summaryrefslogtreecommitdiff
path: root/support/gladtex/tests/test_htmlhandling.py
blob: 359ef36965644787bfa1bddebd40afc883bf3fe7 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
#pylint: disable=too-many-public-methods
from functools import reduce
import os, re, shutil, tempfile
import unittest
from gleetex import htmlhandling


excl_filename = htmlhandling.HtmlImageFormatter.EXCLUSION_FILE_NAME

HTML_SKELETON = '''<html><head><meta http-equiv="Content-Type" content="text/html; charset={0}" />
</head><body>{1}</body>'''

def read(file_name, mode='r', encoding='utf-8'):
    """Read the file, return the string. Close file properly."""
    with open(file_name, mode, encoding=encoding) as handle:
        return handle.read()


class HtmlparserTest(unittest.TestCase):
    def setUp(self):
        self.p = htmlhandling.EqnParser()

    def test_start_tags_are_parsed_literally(self):
        self.p.feed("<p   i='o'>")
        self.assertEqual(self.p.get_data()[0], "<p   i='o'>",
                "The HTML parser should copy start tags literally.")

    def test_that_end_tags_are_copied_literally(self):
        self.p.feed("</ p></P>")
        self.assertEqual(''.join(self.p.get_data()), "</ p></P>")

    def test_entities_are_unchanged(self):
        self.p.feed("&#xa;")
        self.assertEqual(self.p.get_data()[0], '&#xa;')

    def test_charsets_are_copied(self):
        self.p.feed('&gt;&rarr;')
        self.assertEqual(''.join(self.p.get_data()[0]), '&gt;&rarr;')

    def test_without_eqn_all_blocks_are_strings(self):
        self.p.feed("<html>\n<head/><body><p>42</p><h1>blah</h1></body></html>")
        self.assertTrue(reduce(lambda x,y: x and isinstance(y, str),
            self.p.get_data()), "all chunks have to be strings")

    def test_equation_is_detected(self):
        self.p.feed('<eq>foo \\pi</eq>')
        self.assertTrue(isinstance(self.p.get_data()[0], (tuple, list)))
        self.assertEqual(self.p.get_data()[0][2], 'foo \\pi')

    def test_tag_followed_by_eqn_is_correctly_recognized(self):
        self.p.feed('<p foo="bar"><eq>bar</eq>')
        self.assertEqual(self.p.get_data()[0], '<p foo="bar">')
        self.assertTrue(isinstance(self.p.get_data(), list), "second item of data must be equation data list")
    
    def test_document_with_tag_then_eqn_then_tag_works(self):
        self.p.feed('<div style="invalid">bar</div><eq>baz</eq><sometag>')
        eqn = None
        # test should not depend on a specific position of equation, search for
        # it
        data = self.p.get_data()
        for chunk in data:
            if isinstance(chunk, (tuple, list)):
                eqn = chunk
                break
        self.assertTrue(isinstance(data[0], str))
        self.assertTrue(eqn is not None,
                "No equation found, must be tuple/list object.")
        self.assertTrue(isinstance(data[-1], str))

    def test_equation_is_copied_literally(self):
        self.p.feed('<eq ignore="me">my\nlittle\n\\tau</eq>')
        self.assertEqual(self.p.get_data()[0][2], 'my\nlittle\n\\tau')

    def test_unclosed_eqns_are_detected(self):
        self.assertRaises(htmlhandling.ParseException, self.p.feed,
                '<p><eq>\\endless\\formula')

    def test_nested_formulas_trigger_exception(self):
        self.assertRaises(htmlhandling.ParseException, self.p.feed,
                "<eq>\\pi<eq></eq></eq>")
        self.assertRaises(htmlhandling.ParseException, self.p.feed,
                "<eq>\\pi<eq></p></eq>")

    def test_formulas_without_displaymath_attribute_are_detected(self):
        self.p.feed('<p><eq>\frac12</eq><br /><eq env="inline">bar</eq></p>')
        formulas = [c for c in self.p.get_data() if isinstance(c, (tuple, list))]
        self.assertEqual(len(formulas), 2) # there should be _2_ formulas
        self.assertEqual(formulas[0][1], False) # no displaymath
        self.assertEqual(formulas[1][1], False) # no displaymath

    def test_that_unclosed_formulas_detected(self):
            self.assertRaises(htmlhandling.ParseException, self.p.feed,
                "<eq>\\pi<eq></p>")
            self.assertRaises(htmlhandling.ParseException, self.p.feed,
                "<eq>\\pi")


    def test_formula_contains_only_formula(self):
        p = htmlhandling.EqnParser()
        p.feed("<p><eq>1<i<9</eq></p>")
        formula = next(e for e in p.get_data() if isinstance(e, (list, tuple)))
        self.assertEqual(formula[-1], "1<i<9")

