| 55 | self.event = None |
| 56 | |
| 57 | # The current indentation level and the stack of previous indents. |
| 58 | self.indents = [] |
| 59 | self.indent = None |
| 60 | |
| 61 | # Flow level. |
| 62 | self.flow_level = 0 |
| 63 | |
| 64 | # Contexts. |
| 65 | self.root_context = False |
| 66 | self.sequence_context = False |
| 67 | self.mapping_context = False |
| 68 | self.simple_key_context = False |
| 69 | |
| 70 | # Characteristics of the last emitted character: |
| 71 | # - current position. |
| 72 | # - is it a whitespace? |
| 73 | # - is it an indention character |
| 74 | # (indentation space, '-', '?', or ':')? |
| 75 | self.line = 0 |
| 76 | self.column = 0 |
| 77 | self.whitespace = True |
| 78 | self.indention = True |
| 79 | |
| 80 | # Whether the document requires an explicit document indicator |
| 81 | self.open_ended = False |
| 82 | |
| 83 | # Formatting details. |
| 84 | self.canonical = canonical |
| 85 | self.allow_unicode = allow_unicode |
| 86 | self.best_indent = 2 |
| 87 | if indent and 1 < indent < 10: |
| 88 | self.best_indent = indent |
| 89 | self.best_width = 80 |
| 90 | if width and width > self.best_indent*2: |
| 91 | self.best_width = width |
| 92 | self.best_line_break = '\n' |
| 93 | if line_break in ['\r', '\n', '\r\n']: |
| 94 | self.best_line_break = line_break |
| 95 | |
| 96 | # Tag prefixes. |
| 97 | self.tag_prefixes = None |
| 98 | |
| 99 | # Prepared anchor and tag. |
| 100 | self.prepared_anchor = None |
| 101 | self.prepared_tag = None |
| 102 | |
| 103 | # Scalar analysis and style. |
| 104 | self.analysis = None |
| 105 | self.style = None |
| 106 | |
| 107 | def dispose(self): |
| 108 | # Reset the state attributes (to clear self-references) |
| 109 | self.states = [] |
| 110 | self.state = None |
| 111 | |
| 112 | def emit(self, event): |
| 113 | self.events.append(event) |
| 114 | while not self.need_more_events(): |
| 115 | self.event = self.events.pop(0) |
| 116 | self.state() |
| 117 | self.event = None |
| 118 | |
| 119 | # In some cases, we wait for a few next events before emitting. |
| 120 | |
| 121 | def need_more_events(self): |
| 122 | if not self.events: |
| 123 | return True |
| 124 | event = self.events[0] |
| 125 | if isinstance(event, DocumentStartEvent): |
| 126 | return self.need_events(1) |
| 127 | elif isinstance(event, SequenceStartEvent): |
| 128 | return self.need_events(2) |
| 129 | elif isinstance(event, MappingStartEvent): |
| 130 | return self.need_events(3) |
| 131 | else: |
| 132 | return False |
| 133 | |
| 134 | def need_events(self, count): |
| 135 | level = 0 |
| 136 | for event in self.events[1:]: |
| 137 | if isinstance(event, (DocumentStartEvent, CollectionStartEvent)): |
| 138 | level += 1 |
| 139 | elif isinstance(event, (DocumentEndEvent, CollectionEndEvent)): |
| 140 | level -= 1 |
| 141 | elif isinstance(event, StreamEndEvent): |
| 142 | level = -1 |
| 143 | if level < 0: |
| 144 | return False |
| 145 | return (len(self.events) < count+1) |
| 146 | |
| 147 | def increase_indent(self, flow=False, indentless=False): |
| 148 | self.indents.append(self.indent) |
| 149 | if self.indent is None: |
| 150 | if flow: |
| 151 | self.indent = self.best_indent |
| 152 | else: |
| 153 | self.indent = 0 |
| 154 | elif not indentless: |
| 155 | self.indent += self.best_indent |
| 156 | |
| 157 | # States. |
| 158 | |
| 159 | # Stream handlers. |
| 160 | |
| 161 | def expect_stream_start(self): |
| 162 | if isinstance(self.event, StreamStartEvent): |
| 163 | if self.event.encoding and not hasattr(self.stream, 'encoding'): |
| 164 | self.encoding = self.event.encoding |
| 165 | self.write_stream_start() |
| 166 | self.state = self.expect_first_document_start |
| 167 | else: |
| 168 | raise EmitterError("expected StreamStartEvent, but got %s" |
| 169 | % self.event) |
| 170 | |
| 171 | def expect_nothing(self): |
| 172 | raise EmitterError("expected nothing, but got %s" % self.event) |
| 173 | |
| 174 | # Document handlers. |
| 175 | |
| 176 | def expect_first_document_start(self): |
| 177 | return self.expect_document_start(first=True) |
| 178 | |
| 179 | def expect_document_start(self, first=False): |
| 180 | if isinstance(self.event, DocumentStartEvent): |
| 181 | if (self.event.version or self.event.tags) and self.open_ended: |
| 182 | self.write_indicator('...', True) |
| 183 | self.write_indent() |
| 184 | if self.event.version: |
| 185 | version_text = self.prepare_version(self.event.version) |
| 186 | self.write_version_directive(version_text) |
| 187 | self.tag_prefixes = self.DEFAULT_TAG_PREFIXES.copy() |
| 188 | if self.event.tags: |
| 189 | handles = sorted(self.event.tags.keys()) |
| 190 | for handle in handles: |
| 191 | prefix = self.event.tags[handle] |
| 192 | self.tag_prefixes[prefix] = handle |
| 193 | handle_text = self.prepare_tag_handle(handle) |
| 194 | prefix_text = self.prepare_tag_prefix(prefix) |
| 195 | self.write_tag_directive(handle_text, prefix_text) |
