ben
/
mwparserfromhell
镜像自地址 https://github.com/earwig/mwparserfromhell
1
1
派生 0
代码 版本发布 24 动态
A Python parser for MediaWiki wikicode https://mwparserfromhell.readthedocs.io/
您最多选择25个主题 主题必须以字母或数字开头，可以包含连字符 (-)，并且长度不得超过35个字符
408 提交
1 分支
5.1 MiB
 
 
 
 
目录树: b8752a5c15
分支列表 标签列表
${ item.name }
创建分支 ${ searchTerm }
从 'b8752a5c15'
${ noResults }
mwparserfromhell/tests/test_builder.py

365 行
18 KiB
原始文件 Blame 文件历史 

  1. # -*- coding: utf-8  -*-

  2. #

  3. # Copyright (C) 2012-2013 Ben Kurtovic <ben.kurtovic@verizon.net>

  4. #

  5. # Permission is hereby granted, free of charge, to any person obtaining a copy

  6. # of this software and associated documentation files (the "Software"), to deal

  7. # in the Software without restriction, including without limitation the rights

  8. # to use, copy, modify, merge, publish, distribute, sublicense, and/or sell

  9. # copies of the Software, and to permit persons to whom the Software is

  10. # furnished to do so, subject to the following conditions:

  11. #

  12. # The above copyright notice and this permission notice shall be included in

  13. # all copies or substantial portions of the Software.

  14. #

  15. # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR

  16. # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

  17. # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE

  18. # AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER

  19. # LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,

  20. # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE

  21. # SOFTWARE.

  22. 

  23. from __future__ import unicode_literals

  24. import unittest

  25. 

  26. from mwparserfromhell.nodes import (Argument, Comment, Heading, HTMLEntity,

  27.                                     Tag, Template, Text, Wikilink)

  28. from mwparserfromhell.nodes.extras import Attribute, Parameter

  29. from mwparserfromhell.parser import tokens

  30. from mwparserfromhell.parser.builder import Builder

  31. 

  32. from ._test_tree_equality import TreeEqualityTestCase, wrap, wraptext

  33. 

  34. class TestBuilder(TreeEqualityTestCase):

  35.     """Tests for the builder, which turns tokens into Wikicode objects."""

  36. 

  37.     def setUp(self):

  38.         self.builder = Builder()

  39. 

  40.     def test_text(self):

  41.         """tests for building Text nodes"""

  42.         tests = [

  43.             ([tokens.Text(text="foobar")], wraptext("foobar")),

  44.             ([tokens.Text(text="fóóbar")], wraptext("fóóbar")),

  45.             ([tokens.Text(text="spam"), tokens.Text(text="eggs")],

  46.              wraptext("spam", "eggs")),

  47.         ]

  48.         for test, valid in tests:

  49.             self.assertWikicodeEqual(valid, self.builder.build(test))

  50. 

  51.     def test_template(self):

  52.         """tests for building Template nodes"""

  53.         tests = [

  54.             ([tokens.TemplateOpen(), tokens.Text(text="foobar"),

  55.               tokens.TemplateClose()],

  56.              wrap([Template(wraptext("foobar"))])),

  57. 

  58.             ([tokens.TemplateOpen(), tokens.Text(text="spam"),

  59.               tokens.Text(text="eggs"), tokens.TemplateClose()],

  60.              wrap([Template(wraptext("spam", "eggs"))])),

  61. 

  62.             ([tokens.TemplateOpen(), tokens.Text(text="foo"),

  63.               tokens.TemplateParamSeparator(), tokens.Text(text="bar"),

  64.               tokens.TemplateClose()],

  65.              wrap([Template(wraptext("foo"), params=[

  66.                  Parameter(wraptext("1"), wraptext("bar"), showkey=False)])])),

  67. 

  68.             ([tokens.TemplateOpen(), tokens.Text(text="foo"),

  69.               tokens.TemplateParamSeparator(), tokens.Text(text="bar"),

  70.               tokens.TemplateParamEquals(), tokens.Text(text="baz"),

  71.               tokens.TemplateClose()],

  72.              wrap([Template(wraptext("foo"), params=[

  73.                  Parameter(wraptext("bar"), wraptext("baz"))])])),

  74. 

  75.             ([tokens.TemplateOpen(), tokens.TemplateParamSeparator(),

  76.               tokens.TemplateParamSeparator(), tokens.TemplateParamEquals(),

  77.               tokens.TemplateParamSeparator(), tokens.TemplateClose()],

  78.              wrap([Template(wrap([]), params=[

  79.                  Parameter(wraptext("1"), wrap([]), showkey=False),

  80.                  Parameter(wrap([]), wrap([]), showkey=True),

  81.                  Parameter(wraptext("2"), wrap([]), showkey=False)])])),

  82. 

