2 # Scanner produces tokens of the following types:
5 # DIRECTIVE(name, value)
22 # SCALAR(value, plain, style)
24 # Read comments in the Scanner code for more details.
27 __all__ = ['Scanner', 'ScannerError']
29 from error import MarkedYAMLError
32 class ScannerError(MarkedYAMLError):
35 class SimpleKey(object):
36 # See below simple keys treatment.
38 def __init__(self, token_number, required, index, line, column, mark):
39 self.token_number = token_number
40 self.required = required
46 class Scanner(object):
49 """Initialize the scanner."""
50 # It is assumed that Scanner and Reader will have a common descendant.
51 # Reader do the dirty work of checking for BOM and converting the
52 # input data to Unicode. It also adds NUL to the end.
54 # Reader supports the following methods
55 # self.peek(i=0) # peek the next i-th character
56 # self.prefix(l=1) # peek the next l characters
57 # self.forward(l=1) # read the next l characters and move the pointer.
59 # Had we reached the end of the stream?
62 # The number of unclosed '{' and '['. `flow_level == 0` means block
66 # List of processed tokens that are not yet emitted.
69 # Add the STREAM-START token.
70 self.fetch_stream_start()
72 # Number of tokens that were emitted through the `get_token` method.
75 # The current indentation level.
78 # Past indentation levels.
81 # Variables related to simple keys treatment.
83 # A simple key is a key that is not denoted by the '?' indicator.
84 # Example of simple keys:
86 # block simple key: value
88 # : { flow simple key: value }
89 # We emit the KEY token before all keys, so when we find a potential
90 # simple key, we try to locate the corresponding ':' indicator.
91 # Simple keys should be limited to a single line and 1024 characters.
93 # Can a simple key start at the current position? A simple key may
95 # - at the beginning of the line, not counting indentation spaces
97 # - after '{', '[', ',' (in the flow context),
98 # - after '?', ':', '-' (in the block context).
99 # In the block context, this flag also signifies if a block collection
100 # may start at the current position.
101 self.allow_simple_key = True
103 # Keep track of possible simple keys. This is a dictionary. The key
104 # is `flow_level`; there can be no more that one possible simple key
105 # for each level. The value is a SimpleKey record:
106 # (token_number, required, index, line, column, mark)
107 # A simple key may start with ALIAS, ANCHOR, TAG, SCALAR(flow),
108 # '[', or '{' tokens.
109 self.possible_simple_keys = {}
113 def check_token(self, *choices):
114 # Check if the next token is one of the given types.
115 while self.need_more_tokens():
116 self.fetch_more_tokens()
120 for choice in choices:
121 if isinstance(self.tokens[0], choice):
125 def peek_token(self):
126 # Return the next token, but do not delete if from the queue.
127 while self.need_more_tokens():
128 self.fetch_more_tokens()
130 return self.tokens[0]
133 # Return the next token.
134 while self.need_more_tokens():
135 self.fetch_more_tokens()
137 self.tokens_taken += 1
138 return self.tokens.pop(0)
142 def need_more_tokens(self):
147 # The current token may be a potential simple key, so we
148 # need to look further.
149 self.stale_possible_simple_keys()
150 if self.next_possible_simple_key() == self.tokens_taken:
153 def fetch_more_tokens(self):
155 # Eat whitespaces and comments until we reach the next token.
156 self.scan_to_next_token()
158 # Remove obsolete possible simple keys.
159 self.stale_possible_simple_keys()
161 # Compare the current indentation and column. It may add some tokens
162 # and decrease the current indentation level.
163 self.unwind_indent(self.column)
165 # Peek the next character.
168 # Is it the end of stream?
170 return self.fetch_stream_end()
173 if ch == u'%' and self.check_directive():
174 return self.fetch_directive()
176 # Is it the document start?
177 if ch == u'-' and self.check_document_start():
178 return self.fetch_document_start()
180 # Is it the document end?
181 if ch == u'.' and self.check_document_end():
182 return self.fetch_document_end()
184 # TODO: support for BOM within a stream.
186 # return self.fetch_bom() <-- issue BOMToken
188 # Note: the order of the following checks is NOT significant.
190 # Is it the flow sequence start indicator?
192 return self.fetch_flow_sequence_start()
194 # Is it the flow mapping start indicator?
196 return self.fetch_flow_mapping_start()
198 # Is it the flow sequence end indicator?
200 return self.fetch_flow_sequence_end()
202 # Is it the flow mapping end indicator?
204 return self.fetch_flow_mapping_end()
206 # Is it the flow entry indicator?
208 return self.fetch_flow_entry()
210 # Is it the block entry indicator?
211 if ch == u'-' and self.check_block_entry():
212 return self.fetch_block_entry()
214 # Is it the key indicator?
215 if ch == u'?' and self.check_key():
216 return self.fetch_key()
218 # Is it the value indicator?
219 if ch == u':' and self.check_value():
220 return self.fetch_value()
224 return self.fetch_alias()
228 return self.fetch_anchor()
232 return self.fetch_tag()
234 # Is it a literal scalar?