| 196 | implicit = (first and not self.event.explicit and not self.canonical |
| 197 | and not self.event.version and not self.event.tags |
| 198 | and not self.check_empty_document()) |
| 199 | if not implicit: |
| 200 | self.write_indent() |
| 201 | self.write_indicator('---', True) |
| 202 | if self.canonical: |
| 203 | self.write_indent() |
| 204 | self.state = self.expect_document_root |
| 205 | elif isinstance(self.event, StreamEndEvent): |
| 206 | if self.open_ended: |
| 207 | self.write_indicator('...', True) |
| 208 | self.write_indent() |
| 209 | self.write_stream_end() |
| 210 | self.state = self.expect_nothing |
| 211 | else: |
| 212 | raise EmitterError("expected DocumentStartEvent, but got %s" |
| 213 | % self.event) |
| 214 | |
| 215 | def expect_document_end(self): |
| 216 | if isinstance(self.event, DocumentEndEvent): |
| 217 | self.write_indent() |
| 218 | if self.event.explicit: |
| 219 | self.write_indicator('...', True) |
| 220 | self.write_indent() |
| 221 | self.flush_stream() |
| 222 | self.state = self.expect_document_start |
| 223 | else: |
| 224 | raise EmitterError("expected DocumentEndEvent, but got %s" |
| 225 | % self.event) |
| 226 | |
| 227 | def expect_document_root(self): |
| 228 | self.states.append(self.expect_document_end) |
| 229 | self.expect_node(root=True) |
| 230 | |
| 231 | # Node handlers. |
| 232 | |
| 233 | def expect_node(self, root=False, sequence=False, mapping=False, |
| 234 | simple_key=False): |
| 235 | self.root_context = root |
| 236 | self.sequence_context = sequence |
| 237 | self.mapping_context = mapping |
| 238 | self.simple_key_context = simple_key |
| 239 | if isinstance(self.event, AliasEvent): |
| 240 | self.expect_alias() |
| 241 | elif isinstance(self.event, (ScalarEvent, CollectionStartEvent)): |
| 242 | self.process_anchor('&') |
| 243 | self.process_tag() |
| 244 | if isinstance(self.event, ScalarEvent): |
| 245 | self.expect_scalar() |
| 246 | elif isinstance(self.event, SequenceStartEvent): |
| 247 | if self.flow_level or self.canonical or self.event.flow_style \ |
| 248 | or self.check_empty_sequence(): |
| 249 | self.expect_flow_sequence() |
| 250 | else: |
| 251 | self.expect_block_sequence() |
| 252 | elif isinstance(self.event, MappingStartEvent): |
| 253 | if self.flow_level or self.canonical or self.event.flow_style \ |
| 254 | or self.check_empty_mapping(): |
| 255 | self.expect_flow_mapping() |
| 256 | else: |
| 257 | self.expect_block_mapping() |
| 258 | else: |
| 259 | raise EmitterError("expected NodeEvent, but got %s" % self.event) |
| 260 | |
| 261 | def expect_alias(self): |
| 262 | if self.event.anchor is None: |
| 263 | raise EmitterError("anchor is not specified for alias") |
| 264 | self.process_anchor('*') |
| 265 | self.state = self.states.pop() |
| 266 | |
| 267 | def expect_scalar(self): |
| 268 | self.increase_indent(flow=True) |
| 269 | self.process_scalar() |
| 270 | self.indent = self.indents.pop() |
| 271 | self.state = self.states.pop() |
| 272 | |
| 273 | # Flow sequence handlers. |
| 274 | |
| 275 | def expect_flow_sequence(self): |
| 276 | self.write_indicator('[', True, whitespace=True) |
| 277 | self.flow_level += 1 |
| 278 | self.increase_indent(flow=True) |
| 279 | self.state = self.expect_first_flow_sequence_item |
| 280 | |
| 281 | def expect_first_flow_sequence_item(self): |
| 282 | if isinstance(self.event, SequenceEndEvent): |
| 283 | self.indent = self.indents.pop() |
| 284 | self.flow_level -= 1 |
| 285 | self.write_indicator(']', False) |
| 286 | self.state = self.states.pop() |
| 287 | else: |
| 288 | if self.canonical or self.column > self.best_width: |
| 289 | self.write_indent() |
| 290 | self.states.append(self.expect_flow_sequence_item) |
| 291 | self.expect_node(sequence=True) |
| 292 | |
| 293 | def expect_flow_sequence_item(self): |
| 294 | if isinstance(self.event, SequenceEndEvent): |
| 295 | self.indent = self.indents.pop() |
| 296 | self.flow_level -= 1 |
| 297 | if self.canonical: |
| 298 | self.write_indicator(',', False) |
| 299 | self.write_indent() |
| 300 | self.write_indicator(']', False) |
| 301 | self.state = self.states.pop() |
| 302 | else: |
| 303 | self.write_indicator(',', False) |
| 304 | if self.canonical or self.column > self.best_width: |
| 305 | self.write_indent() |
| 306 | self.states.append(self.expect_flow_sequence_item) |
| 307 | self.expect_node(sequence=True) |
| 308 | |
| 309 | # Flow mapping handlers. |
| 310 | |
| 311 | def expect_flow_mapping(self): |
| 312 | self.write_indicator('{', True, whitespace=True) |
| 313 | self.flow_level += 1 |
| 314 | self.increase_indent(flow=True) |
| 315 | self.state = self.expect_first_flow_mapping_key |
| 316 | |
| 317 | def expect_first_flow_mapping_key(self): |
| 318 | if isinstance(self.event, MappingEndEvent): |
| 319 | self.indent = self.indents.pop() |
| 320 | self.flow_level -= 1 |
| 321 | self.write_indicator('}', False) |
| 322 | self.state = self.states.pop() |
| 323 | else: |