  83.             ([tokens.TemplateOpen(), tokens.Text(text="foo"),

  84.               tokens.TemplateParamSeparator(), tokens.Text(text="bar"),

  85.               tokens.TemplateParamEquals(), tokens.Text(text="baz"),

  86.               tokens.TemplateParamSeparator(), tokens.Text(text="biz"),

  87.               tokens.TemplateParamSeparator(), tokens.Text(text="buzz"),

  88.               tokens.TemplateParamSeparator(), tokens.Text(text="3"),

  89.               tokens.TemplateParamEquals(), tokens.Text(text="buff"),

  90.               tokens.TemplateParamSeparator(), tokens.Text(text="baff"),

  91.               tokens.TemplateClose()],

  92.              wrap([Template(wraptext("foo"), params=[

  93.                  Parameter(wraptext("bar"), wraptext("baz")),

  94.                  Parameter(wraptext("1"), wraptext("biz"), showkey=False),

  95.                  Parameter(wraptext("2"), wraptext("buzz"), showkey=False),

  96.                  Parameter(wraptext("3"), wraptext("buff")),

  97.                  Parameter(wraptext("3"), wraptext("baff"),

  98.                            showkey=False)])])),

  99.         ]

  100.         for test, valid in tests:

  101.             self.assertWikicodeEqual(valid, self.builder.build(test))

  102. 

  103.     def test_argument(self):

  104.         """tests for building Argument nodes"""

  105.         tests = [

  106.             ([tokens.ArgumentOpen(), tokens.Text(text="foobar"),

  107.               tokens.ArgumentClose()],

  108.              wrap([Argument(wraptext("foobar"))])),

  109. 

  110.             ([tokens.ArgumentOpen(), tokens.Text(text="spam"),

  111.               tokens.Text(text="eggs"), tokens.ArgumentClose()],

  112.              wrap([Argument(wraptext("spam", "eggs"))])),

  113. 

  114.             ([tokens.ArgumentOpen(), tokens.Text(text="foo"),

  115.               tokens.ArgumentSeparator(), tokens.Text(text="bar"),

  116.               tokens.ArgumentClose()],

  117.              wrap([Argument(wraptext("foo"), wraptext("bar"))])),

  118. 

  119.             ([tokens.ArgumentOpen(), tokens.Text(text="foo"),

  120.               tokens.Text(text="bar"), tokens.ArgumentSeparator(),

  121.               tokens.Text(text="baz"), tokens.Text(text="biz"),

  122.               tokens.ArgumentClose()],

  123.              wrap([Argument(wraptext("foo", "bar"), wraptext("baz", "biz"))])),

  124.         ]

  125.         for test, valid in tests:

  126.             self.assertWikicodeEqual(valid, self.builder.build(test))

  127. 

  128.     def test_wikilink(self):

  129.         """tests for building Wikilink nodes"""

  130.         tests = [

  131.             ([tokens.WikilinkOpen(), tokens.Text(text="foobar"),

  132.               tokens.WikilinkClose()],

  133.              wrap([Wikilink(wraptext("foobar"))])),

  134. 

  135.             ([tokens.WikilinkOpen(), tokens.Text(text="spam"),

  136.               tokens.Text(text="eggs"), tokens.WikilinkClose()],

  137.              wrap([Wikilink(wraptext("spam", "eggs"))])),

  138. 

  139.             ([tokens.WikilinkOpen(), tokens.Text(text="foo"),

  140.               tokens.WikilinkSeparator(), tokens.Text(text="bar"),

  141.               tokens.WikilinkClose()],

  142.              wrap([Wikilink(wraptext("foo"), wraptext("bar"))])),

  143. 

  144.             ([tokens.WikilinkOpen(), tokens.Text(text="foo"),

  145.               tokens.Text(text="bar"), tokens.WikilinkSeparator(),

  146.               tokens.Text(text="baz"), tokens.Text(text="biz"),

  147.               tokens.WikilinkClose()],

  148.              wrap([Wikilink(wraptext("foo", "bar"), wraptext("baz", "biz"))])),

  149.         ]

  150.         for test, valid in tests:

  151.             self.assertWikicodeEqual(valid, self.builder.build(test))

  152. 