235 if ch == u'|' and not self.flow_level:
236 return self.fetch_literal()
238 # Is it a folded scalar?
239 if ch == u'>' and not self.flow_level:
240 return self.fetch_folded()
242 # Is it a single quoted scalar?
244 return self.fetch_single()
246 # Is it a double quoted scalar?
248 return self.fetch_double()
250 # It must be a plain scalar then.
251 if self.check_plain():
252 return self.fetch_plain()
254 # No? It's an error. Let's produce a nice error message.
255 raise ScannerError("while scanning for the next token", None,
256 "found character %r that cannot start any token"
257 % ch.encode('utf-8'), self.get_mark())
259 # Simple keys treatment.
261 def next_possible_simple_key(self):
262 # Return the number of the nearest possible simple key. Actually we
263 # don't need to loop through the whole dictionary. We may replace it
264 # with the following code:
265 # if not self.possible_simple_keys:
267 # return self.possible_simple_keys[
268 # min(self.possible_simple_keys.keys())].token_number
269 min_token_number = None
270 for level in self.possible_simple_keys:
271 key = self.possible_simple_keys[level]
272 if min_token_number is None or key.token_number < min_token_number:
273 min_token_number = key.token_number
274 return min_token_number
276 def stale_possible_simple_keys(self):
277 # Remove entries that are no longer possible simple keys. According to
278 # the YAML specification, simple keys
279 # - should be limited to a single line,
280 # - should be no longer than 1024 characters.
281 # Disabling this procedure will allow simple keys of any length and
282 # height (may cause problems if indentation is broken though).
283 for level in self.possible_simple_keys.keys():
284 key = self.possible_simple_keys[level]
285 if key.line != self.line \
286 or self.index-key.index > 1024:
288 raise ScannerError("while scanning a simple key", key.mark,
289 "could not found expected ':'", self.get_mark())
290 del self.possible_simple_keys[level]
292 def save_possible_simple_key(self):
293 # The next token may start a simple key. We check if it's possible
294 # and save its position. This function is called for
295 # ALIAS, ANCHOR, TAG, SCALAR(flow), '[', and '{'.
297 # Check if a simple key is required at the current position.
298 required = not self.flow_level and self.indent == self.column
300 # A simple key is required only if it is the first token in the current
301 # line. Therefore it is always allowed.
302 assert self.allow_simple_key or not required
304 # The next token might be a simple key. Let's save it's number and
306 if self.allow_simple_key:
307 self.remove_possible_simple_key()
308 token_number = self.tokens_taken+len(self.tokens)
309 key = SimpleKey(token_number, required,
310 self.index, self.line, self.column, self.get_mark())
311 self.possible_simple_keys[self.flow_level] = key
313 def remove_possible_simple_key(self):
314 # Remove the saved possible key position at the current flow level.
315 if self.flow_level in self.possible_simple_keys:
316 key = self.possible_simple_keys[self.flow_level]
319 raise ScannerError("while scanning a simple key", key.mark,
320 "could not found expected ':'", self.get_mark())
322 del self.possible_simple_keys[self.flow_level]
324 # Indentation functions.
326 def unwind_indent(self, column):
328 ## In flow context, tokens should respect indentation.
329 ## Actually the condition should be `self.indent >= column` according to
330 ## the spec. But this condition will prohibit intuitively correct
331 ## constructions such as
334 #if self.flow_level and self.indent > column:
335 # raise ScannerError(None, None,
336 # "invalid intendation or unclosed '[' or '{'",
339 # In the flow context, indentation is ignored. We make the scanner less
340 # restrictive then specification requires.
344 # In block context, we may need to issue the BLOCK-END tokens.
345 while self.indent > column:
346 mark = self.get_mark()
347 self.indent = self.indents.pop()
348 self.tokens.append(BlockEndToken(mark, mark))
350 def add_indent(self, column):
351 # Check if we need to increase indentation.
352 if self.indent < column:
353 self.indents.append(self.indent)
360 def fetch_stream_start(self):
361 # We always add STREAM-START as the first token and STREAM-END as the
365 mark = self.get_mark()
368 self.tokens.append(StreamStartToken(mark, mark,
369 encoding=self.encoding))
372 def fetch_stream_end(self):
374 # Set the current intendation to -1.
375 self.unwind_indent(-1)
378 self.remove_possible_simple_key()
379 self.allow_simple_key = False
380 self.possible_simple_keys = {}
383 mark = self.get_mark()
386 self.tokens.append(StreamEndToken(mark, mark))
388 # The steam is finished.
391 def fetch_directive(self):
393 # Set the current intendation to -1.
394 self.unwind_indent(-1)
397 self.remove_possible_simple_key()
398 self.allow_simple_key = False
400 # Scan and add DIRECTIVE.