| 324 | if self.canonical or self.column > self.best_width: |
| 325 | self.write_indent() |
| 326 | if not self.canonical and self.check_simple_key(): |
| 327 | self.states.append(self.expect_flow_mapping_simple_value) |
| 328 | self.expect_node(mapping=True, simple_key=True) |
| 329 | else: |
| 330 | self.write_indicator('?', True) |
| 331 | self.states.append(self.expect_flow_mapping_value) |
| 332 | self.expect_node(mapping=True) |
| 333 | |
| 334 | def expect_flow_mapping_key(self): |
| 335 | if isinstance(self.event, MappingEndEvent): |
| 336 | self.indent = self.indents.pop() |
| 337 | self.flow_level -= 1 |
| 338 | if self.canonical: |
| 339 | self.write_indicator(',', False) |
| 340 | self.write_indent() |
| 341 | self.write_indicator('}', False) |
| 342 | self.state = self.states.pop() |
| 343 | else: |
| 344 | self.write_indicator(',', False) |
| 345 | if self.canonical or self.column > self.best_width: |
| 346 | self.write_indent() |
| 347 | if not self.canonical and self.check_simple_key(): |
| 348 | self.states.append(self.expect_flow_mapping_simple_value) |
| 349 | self.expect_node(mapping=True, simple_key=True) |
| 350 | else: |
| 351 | self.write_indicator('?', True) |
| 352 | self.states.append(self.expect_flow_mapping_value) |
| 353 | self.expect_node(mapping=True) |
| 354 | |
| 355 | def expect_flow_mapping_simple_value(self): |
| 356 | self.write_indicator(':', False) |
| 357 | self.states.append(self.expect_flow_mapping_key) |
| 358 | self.expect_node(mapping=True) |
| 359 | |
| 360 | def expect_flow_mapping_value(self): |
| 361 | if self.canonical or self.column > self.best_width: |
| 362 | self.write_indent() |
| 363 | self.write_indicator(':', True) |
| 364 | self.states.append(self.expect_flow_mapping_key) |
| 365 | self.expect_node(mapping=True) |
| 366 | |
| 367 | # Block sequence handlers. |
| 368 | |
| 369 | def expect_block_sequence(self): |
| 370 | indentless = (self.mapping_context and not self.indention) |
| 371 | self.increase_indent(flow=False, indentless=indentless) |
| 372 | self.state = self.expect_first_block_sequence_item |
| 373 | |
| 374 | def expect_first_block_sequence_item(self): |
| 375 | return self.expect_block_sequence_item(first=True) |
| 376 | |
| 377 | def expect_block_sequence_item(self, first=False): |
| 378 | if not first and isinstance(self.event, SequenceEndEvent): |
| 379 | self.indent = self.indents.pop() |
| 380 | self.state = self.states.pop() |
| 381 | else: |
| 382 | self.write_indent() |
| 383 | self.write_indicator('-', True, indention=True) |
| 384 | self.states.append(self.expect_block_sequence_item) |
| 385 | self.expect_node(sequence=True) |
| 386 | |
| 387 | # Block mapping handlers. |
| 388 | |
| 389 | def expect_block_mapping(self): |
| 390 | self.increase_indent(flow=False) |
| 391 | self.state = self.expect_first_block_mapping_key |
| 392 | |
| 393 | def expect_first_block_mapping_key(self): |
| 394 | return self.expect_block_mapping_key(first=True) |
| 395 | |
| 396 | def expect_block_mapping_key(self, first=False): |
| 397 | if not first and isinstance(self.event, MappingEndEvent): |
| 398 | self.indent = self.indents.pop() |
| 399 | self.state = self.states.pop() |
| 400 | else: |
| 401 | self.write_indent() |
| 402 | if self.check_simple_key(): |
| 403 | self.states.append(self.expect_block_mapping_simple_value) |
| 404 | self.expect_node(mapping=True, simple_key=True) |
| 405 | else: |
| 406 | self.write_indicator('?', True, indention=True) |
| 407 | self.states.append(self.expect_block_mapping_value) |
| 408 | self.expect_node(mapping=True) |
| 409 | |
| 410 | def expect_block_mapping_simple_value(self): |
| 411 | self.write_indicator(':', False) |
| 412 | self.states.append(self.expect_block_mapping_key) |
| 413 | self.expect_node(mapping=True) |
| 414 | |
| 415 | def expect_block_mapping_value(self): |
| 416 | self.write_indent() |
| 417 | self.write_indicator(':', True, indention=True) |
| 418 | self.states.append(self.expect_block_mapping_key) |
| 419 | self.expect_node(mapping=True) |
| 420 | |
| 421 | # Checkers. |
| 422 | |
| 423 | def check_empty_sequence(self): |
| 424 | return (isinstance(self.event, SequenceStartEvent) and self.events |
| 425 | and isinstance(self.events[0], SequenceEndEvent)) |
| 426 | |
| 427 | def check_empty_mapping(self): |
| 428 | return (isinstance(self.event, MappingStartEvent) and self.events |
| 429 | and isinstance(self.events[0], MappingEndEvent)) |
| 430 | |
| 431 | def check_empty_document(self): |
| 432 | if not isinstance(self.event, DocumentStartEvent) or not self.events: |
| 433 | return False |
| 434 | event = self.events[0] |
| 435 | return (isinstance(event, ScalarEvent) and event.anchor is None |
| 436 | and event.tag is None and event.implicit and event.value == '') |
| 437 | |
| 438 | def check_simple_key(self): |
| 439 | length = 0 |
| 440 | if isinstance(self.event, NodeEvent) and self.event.anchor is not None: |
