github.com/linchen2chris/hugo@v0.0.0-20230307053224-cec209389705/helpers/content_test.go (about)

     1  // Copyright 2019 The Hugo Authors. All rights reserved.
     2  //
     3  // Licensed under the Apache License, Version 2.0 (the "License");
     4  // you may not use this file except in compliance with the License.
     5  // You may obtain a copy of the License at
     6  // http://www.apache.org/licenses/LICENSE-2.0
     7  //
     8  // Unless required by applicable law or agreed to in writing, software
     9  // distributed under the License is distributed on an "AS IS" BASIS,
    10  // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    11  // See the License for the specific language governing permissions and
    12  // limitations under the License.
    13  
    14  package helpers
    15  
    16  import (
    17  	"bytes"
    18  	"html/template"
    19  	"strings"
    20  	"testing"
    21  
    22  	"github.com/spf13/afero"
    23  
    24  	"github.com/gohugoio/hugo/common/loggers"
    25  	"github.com/gohugoio/hugo/config"
    26  
    27  	qt "github.com/frankban/quicktest"
    28  )
    29  
    30  const tstHTMLContent = "<!DOCTYPE html><html><head><script src=\"http://two/foobar.js\"></script></head><body><nav><ul><li hugo-nav=\"section_0\"></li><li hugo-nav=\"section_1\"></li></ul></nav><article>content <a href=\"http://two/foobar\">foobar</a>. Follow up</article><p>This is some text.<br>And some more.</p></body></html>"
    31  
    32  func TestTrimShortHTML(t *testing.T) {
    33  	tests := []struct {
    34  		input, output []byte
    35  	}{
    36  		{[]byte(""), []byte("")},
    37  		{[]byte("Plain text"), []byte("Plain text")},
    38  		{[]byte("  \t\n Whitespace text\n\n"), []byte("Whitespace text")},
    39  		{[]byte("<p>Simple paragraph</p>"), []byte("Simple paragraph")},
    40  		{[]byte("\n  \n \t  <p> \t Whitespace\nHTML  \n\t </p>\n\t"), []byte("Whitespace\nHTML")},
    41  		{[]byte("<p>Multiple</p><p>paragraphs</p>"), []byte("<p>Multiple</p><p>paragraphs</p>")},
    42  		{[]byte("<p>Nested<p>paragraphs</p></p>"), []byte("<p>Nested<p>paragraphs</p></p>")},
    43  		{[]byte("<p>Hello</p>\n<ul>\n<li>list1</li>\n<li>list2</li>\n</ul>"), []byte("<p>Hello</p>\n<ul>\n<li>list1</li>\n<li>list2</li>\n</ul>")},
    44  	}
    45  
    46  	c := newTestContentSpec()
    47  	for i, test := range tests {
    48  		output := c.TrimShortHTML(test.input)
    49  		if !bytes.Equal(test.output, output) {
    50  			t.Errorf("Test %d failed. Expected %q got %q", i, test.output, output)
    51  		}
    52  	}
    53  }
    54  
    55  func TestStripEmptyNav(t *testing.T) {
    56  	c := qt.New(t)
    57  	cleaned := stripEmptyNav([]byte("do<nav>\n</nav>\n\nbedobedo"))
    58  	c.Assert(cleaned, qt.DeepEquals, []byte("dobedobedo"))
    59  }
    60  
    61  func TestBytesToHTML(t *testing.T) {
    62  	c := qt.New(t)
    63  	c.Assert(BytesToHTML([]byte("dobedobedo")), qt.Equals, template.HTML("dobedobedo"))
    64  }
    65  
    66  func TestNewContentSpec(t *testing.T) {
    67  	cfg := config.NewWithTestDefaults()
    68  	c := qt.New(t)
    69  
    70  	cfg.Set("summaryLength", 32)
    71  	cfg.Set("buildFuture", true)
    72  	cfg.Set("buildExpired", true)
    73  	cfg.Set("buildDrafts", true)
    74  