  153.     def test_html_entity(self):

  154.         """tests for building HTMLEntity nodes"""

  155.         tests = [

  156.             ([tokens.HTMLEntityStart(), tokens.Text(text="nbsp"),

  157.               tokens.HTMLEntityEnd()],

  158.              wrap([HTMLEntity("nbsp", named=True, hexadecimal=False)])),

  159. 

  160.             ([tokens.HTMLEntityStart(), tokens.HTMLEntityNumeric(),

  161.               tokens.Text(text="107"), tokens.HTMLEntityEnd()],

  162.              wrap([HTMLEntity("107", named=False, hexadecimal=False)])),

  163. 

  164.             ([tokens.HTMLEntityStart(), tokens.HTMLEntityNumeric(),

  165.               tokens.HTMLEntityHex(char="X"), tokens.Text(text="6B"),

  166.               tokens.HTMLEntityEnd()],

  167.              wrap([HTMLEntity("6B", named=False, hexadecimal=True,

  168.                               hex_char="X")])),

  169.         ]

  170.         for test, valid in tests:

  171.             self.assertWikicodeEqual(valid, self.builder.build(test))

  172. 

  173.     def test_heading(self):

  174.         """tests for building Heading nodes"""

  175.         tests = [

  176.             ([tokens.HeadingStart(level=2), tokens.Text(text="foobar"),

  177.               tokens.HeadingEnd()],

  178.              wrap([Heading(wraptext("foobar"), 2)])),

  179. 

  180.             ([tokens.HeadingStart(level=4), tokens.Text(text="spam"),

  181.               tokens.Text(text="eggs"), tokens.HeadingEnd()],

  182.              wrap([Heading(wraptext("spam", "eggs"), 4)])),

  183.         ]

  184.         for test, valid in tests:

  185.             self.assertWikicodeEqual(valid, self.builder.build(test))

  186. 

  187.     def test_comment(self):

  188.         """tests for building Comment nodes"""

  189.         tests = [

  190.             ([tokens.CommentStart(), tokens.Text(text="foobar"),

  191.               tokens.CommentEnd()],

  192.              wrap([Comment(wraptext("foobar"))])),

  193. 

  194.             ([tokens.CommentStart(), tokens.Text(text="spam"),

  195.               tokens.Text(text="eggs"), tokens.CommentEnd()],

  196.              wrap([Comment(wraptext("spam", "eggs"))])),

  197.         ]

  198.         for test, valid in tests:

  199.             self.assertWikicodeEqual(valid, self.builder.build(test))

  200. 

  201.     def test_tag(self):

  202.         """tests for building Tag nodes"""

  203.         tests = [

  204.             # <ref></ref>

  205.             ([tokens.TagOpenOpen(), tokens.Text(text="ref"),

  206.               tokens.TagCloseOpen(padding=""), tokens.TagOpenClose(),

  207.               tokens.Text(text="ref"), tokens.TagCloseClose()],

  208.              wrap([Tag(wraptext("ref"), wrap([]),

  209.                        closing_tag=wraptext("ref"))])),

  210. 

  211.             # <ref name></ref>

  212.             ([tokens.TagOpenOpen(), tokens.Text(text="ref"),

  213.               tokens.TagAttrStart(pad_first=" ", pad_before_eq="",

  214.                                   pad_after_eq=""),

  215.               tokens.Text(text="name"), tokens.TagCloseOpen(padding=""),

  216.               tokens.TagOpenClose(), tokens.Text(text="ref"),

  217.               tokens.TagCloseClose()],

  218.              wrap([Tag(wraptext("ref"), wrap([]),

  219.                       attrs=[Attribute(wraptext("name"))])])),

  220. 

  221.             # <ref name="abc" />

  222.             ([tokens.TagOpenOpen(), tokens.Text(text="ref"),

  223.               tokens.TagAttrStart(pad_first=" ", pad_before_eq="",

  224.                                   pad_after_eq=""),

  225.               tokens.Text(text="name"), tokens.TagAttrEquals(),

  226.               tokens.TagAttrQuote(), tokens.Text(text="abc"),

  227.               tokens.TagCloseSelfclose(padding=" ")],

  228.              wrap([Tag(wraptext("ref"),

  229.                        attrs=[Attribute(wraptext("name"), wraptext("abc"))],

  230.                        self_closing=True, padding=" ")])),

  231. 

  232.             # <br/>

  233.             ([tokens.TagOpenOpen(), tokens.Text(text="br"),

  234.               tokens.TagCloseSelfclose(padding="")],

  235.              wrap([Tag(wraptext("br"), self_closing=True)])),

  236. 