401 self.tokens.append(self.scan_directive())
403 def fetch_document_start(self):
404 self.fetch_document_indicator(DocumentStartToken)
406 def fetch_document_end(self):
407 self.fetch_document_indicator(DocumentEndToken)
409 def fetch_document_indicator(self, TokenClass):
411 # Set the current intendation to -1.
412 self.unwind_indent(-1)
414 # Reset simple keys. Note that there could not be a block collection
416 self.remove_possible_simple_key()
417 self.allow_simple_key = False
419 # Add DOCUMENT-START or DOCUMENT-END.
420 start_mark = self.get_mark()
422 end_mark = self.get_mark()
423 self.tokens.append(TokenClass(start_mark, end_mark))
425 def fetch_flow_sequence_start(self):
426 self.fetch_flow_collection_start(FlowSequenceStartToken)
428 def fetch_flow_mapping_start(self):
429 self.fetch_flow_collection_start(FlowMappingStartToken)
431 def fetch_flow_collection_start(self, TokenClass):
433 # '[' and '{' may start a simple key.
434 self.save_possible_simple_key()
436 # Increase the flow level.
439 # Simple keys are allowed after '[' and '{'.
440 self.allow_simple_key = True
442 # Add FLOW-SEQUENCE-START or FLOW-MAPPING-START.
443 start_mark = self.get_mark()
445 end_mark = self.get_mark()
446 self.tokens.append(TokenClass(start_mark, end_mark))
448 def fetch_flow_sequence_end(self):
449 self.fetch_flow_collection_end(FlowSequenceEndToken)
451 def fetch_flow_mapping_end(self):
452 self.fetch_flow_collection_end(FlowMappingEndToken)
454 def fetch_flow_collection_end(self, TokenClass):
456 # Reset possible simple key on the current level.
457 self.remove_possible_simple_key()
459 # Decrease the flow level.
462 # No simple keys after ']' or '}'.
463 self.allow_simple_key = False
465 # Add FLOW-SEQUENCE-END or FLOW-MAPPING-END.
466 start_mark = self.get_mark()
468 end_mark = self.get_mark()
469 self.tokens.append(TokenClass(start_mark, end_mark))
471 def fetch_flow_entry(self):
473 # Simple keys are allowed after ','.
474 self.allow_simple_key = True
476 # Reset possible simple key on the current level.
477 self.remove_possible_simple_key()
480 start_mark = self.get_mark()
482 end_mark = self.get_mark()
483 self.tokens.append(FlowEntryToken(start_mark, end_mark))
485 def fetch_block_entry(self):
487 # Block context needs additional checks.
488 if not self.flow_level:
490 # Are we allowed to start a new entry?
491 if not self.allow_simple_key:
492 raise ScannerError(None, None,
493 "sequence entries are not allowed here",
496 # We may need to add BLOCK-SEQUENCE-START.
497 if self.add_indent(self.column):
498 mark = self.get_mark()
499 self.tokens.append(BlockSequenceStartToken(mark, mark))
501 # It's an error for the block entry to occur in the flow context,
502 # but we let the parser detect this.
506 # Simple keys are allowed after '-'.
507 self.allow_simple_key = True
509 # Reset possible simple key on the current level.
510 self.remove_possible_simple_key()
513 start_mark = self.get_mark()
515 end_mark = self.get_mark()
516 self.tokens.append(BlockEntryToken(start_mark, end_mark))
520 # Block context needs additional checks.
521 if not self.flow_level:
523 # Are we allowed to start a key (not nessesary a simple)?
524 if not self.allow_simple_key:
525 raise ScannerError(None, None,
526 "mapping keys are not allowed here",
529 # We may need to add BLOCK-MAPPING-START.
530 if self.add_indent(self.column):
531 mark = self.get_mark()
532 self.tokens.append(BlockMappingStartToken(mark, mark))
534 # Simple keys are allowed after '?' in the block context.
535 self.allow_simple_key = not self.flow_level
537 # Reset possible simple key on the current level.
538 self.remove_possible_simple_key()
541 start_mark = self.get_mark()
543 end_mark = self.get_mark()
544 self.tokens.append(KeyToken(start_mark, end_mark))
546 def fetch_value(self):
548 # Do we determine a simple key?
549 if self.flow_level in self.possible_simple_keys:
552 key = self.possible_simple_keys[self.flow_level]
553 del self.possible_simple_keys[self.flow_level]
554 self.tokens.insert(key.token_number-self.tokens_taken,
555 KeyToken(key.mark, key.mark))
557 # If this key starts a new block mapping, we need to add
558 # BLOCK-MAPPING-START.
559 if not self.flow_level:
560 if self.add_indent(key.column):
561 self.tokens.insert(key.token_number-self.tokens_taken,
562 BlockMappingStartToken(key.mark, key.mark))
564 # There cannot be two simple keys one after another.
565 self.allow_simple_key = False
567 # It must be a part of a complex key.
570 # Block context needs additional checks.
571 # (Do we really need them? They will be catched by the parser
573 if not self.flow_level:
575 # We are allowed to start a complex value if and only if
576 # we can start a simple key.