    75  	spec, err := NewContentSpec(cfg, loggers.NewErrorLogger(), afero.NewMemMapFs(), nil)
    76  
    77  	c.Assert(err, qt.IsNil)
    78  	c.Assert(spec.summaryLength, qt.Equals, 32)
    79  	c.Assert(spec.BuildFuture, qt.Equals, true)
    80  	c.Assert(spec.BuildExpired, qt.Equals, true)
    81  	c.Assert(spec.BuildDrafts, qt.Equals, true)
    82  }
    83  
    84  var benchmarkTruncateString = strings.Repeat("This is a sentence about nothing.", 20)
    85  
    86  func BenchmarkTestTruncateWordsToWholeSentence(b *testing.B) {
    87  	c := newTestContentSpec()
    88  	b.ResetTimer()
    89  	for i := 0; i < b.N; i++ {
    90  		c.TruncateWordsToWholeSentence(benchmarkTruncateString)
    91  	}
    92  }
    93  
    94  func BenchmarkTestTruncateWordsToWholeSentenceOld(b *testing.B) {
    95  	c := newTestContentSpec()
    96  	b.ResetTimer()
    97  	for i := 0; i < b.N; i++ {
    98  		c.truncateWordsToWholeSentenceOld(benchmarkTruncateString)
    99  	}
   100  }
   101  
   102  func TestTruncateWordsToWholeSentence(t *testing.T) {
   103  	c := newTestContentSpec()
   104  	type test struct {
   105  		input, expected string
   106  		max             int
   107  		truncated       bool
   108  	}
   109  	data := []test{
   110  		{"a b c", "a b c", 12, false},
   111  		{"a b c", "a b c", 3, false},
   112  		{"a", "a", 1, false},
   113  		{"This is a sentence.", "This is a sentence.", 5, false},
   114  		{"This is also a sentence!", "This is also a sentence!", 1, false},
   115  		{"To be. Or not to be. That's the question.", "To be.", 1, true},
   116  		{" \nThis is not a sentence\nAnd this is another", "This is not a sentence", 4, true},
   117  		{"", "", 10, false},
   118  		{"This... is a more difficult test?", "This... is a more difficult test?", 1, false},
   119  	}
   120  	for i, d := range data {
   121  		c.summaryLength = d.max
   122  		output, truncated := c.TruncateWordsToWholeSentence(d.input)
   123  		if d.expected != output {
   124  			t.Errorf("Test %d failed. Expected %q got %q", i, d.expected, output)
   125  		}
   126  
   127  		if d.truncated != truncated {
   128  			t.Errorf("Test %d failed. Expected truncated=%t got %t", i, d.truncated, truncated)
   129  		}
   130  	}
   131  }
   132  
   133  func TestTruncateWordsByRune(t *testing.T) {
   134  	c := newTestContentSpec()
   135  	type test struct {
   136  		input, expected string
   137  		max             int
   138  		truncated       bool
   139  	}
   140  	data := []test{
   141  		{"", "", 1, false},
   142  		{"a b c", "a b c", 12, false},
   143  		{"a b c", "a b c", 3, false},
   144  		{"a", "a", 1, false},
   145  		{"Hello 中国", "", 0, true},
   146  		{"这是中文，全中文。", "这是中文，", 5, true},
   147  		{"Hello 中国", "Hello 中", 2, true},
   148  		{"Hello 中国", "Hello 中国", 3, false},
   149  		{"Hello中国 Good 好的", "Hello中国 Good 好", 9, true},
   150  		{"This is a sentence.", "This is", 2, true},
   151  		{"This is also a sentence!", "This", 1, true},
   152  		{"To be. Or not to be. That's the question.", "To be. Or not", 4, true},
   153  		{" \nThis is    not a sentence\n ", "This is not", 3, true},
   154  	}
   155  	for i, d := range data {
   156  		c.summaryLength = d.max
   157  		output, truncated := c.TruncateWordsByRune(strings.Fields(d.input))
   158  		if d.expected != output {