| 441 | if self.prepared_anchor is None: |
| 442 | self.prepared_anchor = self.prepare_anchor(self.event.anchor) |
| 443 | length += len(self.prepared_anchor) |
| 444 | if isinstance(self.event, (ScalarEvent, CollectionStartEvent)) \ |
| 445 | and self.event.tag is not None: |
| 446 | if self.prepared_tag is None: |
| 447 | self.prepared_tag = self.prepare_tag(self.event.tag) |
| 448 | length += len(self.prepared_tag) |
| 449 | if isinstance(self.event, ScalarEvent): |
| 450 | if self.analysis is None: |
| 451 | self.analysis = self.analyze_scalar(self.event.value) |
| 452 | length += len(self.analysis.scalar) |
| 453 | return (length < 128 and (isinstance(self.event, AliasEvent) |
| 454 | or (isinstance(self.event, ScalarEvent) |
| 455 | and not self.analysis.empty and not self.analysis.multiline) |
| 456 | or self.check_empty_sequence() or self.check_empty_mapping())) |
| 457 | |
| 458 | # Anchor, Tag, and Scalar processors. |
| 459 | |
| 460 | def process_anchor(self, indicator): |
| 461 | if self.event.anchor is None: |
| 462 | self.prepared_anchor = None |
| 463 | return |
| 464 | if self.prepared_anchor is None: |
| 465 | self.prepared_anchor = self.prepare_anchor(self.event.anchor) |
| 466 | if self.prepared_anchor: |
| 467 | self.write_indicator(indicator+self.prepared_anchor, True) |
| 468 | self.prepared_anchor = None |
| 469 | |
| 470 | def process_tag(self): |
| 471 | tag = self.event.tag |
| 472 | if isinstance(self.event, ScalarEvent): |
| 473 | if self.style is None: |
| 474 | self.style = self.choose_scalar_style() |
| 475 | if ((not self.canonical or tag is None) and |
| 476 | ((self.style == '' and self.event.implicit[0]) |
| 477 | or (self.style != '' and self.event.implicit[1]))): |
| 478 | self.prepared_tag = None |
| 479 | return |
| 480 | if self.event.implicit[0] and tag is None: |
| 481 | tag = '!' |
| 482 | self.prepared_tag = None |
| 483 | else: |
| 484 | if (not self.canonical or tag is None) and self.event.implicit: |
| 485 | self.prepared_tag = None |
| 486 | return |
| 487 | if tag is None: |
| 488 | raise EmitterError("tag is not specified") |
| 489 | if self.prepared_tag is None: |
| 490 | self.prepared_tag = self.prepare_tag(tag) |
| 491 | if self.prepared_tag: |
| 492 | self.write_indicator(self.prepared_tag, True) |
| 493 | self.prepared_tag = None |
| 494 | |
| 495 | def choose_scalar_style(self): |
| 496 | if self.analysis is None: |
| 497 | self.analysis = self.analyze_scalar(self.event.value) |
| 498 | if self.event.style == '"' or self.canonical: |
| 499 | return '"' |
| 500 | if not self.event.style and self.event.implicit[0]: |
| 501 | if (not (self.simple_key_context and |
| 502 | (self.analysis.empty or self.analysis.multiline)) |
| 503 | and (self.flow_level and self.analysis.allow_flow_plain |
| 504 | or (not self.flow_level and self.analysis.allow_block_plain))): |
| 505 | return '' |
| 506 | if self.event.style and self.event.style in '|>': |
| 507 | if (not self.flow_level and not self.simple_key_context |
| 508 | and self.analysis.allow_block): |
| 509 | return self.event.style |
| 510 | if not self.event.style or self.event.style == '\'': |
| 511 | if (self.analysis.allow_single_quoted and |
| 512 | not (self.simple_key_context and self.analysis.multiline)): |
| 513 | return '\'' |
| 514 | return '"' |
| 515 | |
| 516 | def process_scalar(self): |
| 517 | if self.analysis is None: |
| 518 | self.analysis = self.analyze_scalar(self.event.value) |
| 519 | if self.style is None: |
| 520 | self.style = self.choose_scalar_style() |
| 521 | split = (not self.simple_key_context) |
| 522 | #if self.analysis.multiline and split \ |
| 523 | # and (not self.style or self.style in '\'\"'): |
| 524 | # self.write_indent() |
| 525 | if self.style == '"': |
| 526 | self.write_double_quoted(self.analysis.scalar, split) |
| 527 | elif self.style == '\'': |
| 528 | self.write_single_quoted(self.analysis.scalar, split) |
| 529 | elif self.style == '>': |
| 530 | self.write_folded(self.analysis.scalar) |
| 531 | elif self.style == '|': |
| 532 | self.write_literal(self.analysis.scalar) |
| 533 | else: |
| 534 | self.write_plain(self.analysis.scalar, split) |
| 535 | self.analysis = None |
| 536 | self.style = None |
| 537 | |
| 538 | # Analyzers. |
| 539 | |
| 540 | def prepare_version(self, version): |
| 541 | major, minor = version |
| 542 | if major != 1: |
| 543 | raise EmitterError("unsupported YAML version: %d.%d" % (major, minor)) |
| 544 | return '%d.%d' % (major, minor) |
| 545 | |
| 546 | def prepare_tag_handle(self, handle): |
| 547 | if not handle: |
| 548 | raise EmitterError("tag handle must not be empty") |
| 549 | if handle[0] != '!' or handle[-1] != '!': |
| 550 | raise EmitterError("tag handle must start and end with '!': %r" % handle) |
| 551 | for ch in handle[1:-1]: |
| 552 | if not ('0' <= ch <= '9' or 'A' <= ch <= 'Z' or 'a' <= ch <= 'z' \ |
| 553 | or ch in '-_'): |
| 554 | raise EmitterError("invalid character %r in the tag handle: %r" |