  237.             # <li>

  238.             ([tokens.TagOpenOpen(), tokens.Text(text="li"),

  239.               tokens.TagCloseSelfclose(padding="", implicit=True)],

  240.              wrap([Tag(wraptext("li"), self_closing=True, implicit=True)])),

  241. 

  242.             # </br>

  243.             ([tokens.TagOpenOpen(invalid=True), tokens.Text(text="br"),

  244.               tokens.TagCloseSelfclose(padding="", implicit=True)],

  245.              wrap([Tag(wraptext("br"), self_closing=True, invalid=True,

  246.                        implicit=True)])),

  247. 

  248.             # </br/>

  249.             ([tokens.TagOpenOpen(invalid=True), tokens.Text(text="br"),

  250.               tokens.TagCloseSelfclose(padding="")],

  251.              wrap([Tag(wraptext("br"), self_closing=True, invalid=True)])),

  252. 

  253.             # <ref name={{abc}}   foo="bar {{baz}}" abc={{de}}f ghi=j{{k}}{{l}}

  254.             #      mno =  "{{p}} [[q]] {{r}}">[[Source]]</ref>

  255.             ([tokens.TagOpenOpen(), tokens.Text(text="ref"),

  256.               tokens.TagAttrStart(pad_first=" ", pad_before_eq="",

  257.                                   pad_after_eq=""),

  258.               tokens.Text(text="name"), tokens.TagAttrEquals(),

  259.               tokens.TemplateOpen(), tokens.Text(text="abc"),

  260.               tokens.TemplateClose(),

  261.               tokens.TagAttrStart(pad_first="   ", pad_before_eq="",

  262.                                   pad_after_eq=""),

  263.               tokens.Text(text="foo"), tokens.TagAttrEquals(),

  264.               tokens.TagAttrQuote(), tokens.Text(text="bar "),

  265.               tokens.TemplateOpen(), tokens.Text(text="baz"),

  266.               tokens.TemplateClose(),

  267.               tokens.TagAttrStart(pad_first=" ", pad_before_eq="",

  268.                                   pad_after_eq=""),

  269.               tokens.Text(text="abc"), tokens.TagAttrEquals(),

  270.               tokens.TemplateOpen(), tokens.Text(text="de"),

  271.               tokens.TemplateClose(), tokens.Text(text="f"),

  272.               tokens.TagAttrStart(pad_first=" ", pad_before_eq="",

  273.                                   pad_after_eq=""),

  274.               tokens.Text(text="ghi"), tokens.TagAttrEquals(),

  275.               tokens.Text(text="j"), tokens.TemplateOpen(),

  276.               tokens.Text(text="k"), tokens.TemplateClose(),

  277.               tokens.TemplateOpen(), tokens.Text(text="l"),

  278.               tokens.TemplateClose(),

  279.               tokens.TagAttrStart(pad_first=" \n ", pad_before_eq=" ",

  280.                                   pad_after_eq="  "),

  281.               tokens.Text(text="mno"), tokens.TagAttrEquals(),

  282.               tokens.TagAttrQuote(), tokens.TemplateOpen(),

  283.               tokens.Text(text="p"), tokens.TemplateClose(),

  284.               tokens.Text(text=" "), tokens.WikilinkOpen(),

  285.               tokens.Text(text="q"), tokens.WikilinkClose(),

  286.               tokens.Text(text=" "), tokens.TemplateOpen(),

  287.               tokens.Text(text="r"), tokens.TemplateClose(),

  288.               tokens.TagCloseOpen(padding=""), tokens.WikilinkOpen(),

  289.               tokens.Text(text="Source"), tokens.WikilinkClose(),

  290.               tokens.TagOpenClose(), tokens.Text(text="ref"),

  291.               tokens.TagCloseClose()],

  292.              wrap([Tag(wraptext("ref"), wrap([Wikilink(wraptext("Source"))]), [

  293.                     Attribute(wraptext("name"),

  294.                               wrap([Template(wraptext("abc"))]), False),

  295.                     Attribute(wraptext("foo"), wrap([Text("bar "),

  296.                               Template(wraptext("baz"))]), pad_first="   "),

  297.                     Attribute(wraptext("abc"), wrap([Template(wraptext("de")),

  298.                               Text("f")]), False),

  299.                     Attribute(wraptext("ghi"), wrap([Text("j"),

  300.                               Template(wraptext("k")),

  301.                               Template(wraptext("l"))]), False),

  302.                     Attribute(wraptext("mno"), wrap([Template(wraptext("p")),

  303.                               Text(" "), Wikilink(wraptext("q")), Text(" "),

  304.                               Template(wraptext("r"))]), True, " \n ", " ",

  305.                               "  ")])])),

  306.         ]

  307.         for test, valid in tests:

  308.             self.assertWikicodeEqual(valid, self.builder.build(test))

  309. 