        p = htmlhandling.EqnParser()
        p.feed('<p><eq env="displaymath">test</eq></p>')
        formula = next(e for e in p.get_data() if isinstance(e, (list, tuple)))
        self.assertEqual(formula[-1], "test")

        p = htmlhandling.EqnParser()
        p.feed("<p><eq>1<i<9</eq></p>")
        formula = next(e for e in p.get_data() if isinstance(e, (list, tuple)))
        self.assertEqual(formula[-1], "1<i<9")
    

    def test_formula_with_html_sequences_are_unescaped(self):
        self.p.feed('<eq>a&gt;b</eq>')
        formula = self.p.get_data()[0]
        self.assertEqual(formula[-1], "a>b")
        

    def test_displaymath_is_recognized(self):
        self.p.feed('<eq env="displaymath">\\sum\limits_{n=1}^{e^i} a^nl^n</eq>')
        self.assertEqual(self.p.get_data()[0][1], True) # displaymath flag set

    def test_encoding_is_parsed_from_HTML4(self):
        iso8859_1 = HTML_SKELETON.format('iso-8859-15', 'öäüß').encode('iso-8859-1')
        self.p.feed(iso8859_1)
        self.assertEqual(self.p._EqnParser__encoding, 'iso-8859-15')

    def test_encoding_is_parsed_from_HTML5(self):
        document = r"""<!DOCTYPE html>
            <html xmlns="http://www.w3.org/1999/xhtml" lang="" xml:lang="">
            <head><meta charset="utf-8" />
              <meta name="generator" content="pandoc" />
              </head><body><p>hi</p></body></html>"""
        self.p.feed(document.encode('utf-8'))
        self.assertEqual(self.p._EqnParser__encoding.lower(), 'utf-8')


    def test_strings_can_be_passed_tO_parser_as_well(self):
        # no exception - everything is working as expected
        self.p.feed(HTML_SKELETON.format('utf-8', 'æø'))

class GetPositionTest(unittest.TestCase):
    def test_that_line_number_is_correct(self):
        self.assertEqual(htmlhandling.get_position('jojo', 0)[0], 0)
        self.assertEqual(htmlhandling.get_position('jojo', 3)[0], 0)
        self.assertEqual(htmlhandling.get_position('a\njojo', 3)[0], 1)
        self.assertEqual(htmlhandling.get_position('a\n\njojo', 3)[0], 2)

    def test_that_position_on_line_is_correct(self):
        self.assertEqual(htmlhandling.get_position('jojo', 0)[1], 0)
        self.assertEqual(htmlhandling.get_position('jojo', 3)[1], 3)
        self.assertEqual(htmlhandling.get_position('a\njojo', 3)[1], 2)
        self.assertEqual(htmlhandling.get_position('a\n\njojo', 3)[1], 1)
        



class HtmlImageTest(unittest.TestCase):
    def setUp(self):
        self.pos = {'depth':99, 'height' : 88, 'width' : 77}
        self.original_directory = os.getcwd()
        self.tmpdir = tempfile.mkdtemp()
        os.chdir(self.tmpdir)

    def tearDown(self):
        os.chdir(self.original_directory)
        shutil.rmtree(self.tmpdir, ignore_errors=True)

    def test_that_no_file_is_written_if_no_content(self):
        with htmlhandling.HtmlImageFormatter('foo.html'):
            pass
        self.assertFalse(os.path.exists('foo.html')        )

    def test_file_if_written_when_content_exists(self):
        with htmlhandling.HtmlImageFormatter() as img:
            img.format_excluded(self.pos, '\\tau\\tau', 'foo.png')
        self.assertTrue(os.path.exists(excl_filename)        )

    def test_written_file_starts_and_ends_more_or_less_properly(self):
        with htmlhandling.HtmlImageFormatter('.') as img:
            img.format_excluded(self.pos, '\\tau\\tau', 'foo.png')
        data = read(htmlhandling.HtmlImageFormatter.EXCLUSION_FILE_NAME, 'r', encoding='utf-8')
        self.assertTrue('<html' in data and '</html>' in data)
        self.assertTrue('<body' in data and '</body>' in data)
        # make sure encoding is specified
        self.assertTrue('<meta' in data and 'charset=' in data)

    def test_id_contains_no_special_characters(self):
        data = htmlhandling.gen_id('\\tau!\'{}][~^')
        for character in {'!', "'", '\\', '{', '}'}:
            self.assertFalse(character in data)