| 555 | % (ch, handle)) |
| 556 | return handle |
| 557 | |
| 558 | def prepare_tag_prefix(self, prefix): |
| 559 | if not prefix: |
| 560 | raise EmitterError("tag prefix must not be empty") |
| 561 | chunks = [] |
| 562 | start = end = 0 |
| 563 | if prefix[0] == '!': |
| 564 | end = 1 |
| 565 | while end < len(prefix): |
| 566 | ch = prefix[end] |
| 567 | if '0' <= ch <= '9' or 'A' <= ch <= 'Z' or 'a' <= ch <= 'z' \ |
| 568 | or ch in '-;/?!:@&=+$,_.~*\'()[]': |
| 569 | end += 1 |
| 570 | else: |
| 571 | if start < end: |
| 572 | chunks.append(prefix[start:end]) |
| 573 | start = end = end+1 |
| 574 | data = ch.encode('utf-8') |
| 575 | for ch in data: |
| 576 | chunks.append('%%%02X' % ord(ch)) |
| 577 | if start < end: |
| 578 | chunks.append(prefix[start:end]) |
| 579 | return ''.join(chunks) |
| 580 | |
| 581 | def prepare_tag(self, tag): |
| 582 | if not tag: |
| 583 | raise EmitterError("tag must not be empty") |
| 584 | if tag == '!': |
| 585 | return tag |
| 586 | handle = None |
| 587 | suffix = tag |
| 588 | prefixes = sorted(self.tag_prefixes.keys()) |
| 589 | for prefix in prefixes: |
| 590 | if tag.startswith(prefix) \ |
| 591 | and (prefix == '!' or len(prefix) < len(tag)): |
| 592 | handle = self.tag_prefixes[prefix] |
| 593 | suffix = tag[len(prefix):] |
| 594 | chunks = [] |
| 595 | start = end = 0 |
| 596 | while end < len(suffix): |
| 597 | ch = suffix[end] |
| 598 | if '0' <= ch <= '9' or 'A' <= ch <= 'Z' or 'a' <= ch <= 'z' \ |
| 599 | or ch in '-;/?:@&=+$,_.~*\'()[]' \ |
| 600 | or (ch == '!' and handle != '!'): |
| 601 | end += 1 |
| 602 | else: |
| 603 | if start < end: |
| 604 | chunks.append(suffix[start:end]) |
| 605 | start = end = end+1 |
| 606 | data = ch.encode('utf-8') |
| 607 | for ch in data: |
| 608 | chunks.append('%%%02X' % ord(ch)) |
| 609 | if start < end: |
| 610 | chunks.append(suffix[start:end]) |
| 611 | suffix_text = ''.join(chunks) |
| 612 | if handle: |
| 613 | return '%s%s' % (handle, suffix_text) |
| 614 | else: |
| 615 | return '!<%s>' % suffix_text |
| 616 | |
| 617 | def prepare_anchor(self, anchor): |
| 618 | if not anchor: |
| 619 | raise EmitterError("anchor must not be empty") |
| 620 | for ch in anchor: |
| 621 | if not ('0' <= ch <= '9' or 'A' <= ch <= 'Z' or 'a' <= ch <= 'z' \ |
| 622 | or ch in '-_'): |
| 623 | raise EmitterError("invalid character %r in the anchor: %r" |
| 624 | % (ch, anchor)) |
| 625 | return anchor |
| 626 | |
| 627 | def analyze_scalar(self, scalar): |
| 628 | |
| 629 | # Empty scalar is a special case. |
| 630 | if not scalar: |
| 631 | return ScalarAnalysis(scalar=scalar, empty=True, multiline=False, |
| 632 | allow_flow_plain=False, allow_block_plain=True, |
| 633 | allow_single_quoted=True, allow_double_quoted=True, |
| 634 | allow_block=False) |
| 635 | |
| 636 | # Indicators and special characters. |
| 637 | block_indicators = False |
| 638 | flow_indicators = False |
| 639 | line_breaks = False |
| 640 | special_characters = False |
| 641 | |
| 642 | # Important whitespace combinations. |
| 643 | leading_space = False |
| 644 | leading_break = False |
| 645 | trailing_space = False |
| 646 | trailing_break = False |
| 647 | break_space = False |
| 648 | space_break = False |
| 649 | |
| 650 | # Check document indicators. |
| 651 | if scalar.startswith('---') or scalar.startswith('...'): |
| 652 | block_indicators = True |
| 653 | flow_indicators = True |
| 654 | |
| 655 | # First character or preceded by a whitespace. |
| 656 | preceeded_by_whitespace = True |
| 657 | |
| 658 | # Last character or followed by a whitespace. |
| 659 | followed_by_whitespace = (len(scalar) == 1 or |
| 660 | scalar[1] in '\0 \t\r\n\x85\u2028\u2029') |
| 661 | |
| 662 | # The previous character is a space. |
| 663 | previous_space = False |
| 664 | |
| 665 | # The previous character is a break. |
| 666 | previous_break = False |
| 667 | |
| 668 | index = 0 |
| 669 | while index < len(scalar): |
| 670 | ch = scalar[index] |
| 671 | |
| 672 | # Check for indicators. |
| 673 | if index == 0: |
| 674 | # Leading indicators are special characters. |
| 675 | if ch in '#,[]{}&*!|>\'\"%@`': |
| 676 | flow_indicators = True |
| 677 | block_indicators = True |
| 678 | if ch in '?:': |
| 679 | flow_indicators = True |
| 680 | if followed_by_whitespace: |
| 681 | block_indicators = True |
| 682 | if ch == '-' and followed_by_whitespace: |
| 683 | flow_indicators = True |
| 684 | block_indicators = True |
| 685 | else: |
| 686 | # Some indicators cannot appear within a scalar as well. |
| 687 | if ch in ',?[]{}': |
| 688 | flow_indicators = True |
| 689 | if ch == ':': |
| 690 | flow_indicators = True |
| 691 | if followed_by_whitespace: |
| 692 | block_indicators = True |
| 693 | if ch == '#' and preceeded_by_whitespace: |
| 694 | flow_indicators = True |
| 695 | block_indicators = True |
| 696 | |
| 697 | # Check for line breaks, special, and unicode characters. |
| 698 | if ch in '\n\x85\u2028\u2029': |