   159  			t.Errorf("Test %d failed. Expected %q got %q", i, d.expected, output)
   160  		}
   161  
   162  		if d.truncated != truncated {
   163  			t.Errorf("Test %d failed. Expected truncated=%t got %t", i, d.truncated, truncated)
   164  		}
   165  	}
   166  }
   167  
   168  func TestExtractTOCNormalContent(t *testing.T) {
   169  	content := []byte("<nav>\n<ul>\nTOC<li><a href=\"#")
   170  
   171  	actualTocLessContent, actualToc := ExtractTOC(content)
   172  	expectedTocLess := []byte("TOC<li><a href=\"#")
   173  	expectedToc := []byte("<nav id=\"TableOfContents\">\n<ul>\n")
   174  
   175  	if !bytes.Equal(actualTocLessContent, expectedTocLess) {
   176  		t.Errorf("Actual tocless (%s) did not equal expected (%s) tocless content", actualTocLessContent, expectedTocLess)
   177  	}
   178  
   179  	if !bytes.Equal(actualToc, expectedToc) {
   180  		t.Errorf("Actual toc (%s) did not equal expected (%s) toc content", actualToc, expectedToc)
   181  	}
   182  }
   183  
   184  func TestExtractTOCGreaterThanSeventy(t *testing.T) {
   185  	content := []byte("<nav>\n<ul>\nTOC This is a very long content which will definitely be greater than seventy, I promise you that.<li><a href=\"#")
   186  
   187  	actualTocLessContent, actualToc := ExtractTOC(content)
   188  	// Because the start of Toc is greater than 70+startpoint of <li> content and empty TOC will be returned
   189  	expectedToc := []byte("")
   190  
   191  	if !bytes.Equal(actualTocLessContent, content) {
   192  		t.Errorf("Actual tocless (%s) did not equal expected (%s) tocless content", actualTocLessContent, content)
   193  	}
   194  
   195  	if !bytes.Equal(actualToc, expectedToc) {
   196  		t.Errorf("Actual toc (%s) did not equal expected (%s) toc content", actualToc, expectedToc)
   197  	}
   198  }
   199  
   200  func TestExtractNoTOC(t *testing.T) {
   201  	content := []byte("TOC")
   202  
   203  	actualTocLessContent, actualToc := ExtractTOC(content)
   204  	expectedToc := []byte("")
   205  
   206  	if !bytes.Equal(actualTocLessContent, content) {
   207  		t.Errorf("Actual tocless (%s) did not equal expected (%s) tocless content", actualTocLessContent, content)
   208  	}
   209  
   210  	if !bytes.Equal(actualToc, expectedToc) {
   211  		t.Errorf("Actual toc (%s) did not equal expected (%s) toc content", actualToc, expectedToc)
   212  	}
   213  }
   214  
   215  var totalWordsBenchmarkString = strings.Repeat("Hugo Rocks ", 200)
   216  
   217  func TestTotalWords(t *testing.T) {
   218  	for i, this := range []struct {
   219  		s     string
   220  		words int
   221  	}{
   222  		{"Two, Words!", 2},
   223  		{"Word", 1},
   224  		{"", 0},
   225  		{"One, Two,      Three", 3},
   226  		{totalWordsBenchmarkString, 400},
   227  	} {
   228  		actualWordCount := TotalWords(this.s)
   229  
   230  		if actualWordCount != this.words {
   231  			t.Errorf("[%d] Actual word count (%d) for test string (%s) did not match %d", i, actualWordCount, this.s, this.words)
   232  		}
   233  	}
   234  }
   235  
   236  func BenchmarkTotalWords(b *testing.B) {
   237  	b.ResetTimer()
   238  	for i := 0; i < b.N; i++ {
   239  		wordCount := TotalWords(totalWordsBenchmarkString)
   240  		if wordCount != 400 {
   241  			b.Fatal("Wordcount error")
   242  		}
   243  	}
   244  }