    def test_formula_can_consist_only_of_numbers_and_id_is_generated(self):
        data = htmlhandling.gen_id('9*8*7=504')
        self.assertTrue(data.startswith('form'))
        self.assertTrue(data.endswith('504'))

    def test_that_empty_ids_raise_exception(self):
        self.assertRaises(ValueError, htmlhandling.gen_id, '')

    def test_that_same_characters_are_not_repeated(self):
        id = htmlhandling.gen_id("jo{{{{{{{{ha")
        self.assertEqual(id, "jo_ha")

    def test_that_ids_are_max_150_characters_wide(self):
        id = htmlhandling.gen_id('\\alpha\\cdot\\gamma + ' * 999)
        self.assertTrue(len(id) == 150)

    def test_that_ids_start_with_letter(self):
        id = htmlhandling.gen_id('{}\\[]ÖÖÖö9343...·tau')
        self.assertTrue(id[0].isalpha())

    def test_that_link_to_external_image_points_to_file_and_formula(self):
        with htmlhandling.HtmlImageFormatter() as img:
            formatted_img = img.format_excluded(self.pos, '\\tau\\tau', 'foo.png')
            expected_id = htmlhandling.gen_id('\\tau\\tau')
        external_file = read(excl_filename, 'r', encoding='utf-8')
        # find linked formula path
        href = re.search('href="(.*?)"', formatted_img)
        self.assertTrue(href != None)
        # extract path and id from it
        self.assertTrue('#' in href.groups()[0])
        path, id = href.groups()[0].split('#')
        self.assertEqual(path, excl_filename)
        self.assertEqual(id, expected_id)

        # check external file
        self.assertTrue('<p id' in external_file)
        self.assertTrue('="'+expected_id in external_file)


    def test_that_link_to_external_image_points_to_file_basepath_and_formula(self):
        os.mkdir('basepath')
        with htmlhandling.HtmlImageFormatter('basepath') as img:
            formatted_img = img.format_excluded(self.pos, '\\tau\\tau', 'foo.png')
            expected_id = htmlhandling.gen_id('\\tau\\tau')
        # find linked formula path
        href = re.search('href="(.*?)"', formatted_img)
        self.assertTrue(href != None)
        # extract path and id from it
        self.assertTrue('#' in href.groups()[0])
        path, id = href.groups()[0].split('#')
        self.assertEqual(path, 'basepath/' + excl_filename)
        self.assertEqual(id, expected_id)

    def test_height_and_width_is_in_formatted_html_img_tag(self):
        data = None
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            data = img.get_html_img(self.pos, '\\tau\\tau', 'foo.png')
        self.assertTrue('height=' in data and str(self.pos['height']) in data)
        self.assertTrue('width=' in data and str(self.pos['width']) in data)

    def test_no_formula_gets_lost_when_reparsing_external_formula_file(self):
        with htmlhandling.HtmlImageFormatter() as img:
            img.format_excluded(self.pos, '\\tau' * 999, 'foo.png')
        with htmlhandling.HtmlImageFormatter() as img:
            img.format_excluded(self.pos, '\\pi' * 666, 'foo_2.png')
        data = read(excl_filename)
        self.assertTrue('\\tau' in data)
        self.assertTrue('\\pi' in data)

    def test_too_long_formulas_are_not_outsourced_if_not_configured(self):
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            img.format(self.pos, '\\tau' * 999, 'foo.png')
        self.assertFalse(os.path.exists('foo.html'))

    def test_that_too_long_formulas_get_outsourced_if_configured(self):
        with htmlhandling.HtmlImageFormatter() as img:
            img.set_max_formula_length(90)
            img.set_exclude_long_formulas(True)
            img.format(self.pos, '\\tau' * 999, 'foo.png')
        self.assertTrue(os.path.exists(excl_filename))
        data = read(htmlhandling.HtmlImageFormatter.EXCLUSION_FILE_NAME)
        self.assertTrue('\\tau\\tau' in data)

    def test_url_is_included(self):
        prefix = "http://crustulus.de/blog"
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            img.set_url(prefix)
            data = img.format(self.pos, '\epsilon<0', 'foo.png')
            self.assertTrue( prefix in data)

    def test_url_doesnt_contain_double_slashes(self):
        prefix = "http://crustulus.de/blog/"
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            img.set_url(prefix)
            data = img.format(self.pos, r'\gamma\text{strahlung}', 'foo.png')
            self.assertFalse('//' in data.replace('http://','ignore'))