| 699 | line_breaks = True |
| 700 | if not (ch == '\n' or '\x20' <= ch <= '\x7E'): |
| 701 | if (ch == '\x85' or '\xA0' <= ch <= '\uD7FF' |
| 702 | or '\uE000' <= ch <= '\uFFFD') and ch != '\uFEFF': |
| 703 | unicode_characters = True |
| 704 | if not self.allow_unicode: |
| 705 | special_characters = True |
| 706 | else: |
| 707 | special_characters = True |
| 708 | |
| 709 | # Detect important whitespace combinations. |
| 710 | if ch == ' ': |
| 711 | if index == 0: |
| 712 | leading_space = True |
| 713 | if index == len(scalar)-1: |
| 714 | trailing_space = True |
| 715 | if previous_break: |
| 716 | break_space = True |
| 717 | previous_space = True |
| 718 | previous_break = False |
| 719 | elif ch in '\n\x85\u2028\u2029': |
| 720 | if index == 0: |
| 721 | leading_break = True |
| 722 | if index == len(scalar)-1: |
| 723 | trailing_break = True |
| 724 | if previous_space: |
| 725 | space_break = True |
| 726 | previous_space = False |
| 727 | previous_break = True |
| 728 | else: |
| 729 | previous_space = False |
| 730 | previous_break = False |
| 731 | |
| 732 | # Prepare for the next character. |
| 733 | index += 1 |
| 734 | preceeded_by_whitespace = (ch in '\0 \t\r\n\x85\u2028\u2029') |
| 735 | followed_by_whitespace = (index+1 >= len(scalar) or |
| 736 | scalar[index+1] in '\0 \t\r\n\x85\u2028\u2029') |
| 737 | |
| 738 | # Let's decide what styles are allowed. |
| 739 | allow_flow_plain = True |
| 740 | allow_block_plain = True |
| 741 | allow_single_quoted = True |
| 742 | allow_double_quoted = True |
| 743 | allow_block = True |
| 744 | |
| 745 | # Leading and trailing whitespaces are bad for plain scalars. |
| 746 | if (leading_space or leading_break |
| 747 | or trailing_space or trailing_break): |
| 748 | allow_flow_plain = allow_block_plain = False |
| 749 | |
| 750 | # We do not permit trailing spaces for block scalars. |
| 751 | if trailing_space: |
| 752 | allow_block = False |
| 753 | |
| 754 | # Spaces at the beginning of a new line are only acceptable for block |
| 755 | # scalars. |
| 756 | if break_space: |
| 757 | allow_flow_plain = allow_block_plain = allow_single_quoted = False |
| 758 | |
| 759 | # Spaces followed by breaks, as well as special character are only |
| 760 | # allowed for double quoted scalars. |
| 761 | if space_break or special_characters: |
| 762 | allow_flow_plain = allow_block_plain = \ |
| 763 | allow_single_quoted = allow_block = False |
| 764 | |
| 765 | # Although the plain scalar writer supports breaks, we never emit |
| 766 | # multiline plain scalars. |
| 767 | if line_breaks: |
| 768 | allow_flow_plain = allow_block_plain = False |
| 769 | |
| 770 | # Flow indicators are forbidden for flow plain scalars. |
| 771 | if flow_indicators: |
| 772 | allow_flow_plain = False |
| 773 | |
| 774 | # Block indicators are forbidden for block plain scalars. |
| 775 | if block_indicators: |
| 776 | allow_block_plain = False |
| 777 | |
| 778 | return ScalarAnalysis(scalar=scalar, |
| 779 | empty=False, multiline=line_breaks, |
| 780 | allow_flow_plain=allow_flow_plain, |
| 781 | allow_block_plain=allow_block_plain, |
| 782 | allow_single_quoted=allow_single_quoted, |
| 783 | allow_double_quoted=allow_double_quoted, |
| 784 | allow_block=allow_block) |
| 785 | |
| 786 | # Writers. |
| 787 | |
| 788 | def flush_stream(self): |
| 789 | if hasattr(self.stream, 'flush'): |
| 790 | self.stream.flush() |
| 791 | |
| 792 | def write_stream_start(self): |
| 793 | # Write BOM if needed. |
| 794 | if self.encoding and self.encoding.startswith('utf-16'): |
| 795 | self.stream.write('\uFEFF'.encode(self.encoding)) |
| 796 | |
| 797 | def write_stream_end(self): |
| 798 | self.flush_stream() |
| 799 | |
| 800 | def write_indicator(self, indicator, need_whitespace, |
| 801 | whitespace=False, indention=False): |
| 802 | if self.whitespace or not need_whitespace: |
| 803 | data = indicator |
| 804 | else: |
| 805 | data = ' '+indicator |
| 806 | self.whitespace = whitespace |
| 807 | self.indention = self.indention and indention |
| 808 | self.column += len(data) |
| 809 | self.open_ended = False |
| 810 | if self.encoding: |
| 811 | data = data.encode(self.encoding) |
| 812 | self.stream.write(data) |
| 813 | |
| 814 | def write_indent(self): |
| 815 | indent = self.indent or 0 |
| 816 | if not self.indention or self.column > indent \ |
| 817 | or (self.column == indent and not self.whitespace): |
| 818 | self.write_line_break() |
| 819 | if self.column < indent: |
| 820 | self.whitespace = True |
| 821 | data = ' '*(indent-self.column) |
| 822 | self.column = indent |
| 823 | if self.encoding: |
| 824 | data = data.encode(self.encoding) |
| 825 | self.stream.write(data) |
| 826 | |
| 827 | def write_line_break(self, data=None): |
| 828 | if data is None: |
| 829 | data = self.best_line_break |
| 830 | self.whitespace = True |
| 831 | self.indention = True |
| 832 | self.line += 1 |
| 833 | self.column = 0 |
| 834 | if self.encoding: |