577 if not self.allow_simple_key:
578 raise ScannerError(None, None,
579 "mapping values are not allowed here",
582 # If this value starts a new block mapping, we need to add
583 # BLOCK-MAPPING-START. It will be detected as an error later by
585 if not self.flow_level:
586 if self.add_indent(self.column):
587 mark = self.get_mark()
588 self.tokens.append(BlockMappingStartToken(mark, mark))
590 # Simple keys are allowed after ':' in the block context.
591 self.allow_simple_key = not self.flow_level
593 # Reset possible simple key on the current level.
594 self.remove_possible_simple_key()
597 start_mark = self.get_mark()
599 end_mark = self.get_mark()
600 self.tokens.append(ValueToken(start_mark, end_mark))
602 def fetch_alias(self):
604 # ALIAS could be a simple key.
605 self.save_possible_simple_key()
607 # No simple keys after ALIAS.
608 self.allow_simple_key = False
610 # Scan and add ALIAS.
611 self.tokens.append(self.scan_anchor(AliasToken))
613 def fetch_anchor(self):
615 # ANCHOR could start a simple key.
616 self.save_possible_simple_key()
618 # No simple keys after ANCHOR.
619 self.allow_simple_key = False
621 # Scan and add ANCHOR.
622 self.tokens.append(self.scan_anchor(AnchorToken))
626 # TAG could start a simple key.
627 self.save_possible_simple_key()
629 # No simple keys after TAG.
630 self.allow_simple_key = False
633 self.tokens.append(self.scan_tag())
635 def fetch_literal(self):
636 self.fetch_block_scalar(style='|')
638 def fetch_folded(self):
639 self.fetch_block_scalar(style='>')
641 def fetch_block_scalar(self, style):
643 # A simple key may follow a block scalar.
644 self.allow_simple_key = True
646 # Reset possible simple key on the current level.
647 self.remove_possible_simple_key()
649 # Scan and add SCALAR.
650 self.tokens.append(self.scan_block_scalar(style))
652 def fetch_single(self):
653 self.fetch_flow_scalar(style='\'')
655 def fetch_double(self):
656 self.fetch_flow_scalar(style='"')
658 def fetch_flow_scalar(self, style):
660 # A flow scalar could be a simple key.
661 self.save_possible_simple_key()
663 # No simple keys after flow scalars.
664 self.allow_simple_key = False
666 # Scan and add SCALAR.
667 self.tokens.append(self.scan_flow_scalar(style))
669 def fetch_plain(self):
671 # A plain scalar could be a simple key.
672 self.save_possible_simple_key()
674 # No simple keys after plain scalars. But note that `scan_plain` will
675 # change this flag if the scan is finished at the beginning of the
677 self.allow_simple_key = False
679 # Scan and add SCALAR. May change `allow_simple_key`.
680 self.tokens.append(self.scan_plain())
684 def check_directive(self):
686 # DIRECTIVE: ^ '%' ...
687 # The '%' indicator is already checked.
691 def check_document_start(self):
693 # DOCUMENT-START: ^ '---' (' '|'\n')
695 if self.prefix(3) == u'---' \
696 and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
699 def check_document_end(self):
701 # DOCUMENT-END: ^ '...' (' '|'\n')
703 if self.prefix(3) == u'...' \
704 and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
707 def check_block_entry(self):
709 # BLOCK-ENTRY: '-' (' '|'\n')
710 return self.peek(1) in u'\0 \t\r\n\x85\u2028\u2029'
714 # KEY(flow context): '?'
718 # KEY(block context): '?' (' '|'\n')
720 return self.peek(1) in u'\0 \t\r\n\x85\u2028\u2029'
722 def check_value(self):
724 # VALUE(flow context): ':'
728 # VALUE(block context): ':' (' '|'\n')
730 return self.peek(1) in u'\0 \t\r\n\x85\u2028\u2029'
732 def check_plain(self):
734 # A plain scalar may start with any non-space character except:
735 # '-', '?', ':', ',', '[', ']', '{', '}',
736 # '#', '&', '*', '!', '|', '>', '\'', '\"',
739 # It may also start with
741 # if it is followed by a non-space character.
743 # Note that we limit the last rule to the block context (except the
744 # '-' character) because we want the flow context to be space
747 return ch not in u'\0 \t\r\n\x85\u2028\u2029-?:,[]{}#&*!|>\'\"%@`' \
748 or (self.peek(1) not in u'\0 \t\r\n\x85\u2028\u2029'
749 and (ch == u'-' or (not self.flow_level and ch in u'?:')))
753 def scan_to_next_token(self):
754 # We ignore spaces, line breaks and comments.
755 # If we find a line break in the block context, we set the flag
756 # `allow_simple_key` on.