    # depth is used as negative offset, so negative depth should result in
    # positive offset
    def test_that_negative_depth_results_in_positive_offset(self):
        self.pos['depth'] = '-999'
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            data = img.format(self.pos, r'\gamma\text{strahlung}', 'foo.png')
            self.assertTrue('align: ' + str(self.pos['depth'])[1:] in data)

    def test_that_displaymath_is_set_or_unset(self):
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            data = img.format(self.pos, r'\gamma\text{strahlung}', 'foo.png',
                    True)
            self.assertTrue('="displaymath' in data)
            data = img.format(self.pos, r'\gamma\text{strahlung}', 'foo.png',
                    False)
            self.assertTrue('="inlinemath' in data)

    def test_that_alternative_css_class_is_set_correctly(self):
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            img.set_display_math_css_class('no1')
            img.set_inline_math_css_class('no2')
            data = img.format(self.pos, r'\gamma\text{strahlung}', 'foo.png',
                    True)
            self.assertTrue('="no1"' in data)
            data = img.format(self.pos, r'\gamma\text{strahlung}', 'foo.png',
                    False)
            self.assertTrue('="no2' in data)

    def test_that_unicode_is_replaced_if_requested(self):
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            img.set_replace_nonascii(True)
            data = img.format(self.pos, '←', 'foo.png')
            self.assertTrue('\\leftarrow' in data,
                    'expected: "\\leftarrow" to be in "%s"' % data)
        
    def test_that_unicode_is_kept_if_not_requested_to_replace(self):
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            img.set_replace_nonascii(False)
            data = img.format(self.pos, '←', 'foo.png')
            self.assertTrue('←' in data)
 
    def test_formatting_commands_are_stripped(self):
        with htmlhandling.HtmlImageFormatter('foo.html') as img:
            data = img.format(self.pos, 'a\,b\,c\,d', 'foo.png')
            self.assertTrue('a b c d' in data)
            data = img.format(self.pos, 'a\,b\;c\ d', 'foo.png')
            self.assertTrue('a b c d' in data)

            data = img.format(self.pos, '\Big\{foo\Big\}', 'foo.png')
            self.assertTrue('\{foo' in data and '\}' in data)
            data = img.format(self.pos, r'\left\{foo\right\}', 'foo.png')
            self.assertTrue('\{' in data and 'foo' in data and '\}' in data)
         

def htmleqn(formula, hr=True):
    """Format a formula to appear as if it would have been outsourced into an
    external file."""
    return '%s\n<p id="%s"><pre>%s</pre></span></p>\n' % (\
            ('<hr/>' if hr else ''), htmlhandling.gen_id(formula), formula)


class OutsourcingParserTest(unittest.TestCase):
    def setUp(self):
        self.html = ('<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"' +
        '\n  "http://www.w3.org/TR/html4/strict.dtd">\n<html>\n<head>\n' +
        '<meta http-equiv="content-type" content="text/html; charset=utf-8"/>' +
        '<title>Outsourced Formulas</title></head>\n<body>\n<h1>heading</h1>')

    def get_html(self, string):
        """Create html string with head / tail und put the specified string into
        it."""
        return self.html + string + '\n</body>\n</html>'


    def test_formulas_are_recognized(self):
        data = self.get_html(htmleqn('\\tau'))
        parser = htmlhandling.OutsourcedFormulaParser()
        parser.feed(data)
        self.assertEqual(len(parser.get_formulas()), 1)

    def test_formula_doesnt_contain_surrounding_rubbish(self):
        data = self.get_html(htmleqn('\\gamma'))
        parser = htmlhandling.OutsourcedFormulaParser()
        parser.feed(data)
        self.assertEqual(len(parser.get_formulas()), 1)
        key = next(iter(parser.get_formulas()))
        par = parser.get_formulas()[key]
        self.assertFalse('<h1' in par)
        self.assertFalse('body>' in par)
        self.assertFalse('hr' in par)

    def test_that_header_is_parsed_correctly(self):
        p = htmlhandling.OutsourcedFormulaParser()
        p.feed(self.get_html(htmleqn('test123', False)))
        head = p.get_head()
        self.assertTrue('DOCTYPE' in head)
        self.assertTrue('<html' in head)
        self.assertTrue('<title' in head)
        self.assertTrue('</title' in head)
        self.assertTrue('</head' in head)
        self.assertTrue('<meta' in head)
        self.assertTrue('charset=' in head)

    def test_multiple_formulas_are_recognized_correctly(self):
        p = htmlhandling.OutsourcedFormulaParser()
        p.feed(self.get_html(htmleqn('\\tau', False) + '\n' + 
            htmleqn('\\gamma') + '\n' + htmleqn('\\epsilon<0')))
        forms = p.get_formulas()
        self.assertEqual(len(forms), 3)
        self.assertTrue('\\gamma' in forms.values())
        self.assertTrue('\\gamma' in forms.values())
        self.assertTrue('\\epsilon<0' in forms.values())