| 835 | data = data.encode(self.encoding) |
| 836 | self.stream.write(data) |
| 837 | |
| 838 | def write_version_directive(self, version_text): |
| 839 | data = '%%YAML %s' % version_text |
| 840 | if self.encoding: |
| 841 | data = data.encode(self.encoding) |
| 842 | self.stream.write(data) |
| 843 | self.write_line_break() |
| 844 | |
| 845 | def write_tag_directive(self, handle_text, prefix_text): |
| 846 | data = '%%TAG %s %s' % (handle_text, prefix_text) |
| 847 | if self.encoding: |
| 848 | data = data.encode(self.encoding) |
| 849 | self.stream.write(data) |
| 850 | self.write_line_break() |
| 851 | |
| 852 | # Scalar streams. |
| 853 | |
| 854 | def write_single_quoted(self, text, split=True): |
| 855 | self.write_indicator('\'', True) |
| 856 | spaces = False |
| 857 | breaks = False |
| 858 | start = end = 0 |
| 859 | while end <= len(text): |
| 860 | ch = None |
| 861 | if end < len(text): |
| 862 | ch = text[end] |
| 863 | if spaces: |
| 864 | if ch is None or ch != ' ': |
| 865 | if start+1 == end and self.column > self.best_width and split \ |
| 866 | and start != 0 and end != len(text): |
| 867 | self.write_indent() |
| 868 | else: |
| 869 | data = text[start:end] |
| 870 | self.column += len(data) |
| 871 | if self.encoding: |
| 872 | data = data.encode(self.encoding) |
| 873 | self.stream.write(data) |
| 874 | start = end |
| 875 | elif breaks: |
| 876 | if ch is None or ch not in '\n\x85\u2028\u2029': |
| 877 | if text[start] == '\n': |
| 878 | self.write_line_break() |
| 879 | for br in text[start:end]: |
| 880 | if br == '\n': |
| 881 | self.write_line_break() |
| 882 | else: |
| 883 | self.write_line_break(br) |
| 884 | self.write_indent() |
| 885 | start = end |
| 886 | else: |
| 887 | if ch is None or ch in ' \n\x85\u2028\u2029' or ch == '\'': |
| 888 | if start < end: |
| 889 | data = text[start:end] |
| 890 | self.column += len(data) |
| 891 | if self.encoding: |
| 892 | data = data.encode(self.encoding) |
| 893 | self.stream.write(data) |
| 894 | start = end |
| 895 | if ch == '\'': |
| 896 | data = '\'\'' |
| 897 | self.column += 2 |
| 898 | if self.encoding: |
| 899 | data = data.encode(self.encoding) |
| 900 | self.stream.write(data) |
| 901 | start = end + 1 |
| 902 | if ch is not None: |
| 903 | spaces = (ch == ' ') |
| 904 | breaks = (ch in '\n\x85\u2028\u2029') |
| 905 | end += 1 |
| 906 | self.write_indicator('\'', False) |
| 907 | |
| 908 | ESCAPE_REPLACEMENTS = { |
| 909 | '\0': '0', |
| 910 | '\x07': 'a', |
| 911 | '\x08': 'b', |
| 912 | '\x09': 't', |
| 913 | '\x0A': 'n', |
| 914 | '\x0B': 'v', |
| 915 | '\x0C': 'f', |
| 916 | '\x0D': 'r', |
| 917 | '\x1B': 'e', |
| 918 | '\"': '\"', |
| 919 | '\\': '\\', |
| 920 | '\x85': 'N', |
| 921 | '\xA0': '_', |
| 922 | '\u2028': 'L', |
| 923 | '\u2029': 'P', |
| 924 | } |
| 925 | |
| 926 | def write_double_quoted(self, text, split=True): |
| 927 | self.write_indicator('"', True) |
| 928 | start = end = 0 |
| 929 | while end <= len(text): |
| 930 | ch = None |
| 931 | if end < len(text): |
| 932 | ch = text[end] |
| 933 | if ch is None or ch in '"\\\x85\u2028\u2029\uFEFF' \ |
| 934 | or not ('\x20' <= ch <= '\x7E' |
| 935 | or (self.allow_unicode |
| 936 | and ('\xA0' <= ch <= '\uD7FF' |
| 937 | or '\uE000' <= ch <= '\uFFFD'))): |
| 938 | if start < end: |
| 939 | data = text[start:end] |
| 940 | self.column += len(data) |
| 941 | if self.encoding: |
| 942 | data = data.encode(self.encoding) |
| 943 | self.stream.write(data) |
| 944 | start = end |
| 945 | if ch is not None: |
| 946 | if ch in self.ESCAPE_REPLACEMENTS: |
| 947 | data = '\\'+self.ESCAPE_REPLACEMENTS[ch] |
| 948 | elif ch <= '\xFF': |
| 949 | data = '\\x%02X' % ord(ch) |
| 950 | elif ch <= '\uFFFF': |
| 951 | data = '\\u%04X' % ord(ch) |
| 952 | else: |
| 953 | data = '\\U%08X' % ord(ch) |
| 954 | self.column += len(data) |
| 955 | if self.encoding: |
| 956 | data = data.encode(self.encoding) |
| 957 | self.stream.write(data) |
| 958 | start = end+1 |
| 959 | if 0 < end < len(text)-1 and (ch == ' ' or start >= end) \ |
| 960 | and self.column+(end-start) > self.best_width and split: |
| 961 | data = text[start:end]+'\\' |
| 962 | if start < end: |
| 963 | start = end |
| 964 | self.column += len(data) |
| 965 | if self.encoding: |
| 966 | data = data.encode(self.encoding) |
| 967 | self.stream.write(data) |
| 968 | self.write_indent() |
| 969 | self.whitespace = False |
| 970 | self.indention = False |
| 971 | if text[start] == ' ': |
| 972 | data = '\\' |
| 973 | self.column += len(data) |
| 974 | if self.encoding: |
| 975 | data = data.encode(self.encoding) |
| 976 | self.stream.write(data) |
| 977 | end += 1 |
| 978 | self.write_indicator('"', False) |
| 979 | |
| 980 | def determine_block_hints(self, text): |
| 981 | hints = '' |
| 982 | if text: |
| 983 | if text[0] in ' \n\x85\u2028\u2029': |
| 984 | hints += str(self.best_indent) |
| 985 | if text[-1] not in '\n\x85\u2028\u2029': |
| 986 | hints += '-' |
