3 from .exceptions import ParseException
9 Wrapper for parse actions, to ensure they are only called once.
12 def __init__(self, method_call):
13 from .core import _trim_arity
15 self.callable = _trim_arity(method_call)
18 def __call__(self, s, l, t):
20 results = self.callable(s, l, t)
23 raise ParseException(s, l, "OnlyOnce obj called multiple times w/out reset")
27 Allow the associated parse action to be called once more.
33 def match_only_at_col(n):
35 Helper method for defining parse actions that require matching at
36 a specific column in the input text.
39 def verify_col(strg, locn, toks):
40 if col(locn, strg) != n:
41 raise ParseException(strg, locn, "matched token not at column {}".format(n))
46 def replace_with(repl_str):
48 Helper method for common parse actions that simply return
49 a literal value. Especially useful when used with
50 :class:`transform_string<ParserElement.transform_string>` ().
54 num = Word(nums).set_parse_action(lambda toks: int(toks[0]))
55 na = one_of("N/A NA").set_parse_action(replace_with(math.nan))
58 term[1, ...].parse_string("324 234 N/A 234") # -> [324, 234, nan, 234]
60 return lambda s, l, t: [repl_str]
63 def remove_quotes(s, l, t):
65 Helper parse action for removing quotation marks from parsed
70 # by default, quotation marks are included in parsed results
71 quoted_string.parse_string("'Now is the Winter of our Discontent'") # -> ["'Now is the Winter of our Discontent'"]
73 # use remove_quotes to strip quotation marks from parsed results
74 quoted_string.set_parse_action(remove_quotes)
75 quoted_string.parse_string("'Now is the Winter of our Discontent'") # -> ["Now is the Winter of our Discontent"]
80 def with_attribute(*args, **attr_dict):
82 Helper to create a validating parse action to be used with start
83 tags created with :class:`make_xml_tags` or
84 :class:`make_html_tags`. Use ``with_attribute`` to qualify
85 a starting tag with a required attribute value, to avoid false
86 matches on common tags such as ``<TD>`` or ``<DIV>``.
88 Call ``with_attribute`` with a series of attribute names and
89 values. Specify the list of filter attributes names and values as:
91 - keyword arguments, as in ``(align="right")``, or
92 - as an explicit dict with ``**`` operator, when an attribute
93 name is also a Python reserved word, as in ``**{"class":"Customer", "align":"right"}``
94 - a list of name-value tuples, as in ``(("ns1:class", "Customer"), ("ns2:align", "right"))``
96 For attribute names with a namespace prefix, you must use the second
97 form. Attribute names are matched insensitive to upper/lower case.
99 If just testing for ``class`` (with or without a namespace), use
102 To verify that the attribute exists, but without specifying a value,
103 pass ``with_attribute.ANY_VALUE`` as the value.
110 <div type="grid">1 4 0 1 0</div>
111 <div type="graph">1,3 2,3 1,1</div>
112 <div>this has no type</div>
116 div,div_end = make_html_tags("div")
118 # only match div tag having a type attribute with value "grid"
119 div_grid = div().set_parse_action(with_attribute(type="grid"))
120 grid_expr = div_grid + SkipTo(div | div_end)("body")
121 for grid_header in grid_expr.search_string(html):
122 print(grid_header.body)
124 # construct a match with any div tag having a type attribute, regardless of the value
125 div_any_type = div().set_parse_action(with_attribute(type=with_attribute.ANY_VALUE))
126 div_expr = div_any_type + SkipTo(div | div_end)("body")
127 for div_header in div_expr.search_string(html):
128 print(div_header.body)
140 attrs = attr_dict.items()
141 attrs = [(k, v) for k, v in attrs]
143 def pa(s, l, tokens):
144 for attrName, attrValue in attrs:
145 if attrName not in tokens:
146 raise ParseException(s, l, "no matching attribute " + attrName)
147 if attrValue != with_attribute.ANY_VALUE and tokens[attrName] != attrValue:
148 raise ParseException(
151 "attribute {!r} has value {!r}, must be {!r}".format(
152 attrName, tokens[attrName], attrValue
159 with_attribute.ANY_VALUE = object()
162 def with_class(classname, namespace=""):
164 Simplified version of :class:`with_attribute` when
165 matching on a div class - made difficult because ``class`` is
166 a reserved word in Python.
173 <div class="grid">1 4 0 1 0</div>
174 <div class="graph">1,3 2,3 1,1</div>
175 <div>this <div> has no class</div>
179 div,div_end = make_html_tags("div")
180 div_grid = div().set_parse_action(with_class("grid"))
182 grid_expr = div_grid + SkipTo(div | div_end)("body")
183 for grid_header in grid_expr.search_string(html):
184 print(grid_header.body)
186 div_any_type = div().set_parse_action(with_class(withAttribute.ANY_VALUE))
187 div_expr = div_any_type + SkipTo(div | div_end)("body")
188 for div_header in div_expr.search_string(html):
189 print(div_header.body)
198 classattr = "{}:class".format(namespace) if namespace else "class"
199 return with_attribute(**{classattr: classname})
202 # pre-PEP8 compatibility symbols
203 replaceWith = replace_with
204 removeQuotes = remove_quotes
205 withAttribute = with_attribute
206 withClass = with_class
207 matchOnlyAtCol = match_only_at_col