757 # The byte order mark is stripped if it's the first character in the
758 # stream. We do not yet support BOM inside the stream as the
759 # specification requires. Any such mark will be considered as a part
762 # TODO: We need to make tab handling rules more sane. A good rule is
763 # Tabs cannot precede tokens
764 # BLOCK-SEQUENCE-START, BLOCK-MAPPING-START, BLOCK-END,
765 # KEY(block), VALUE(block), BLOCK-ENTRY
766 # So the checking code is
768 # self.allow_simple_keys = False
769 # We also need to add the check for `allow_simple_keys == True` to
770 # `unwind_indent` before issuing BLOCK-END.
771 # Scanners for block, flow, and plain scalars need to be modified.
773 if self.index == 0 and self.peek() == u'\uFEFF':
777 while self.peek() == u' ':
779 if self.peek() == u'#':
780 while self.peek() not in u'\0\r\n\x85\u2028\u2029':
782 if self.scan_line_break():
783 if not self.flow_level:
784 self.allow_simple_key = True
788 def scan_directive(self):
789 # See the specification for details.
790 start_mark = self.get_mark()
792 name = self.scan_directive_name(start_mark)
795 value = self.scan_yaml_directive_value(start_mark)
796 end_mark = self.get_mark()
798 value = self.scan_tag_directive_value(start_mark)
799 end_mark = self.get_mark()
801 end_mark = self.get_mark()
802 while self.peek() not in u'\0\r\n\x85\u2028\u2029':
804 self.scan_directive_ignored_line(start_mark)
805 return DirectiveToken(name, value, start_mark, end_mark)
807 def scan_directive_name(self, start_mark):
808 # See the specification for details.
810 ch = self.peek(length)
811 while u'0' <= ch <= u'9' or u'A' <= ch <= u'Z' or u'a' <= ch <= u'z' \
814 ch = self.peek(length)
816 raise ScannerError("while scanning a directive", start_mark,
817 "expected alphabetic or numeric character, but found %r"
818 % ch.encode('utf-8'), self.get_mark())
819 value = self.prefix(length)
822 if ch not in u'\0 \r\n\x85\u2028\u2029':
823 raise ScannerError("while scanning a directive", start_mark,
824 "expected alphabetic or numeric character, but found %r"
825 % ch.encode('utf-8'), self.get_mark())
828 def scan_yaml_directive_value(self, start_mark):
829 # See the specification for details.
830 while self.peek() == u' ':
832 major = self.scan_yaml_directive_number(start_mark)
833 if self.peek() != '.':
834 raise ScannerError("while scanning a directive", start_mark,
835 "expected a digit or '.', but found %r"
836 % self.peek().encode('utf-8'),
839 minor = self.scan_yaml_directive_number(start_mark)
840 if self.peek() not in u'\0 \r\n\x85\u2028\u2029':
841 raise ScannerError("while scanning a directive", start_mark,
842 "expected a digit or ' ', but found %r"
843 % self.peek().encode('utf-8'),
845 return (major, minor)
847 def scan_yaml_directive_number(self, start_mark):
848 # See the specification for details.
850 if not (u'0' <= ch <= u'9'):
851 raise ScannerError("while scanning a directive", start_mark,
852 "expected a digit, but found %r" % ch.encode('utf-8'),
855 while u'0' <= self.peek(length) <= u'9':
857 value = int(self.prefix(length))
861 def scan_tag_directive_value(self, start_mark):
862 # See the specification for details.
863 while self.peek() == u' ':
865 handle = self.scan_tag_directive_handle(start_mark)
866 while self.peek() == u' ':
868 prefix = self.scan_tag_directive_prefix(start_mark)
869 return (handle, prefix)
871 def scan_tag_directive_handle(self, start_mark):
872 # See the specification for details.
873 value = self.scan_tag_handle('directive', start_mark)
876 raise ScannerError("while scanning a directive", start_mark,
877 "expected ' ', but found %r" % ch.encode('utf-8'),
881 def scan_tag_directive_prefix(self, start_mark):
882 # See the specification for details.
883 value = self.scan_tag_uri('directive', start_mark)
885 if ch not in u'\0 \r\n\x85\u2028\u2029':
886 raise ScannerError("while scanning a directive", start_mark,
887 "expected ' ', but found %r" % ch.encode('utf-8'),
891 def scan_directive_ignored_line(self, start_mark):
892 # See the specification for details.
893 while self.peek() == u' ':
895 if self.peek() == u'#':
896 while self.peek() not in u'\0\r\n\x85\u2028\u2029':
899 if ch not in u'\0\r\n\x85\u2028\u2029':
900 raise ScannerError("while scanning a directive", start_mark,
901 "expected a comment or a line break, but found %r"
902 % ch.encode('utf-8'), self.get_mark())
903 self.scan_line_break()
905 def scan_anchor(self, TokenClass):
906 # The specification does not restrict characters for anchors and
907 # aliases. This may lead to problems, for instance, the document:
909 # can be interpteted in two ways, as
912 # [ *alias , "value" ]
913 # Therefore we restrict aliases to numbers and ASCII letters.