| 987 | elif len(text) == 1 or text[-2] in '\n\x85\u2028\u2029': |
| 988 | hints += '+' |
| 989 | return hints |
| 990 | |
| 991 | def write_folded(self, text): |
| 992 | hints = self.determine_block_hints(text) |
| 993 | self.write_indicator('>'+hints, True) |
| 994 | if hints[-1:] == '+': |
| 995 | self.open_ended = True |
| 996 | self.write_line_break() |
| 997 | leading_space = True |
| 998 | spaces = False |
| 999 | breaks = True |
| 1000 | start = end = 0 |
| 1001 | while end <= len(text): |
| 1002 | ch = None |
| 1003 | if end < len(text): |
| 1004 | ch = text[end] |
| 1005 | if breaks: |
| 1006 | if ch is None or ch not in '\n\x85\u2028\u2029': |
| 1007 | if not leading_space and ch is not None and ch != ' ' \ |
| 1008 | and text[start] == '\n': |
| 1009 | self.write_line_break() |
| 1010 | leading_space = (ch == ' ') |
| 1011 | for br in text[start:end]: |
| 1012 | if br == '\n': |
| 1013 | self.write_line_break() |
| 1014 | else: |
| 1015 | self.write_line_break(br) |
| 1016 | if ch is not None: |
| 1017 | self.write_indent() |
| 1018 | start = end |
| 1019 | elif spaces: |
| 1020 | if ch != ' ': |
| 1021 | if start+1 == end and self.column > self.best_width: |
| 1022 | self.write_indent() |
| 1023 | else: |
| 1024 | data = text[start:end] |
| 1025 | self.column += len(data) |
| 1026 | if self.encoding: |
| 1027 | data = data.encode(self.encoding) |
| 1028 | self.stream.write(data) |
| 1029 | start = end |
| 1030 | else: |
| 1031 | if ch is None or ch in ' \n\x85\u2028\u2029': |
| 1032 | data = text[start:end] |
| 1033 | self.column += len(data) |
| 1034 | if self.encoding: |
| 1035 | data = data.encode(self.encoding) |
| 1036 | self.stream.write(data) |
| 1037 | if ch is None: |
| 1038 | self.write_line_break() |
| 1039 | start = end |
| 1040 | if ch is not None: |
| 1041 | breaks = (ch in '\n\x85\u2028\u2029') |
| 1042 | spaces = (ch == ' ') |
| 1043 | end += 1 |
| 1044 | |
| 1045 | def write_literal(self, text): |
| 1046 | hints = self.determine_block_hints(text) |
| 1047 | self.write_indicator('|'+hints, True) |
| 1048 | if hints[-1:] == '+': |
| 1049 | self.open_ended = True |
| 1050 | self.write_line_break() |
| 1051 | breaks = True |
| 1052 | start = end = 0 |
| 1053 | while end <= len(text): |
| 1054 | ch = None |
| 1055 | if end < len(text): |
| 1056 | ch = text[end] |
| 1057 | if breaks: |
| 1058 | if ch is None or ch not in '\n\x85\u2028\u2029': |
| 1059 | for br in text[start:end]: |
| 1060 | if br == '\n': |
| 1061 | self.write_line_break() |
| 1062 | else: |
| 1063 | self.write_line_break(br) |
| 1064 | if ch is not None: |
| 1065 | self.write_indent() |
| 1066 | start = end |
| 1067 | else: |
| 1068 | if ch is None or ch in '\n\x85\u2028\u2029': |
| 1069 | data = text[start:end] |
| 1070 | if self.encoding: |
| 1071 | data = data.encode(self.encoding) |
| 1072 | self.stream.write(data) |
| 1073 | if ch is None: |
| 1074 | self.write_line_break() |
| 1075 | start = end |
| 1076 | if ch is not None: |
| 1077 | breaks = (ch in '\n\x85\u2028\u2029') |
| 1078 | end += 1 |
| 1079 | |
| 1080 | def write_plain(self, text, split=True): |
| 1081 | if self.root_context: |
| 1082 | self.open_ended = True |
| 1083 | if not text: |
| 1084 | return |
| 1085 | if not self.whitespace: |
| 1086 | data = ' ' |
| 1087 | self.column += len(data) |
| 1088 | if self.encoding: |
| 1089 | data = data.encode(self.encoding) |
| 1090 | self.stream.write(data) |
| 1091 | self.whitespace = False |
| 1092 | self.indention = False |
| 1093 | spaces = False |
| 1094 | breaks = False |
| 1095 | start = end = 0 |
| 1096 | while end <= len(text): |
| 1097 | ch = None |
| 1098 | if end < len(text): |
| 1099 | ch = text[end] |
| 1100 | if spaces: |
| 1101 | if ch != ' ': |
| 1102 | if start+1 == end and self.column > self.best_width and split: |
| 1103 | self.write_indent() |
| 1104 | self.whitespace = False |
| 1105 | self.indention = False |
| 1106 | else: |
| 1107 | data = text[start:end] |
| 1108 | self.column += len(data) |
| 1109 | if self.encoding: |
| 1110 | data = data.encode(self.encoding) |
| 1111 | self.stream.write(data) |
| 1112 | start = end |
| 1113 | elif breaks: |
| 1114 | if ch not in '\n\x85\u2028\u2029': |
| 1115 | if text[start] == '\n': |
| 1116 | self.write_line_break() |
| 1117 | for br in text[start:end]: |
| 1118 | if br == '\n': |
| 1119 | self.write_line_break() |
| 1120 | else: |
| 1121 | self.write_line_break(br) |
| 1122 | self.write_indent() |
| 1123 | self.whitespace = False |
| 1124 | self.indention = False |
| 1125 | start = end |
| 1126 | else: |
| 1127 | if ch is None or ch in ' \n\x85\u2028\u2029': |
| 1128 | data = text[start:end] |
| 1129 | self.column += len(data) |
| 1130 | if self.encoding: |
| 1131 | data = data.encode(self.encoding) |
| 1132 | self.stream.write(data) |
| 1133 | start = end |
| 1134 | if ch is not None: |
| 1135 | spaces = (ch == ' ') |
| 1136 | breaks = (ch in '\n\x85\u2028\u2029') |
| 1137 | end += 1 |
| 1138 |