  310.     def test_integration(self):

  311.         """a test for building a combination of templates together"""

  312.         # {{{{{{{{foo}}bar|baz=biz}}buzz}}usr|{{bin}}}}

  313.         test = [tokens.TemplateOpen(), tokens.TemplateOpen(),

  314.                 tokens.TemplateOpen(), tokens.TemplateOpen(),

  315.                 tokens.Text(text="foo"), tokens.TemplateClose(),

  316.                 tokens.Text(text="bar"), tokens.TemplateParamSeparator(),

  317.                 tokens.Text(text="baz"), tokens.TemplateParamEquals(),

  318.                 tokens.Text(text="biz"), tokens.TemplateClose(),

  319.                 tokens.Text(text="buzz"), tokens.TemplateClose(),

  320.                 tokens.Text(text="usr"), tokens.TemplateParamSeparator(),

  321.                 tokens.TemplateOpen(), tokens.Text(text="bin"),

  322.                 tokens.TemplateClose(), tokens.TemplateClose()]

  323.         valid = wrap(

  324.             [Template(wrap([Template(wrap([Template(wrap([Template(wraptext(

  325.             "foo")), Text("bar")]), params=[Parameter(wraptext("baz"),

  326.             wraptext("biz"))]), Text("buzz")])), Text("usr")]), params=[

  327.             Parameter(wraptext("1"), wrap([Template(wraptext("bin"))]),

  328.             showkey=False)])])

  329.         self.assertWikicodeEqual(valid, self.builder.build(test))

  330. 

  331.     def test_integration2(self):

  332.         """an even more audacious test for building a horrible wikicode mess"""

  333.         # {{a|b|{{c|[[d]]{{{e}}}}}}}[[f|{{{g}}}<!--h-->]]{{i|j=&nbsp;}}

  334.         test = [tokens.TemplateOpen(), tokens.Text(text="a"),

  335.                 tokens.TemplateParamSeparator(), tokens.Text(text="b"),

  336.                 tokens.TemplateParamSeparator(), tokens.TemplateOpen(),

  337.                 tokens.Text(text="c"), tokens.TemplateParamSeparator(),

  338.                 tokens.WikilinkOpen(), tokens.Text(text="d"),

  339.                 tokens.WikilinkClose(), tokens.ArgumentOpen(),

  340.                 tokens.Text(text="e"), tokens.ArgumentClose(),

  341.                 tokens.TemplateClose(), tokens.TemplateClose(),

  342.                 tokens.WikilinkOpen(), tokens.Text(text="f"),

  343.                 tokens.WikilinkSeparator(), tokens.ArgumentOpen(),

  344.                 tokens.Text(text="g"), tokens.ArgumentClose(),

  345.                 tokens.CommentStart(), tokens.Text(text="h"),

  346.                 tokens.CommentEnd(), tokens.WikilinkClose(),

  347.                 tokens.TemplateOpen(), tokens.Text(text="i"),

  348.                 tokens.TemplateParamSeparator(), tokens.Text(text="j"),

  349.                 tokens.TemplateParamEquals(), tokens.HTMLEntityStart(),

  350.                 tokens.Text(text="nbsp"), tokens.HTMLEntityEnd(),

  351.                 tokens.TemplateClose()]

  352.         valid = wrap(

  353.             [Template(wraptext("a"), params=[Parameter(wraptext("1"), wraptext(

  354.             "b"), showkey=False), Parameter(wraptext("2"), wrap([Template(

  355.             wraptext("c"), params=[Parameter(wraptext("1"), wrap([Wikilink(

  356.             wraptext("d")), Argument(wraptext("e"))]), showkey=False)])]),

  357.             showkey=False)]), Wikilink(wraptext("f"), wrap([Argument(wraptext(

  358.             "g")), Comment(wraptext("h"))])), Template(wraptext("i"), params=[

  359.             Parameter(wraptext("j"), wrap([HTMLEntity("nbsp",

  360.             named=True)]))])])

  361.         self.assertWikicodeEqual(valid, self.builder.build(test))

  362. 

  363. if __name__ == "__main__":

  364.     unittest.main(verbosity=2)