914 start_mark = self.get_mark()
915 indicator = self.peek()
916 if indicator == u'*':
922 ch = self.peek(length)
923 while u'0' <= ch <= u'9' or u'A' <= ch <= u'Z' or u'a' <= ch <= u'z' \
926 ch = self.peek(length)
928 raise ScannerError("while scanning an %s" % name, start_mark,
929 "expected alphabetic or numeric character, but found %r"
930 % ch.encode('utf-8'), self.get_mark())
931 value = self.prefix(length)
934 if ch not in u'\0 \t\r\n\x85\u2028\u2029?:,]}%@`':
935 raise ScannerError("while scanning an %s" % name, start_mark,
936 "expected alphabetic or numeric character, but found %r"
937 % ch.encode('utf-8'), self.get_mark())
938 end_mark = self.get_mark()
939 return TokenClass(value, start_mark, end_mark)
942 # See the specification for details.
943 start_mark = self.get_mark()
948 suffix = self.scan_tag_uri('tag', start_mark)
949 if self.peek() != u'>':
950 raise ScannerError("while parsing a tag", start_mark,
951 "expected '>', but found %r" % self.peek().encode('utf-8'),
954 elif ch in u'\0 \t\r\n\x85\u2028\u2029':
961 while ch not in u'\0 \r\n\x85\u2028\u2029':
966 ch = self.peek(length)
969 handle = self.scan_tag_handle('tag', start_mark)
973 suffix = self.scan_tag_uri('tag', start_mark)
975 if ch not in u'\0 \r\n\x85\u2028\u2029':
976 raise ScannerError("while scanning a tag", start_mark,
977 "expected ' ', but found %r" % ch.encode('utf-8'),
979 value = (handle, suffix)
980 end_mark = self.get_mark()
981 return TagToken(value, start_mark, end_mark)
983 def scan_block_scalar(self, style):
984 # See the specification for details.
992 start_mark = self.get_mark()
996 chomping, increment = self.scan_block_scalar_indicators(start_mark)
997 self.scan_block_scalar_ignored_line(start_mark)
999 # Determine the indentation level and go to the first non-empty line.
1000 min_indent = self.indent+1
1003 if increment is None:
1004 breaks, max_indent, end_mark = self.scan_block_scalar_indentation()
1005 indent = max(min_indent, max_indent)
1007 indent = min_indent+increment-1
1008 breaks, end_mark = self.scan_block_scalar_breaks(indent)
1011 # Scan the inner part of the block scalar.
1012 while self.column == indent and self.peek() != u'\0':
1013 chunks.extend(breaks)
1014 leading_non_space = self.peek() not in u' \t'
1016 while self.peek(length) not in u'\0\r\n\x85\u2028\u2029':
1018 chunks.append(self.prefix(length))
1019 self.forward(length)
1020 line_break = self.scan_line_break()
1021 breaks, end_mark = self.scan_block_scalar_breaks(indent)
1022 if self.column == indent and self.peek() != u'\0':
1024 # Unfortunately, folding rules are ambiguous.
1026 # This is the folding according to the specification:
1028 if folded and line_break == u'\n' \
1029 and leading_non_space and self.peek() not in u' \t':
1033 chunks.append(line_break)
1035 # This is Clark Evans's interpretation (also in the spec
1038 #if folded and line_break == u'\n':
1040 # if self.peek() not in ' \t':
1041 # chunks.append(u' ')
1043 # chunks.append(line_break)
1045 # chunks.append(line_break)
1050 if chomping is not False:
1051 chunks.append(line_break)
1052 if chomping is True:
1053 chunks.extend(breaks)
1056 return ScalarToken(u''.join(chunks), False, start_mark, end_mark,
1059 def scan_block_scalar_indicators(self, start_mark):
1060 # See the specification for details.
1071 if ch in u'0123456789':
1074 raise ScannerError("while scanning a block scalar", start_mark,
1075 "expected indentation indicator in the range 1-9, but found 0",
1078 elif ch in u'0123456789':
1081 raise ScannerError("while scanning a block scalar", start_mark,
1082 "expected indentation indicator in the range 1-9, but found 0",
1093 if ch not in u'\0 \r\n\x85\u2028\u2029':
1094 raise ScannerError("while scanning a block scalar", start_mark,
1095 "expected chomping or indentation indicators, but found %r"
1096 % ch.encode('utf-8'), self.get_mark())
1097 return chomping, increment
1099 def scan_block_scalar_ignored_line(self, start_mark):
1100 # See the specification for details.
1101 while self.peek() == u' ':
1103 if self.peek() == u'#':
1104 while self.peek() not in u'\0\r\n\x85\u2028\u2029':
1107 if ch not in u'\0\r\n\x85\u2028\u2029':
1108 raise ScannerError("while scanning a block scalar", start_mark,
1109 "expected a comment or a line break, but found %r"
1110 % ch.encode('utf-8'), self.get_mark())
1111 self.scan_line_break()
1113 def scan_block_scalar_indentation(self):
1114 # See the specification for details.
1117 end_mark = self.get_mark()
1118 while self.peek() in u' \r\n\x85\u2028\u2029':
1119 if self.peek() != u' ':
1120 chunks.append(self.scan_line_break())
1121 end_mark = self.get_mark()
1124 if self.column > max_indent:
1125 max_indent = self.column
1126 return chunks, max_indent, end_mark
1128 def scan_block_scalar_breaks(self, indent):
1129 # See the specification for details.
1131 end_mark = self.get_mark()
1132 while self.column < indent and self.peek() == u' ':
1134 while self.peek() in u'\r\n\x85\u2028\u2029':
1135 chunks.append(self.scan_line_break())
1136 end_mark = self.get_mark()
1137 while self.column < indent and self.peek() == u' ':
1139 return chunks, end_mark
1141 def scan_flow_scalar(self, style):
1142 # See the specification for details.
1143 # Note that we loose indentation rules for quoted scalars. Quoted
1144 # scalars don't need to adhere indentation because " and ' clearly
1145 # mark the beginning and the end of them. Therefore we are less
1146 # restrictive then the specification requires. We only need to check
1147 # that document separators are not included in scalars.
1153 start_mark = self.get_mark()
1156 chunks.extend(self.scan_flow_scalar_non_spaces(double, start_mark))
1157 while self.peek() != quote:
1158 chunks.extend(self.scan_flow_scalar_spaces(double, start_mark))
1159 chunks.extend(self.scan_flow_scalar_non_spaces(double, start_mark))
1161 end_mark = self.get_mark()
1162 return ScalarToken(u''.join(chunks), False, start_mark, end_mark,
1165 ESCAPE_REPLACEMENTS = {
1191 def scan_flow_scalar_non_spaces(self, double, start_mark):
1192 # See the specification for details.
1196 while self.peek(length) not in u'\'\"\\\0 \t\r\n\x85\u2028\u2029':
1199 chunks.append(self.prefix(length))
1200 self.forward(length)
1202 if not double and ch == u'\'' and self.peek(1) == u'\'':
1203 chunks.append(u'\'')
1205 elif (double and ch == u'\'') or (not double and ch in u'\"\\'):
1208 elif double and ch == u'\\':
1211 if ch in self.ESCAPE_REPLACEMENTS:
1212 chunks.append(self.ESCAPE_REPLACEMENTS[ch])
1214 elif ch in self.ESCAPE_CODES:
1215 length = self.ESCAPE_CODES[ch]
1217 for k in range(length):
1218 if self.peek(k) not in u'0123456789ABCDEFabcdef':
1219 raise ScannerError("while scanning a double-quoted scalar", start_mark,
1220 "expected escape sequence of %d hexdecimal numbers, but found %r" %
1221 (length, self.peek(k).encode('utf-8')), self.get_mark())
1222 code = int(self.prefix(length), 16)
1223 chunks.append(unichr(code))
1224 self.forward(length)
1225 elif ch in u'\r\n\x85\u2028\u2029':
1226 self.scan_line_break()
1227 chunks.extend(self.scan_flow_scalar_breaks(double, start_mark))
1229 raise ScannerError("while scanning a double-quoted scalar", start_mark,
1230 "found unknown escape character %r" % ch.encode('utf-8'), self.get_mark())
1234 def scan_flow_scalar_spaces(self, double, start_mark):
1235 # See the specification for details.
1238 while self.peek(length) in u' \t':
1240 whitespaces = self.prefix(length)
1241 self.forward(length)
1244 raise ScannerError("while scanning a quoted scalar", start_mark,
1245 "found unexpected end of stream", self.get_mark())
1246 elif ch in u'\r\n\x85\u2028\u2029':
1247 line_break = self.scan_line_break()
1248 breaks = self.scan_flow_scalar_breaks(double, start_mark)
1249 if line_break != u'\n':
1250 chunks.append(line_break)
1253 chunks.extend(breaks)
1255 chunks.append(whitespaces)
1258 def scan_flow_scalar_breaks(self, double, start_mark):
1259 # See the specification for details.
1262 # Instead of checking indentation, we check for document
1264 prefix = self.prefix(3)
1265 if (prefix == u'---' or prefix == u'...') \
1266 and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
1267 raise ScannerError("while scanning a quoted scalar", start_mark,
1268 "found unexpected document separator", self.get_mark())
1269 while self.peek() in u' \t':
1271 if self.peek() in u'\r\n\x85\u2028\u2029':
1272 chunks.append(self.scan_line_break())
1276 def scan_plain(self):
1277 # See the specification for details.
1278 # We add an additional restriction for the flow context:
1279 # plain scalars in the flow context cannot contain ',', ':' and '?'.
1280 # We also keep track of the `allow_simple_key` flag here.
1281 # Indentation rules are loosed for the flow context.
1283 start_mark = self.get_mark()
1284 end_mark = start_mark
1285 indent = self.indent+1
1286 # We allow zero indentation for scalars, but then we need to check for
1287 # document separators at the beginning of the line.
1293 if self.peek() == u'#':
1296 ch = self.peek(length)
1297 if ch in u'\0 \t\r\n\x85\u2028\u2029' \
1298 or (not self.flow_level and ch == u':' and
1299 self.peek(length+1) in u'\0 \t\r\n\x85\u2028\u2029') \
1300 or (self.flow_level and ch in u',:?[]{}'):
1303 # It's not clear what we should do with ':' in the flow context.
1304 if (self.flow_level and ch == u':'
1305 and self.peek(length+1) not in u'\0 \t\r\n\x85\u2028\u2029,[]{}'):
1306 self.forward(length)
1307 raise ScannerError("while scanning a plain scalar", start_mark,
1308 "found unexpected ':'", self.get_mark(),
1309 "Please check http://pyyaml.org/wiki/YAMLColonInFlowContext for details.")
1312 self.allow_simple_key = False
1313 chunks.extend(spaces)
1314 chunks.append(self.prefix(length))
1315 self.forward(length)
1316 end_mark = self.get_mark()
1317 spaces = self.scan_plain_spaces(indent, start_mark)
1318 if not spaces or self.peek() == u'#' \
1319 or (not self.flow_level and self.column < indent):
1321 return ScalarToken(u''.join(chunks), True, start_mark, end_mark)
1323 def scan_plain_spaces(self, indent, start_mark):
1324 # See the specification for details.
1325 # The specification is really confusing about tabs in plain scalars.
1326 # We just forbid them completely. Do not use tabs in YAML!
1329 while self.peek(length) in u' ':
1331 whitespaces = self.prefix(length)
1332 self.forward(length)
1334 if ch in u'\r\n\x85\u2028\u2029':
1335 line_break = self.scan_line_break()
1336 self.allow_simple_key = True
1337 prefix = self.prefix(3)
1338 if (prefix == u'---' or prefix == u'...') \
1339 and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
1342 while self.peek() in u' \r\n\x85\u2028\u2029':
1343 if self.peek() == ' ':
1346 breaks.append(self.scan_line_break())
1347 prefix = self.prefix(3)
1348 if (prefix == u'---' or prefix == u'...') \
1349 and self.peek(3) in u'\0 \t\r\n\x85\u2028\u2029':
1351 if line_break != u'\n':
1352 chunks.append(line_break)
1355 chunks.extend(breaks)
1357 chunks.append(whitespaces)
1360 def scan_tag_handle(self, name, start_mark):
1361 # See the specification for details.
1362 # For some strange reasons, the specification does not allow '_' in
1363 # tag handles. I have allowed it anyway.
1366 raise ScannerError("while scanning a %s" % name, start_mark,
1367 "expected '!', but found %r" % ch.encode('utf-8'),
1370 ch = self.peek(length)
1372 while u'0' <= ch <= u'9' or u'A' <= ch <= u'Z' or u'a' <= ch <= u'z' \
1375 ch = self.peek(length)
1377 self.forward(length)
1378 raise ScannerError("while scanning a %s" % name, start_mark,
1379 "expected '!', but found %r" % ch.encode('utf-8'),
1382 value = self.prefix(length)
1383 self.forward(length)
1386 def scan_tag_uri(self, name, start_mark):
1387 # See the specification for details.
1388 # Note: we do not check if URI is well-formed.
1391 ch = self.peek(length)
1392 while u'0' <= ch <= u'9' or u'A' <= ch <= u'Z' or u'a' <= ch <= u'z' \
1393 or ch in u'-;/?:@&=+$,_.!~*\'()[]%':
1395 chunks.append(self.prefix(length))
1396 self.forward(length)
1398 chunks.append(self.scan_uri_escapes(name, start_mark))
1401 ch = self.peek(length)
1403 chunks.append(self.prefix(length))
1404 self.forward(length)
1407 raise ScannerError("while parsing a %s" % name, start_mark,
1408 "expected URI, but found %r" % ch.encode('utf-8'),
1410 return u''.join(chunks)
1412 def scan_uri_escapes(self, name, start_mark):
1413 # See the specification for details.
1415 mark = self.get_mark()
1416 while self.peek() == u'%':
1419 if self.peek(k) not in u'0123456789ABCDEFabcdef':
1420 raise ScannerError("while scanning a %s" % name, start_mark,
1421 "expected URI escape sequence of 2 hexdecimal numbers, but found %r" %
1422 (self.peek(k).encode('utf-8')), self.get_mark())
1423 bytes.append(chr(int(self.prefix(2), 16)))
1426 value = unicode(''.join(bytes), 'utf-8')
1427 except UnicodeDecodeError, exc:
1428 raise ScannerError("while scanning a %s" % name, start_mark, str(exc), mark)
1431 def scan_line_break(self):
1437 # '\u2028' : '\u2028'
1438 # '\u2029 : '\u2029'
1441 if ch in u'\r\n\x85':
1442 if self.prefix(2) == u'\r\n':
1447 elif ch in u'\u2028\u2029':
1454 # psyco.bind(Scanner)
1455 #except ImportError: