third_party/WebKit/Source/core/dom/DocumentStatisticsCollectorTest.cpp - Issue 1419033004: Add feature extraction for distillability to Blink

Side by Side Diff: third_party/WebKit/Source/core/dom/DocumentStatisticsCollectorTest.cpp

Issue 1419033004: Add feature extraction for distillability to Blink (Closed) Base URL: https://chromium.googlesource.com/chromium/src.git@master

Patch Set: address comments, add saturation Created 5 years, 1 month ago

Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.

Jump to:

View unified diff | Download patch

« third_party/WebKit/Source/core/dom/DocumentStatisticsCollector.cpp ('K') | « third_party/WebKit/Source/core/dom/DocumentStatisticsCollector.cpp ('k') | third_party/WebKit/Source/web/WebDocument.cpp » ('j') | no next file with comments »
Toggle Intra-line Diffs ('i') | Expand Comments ('e') | Collapse Comments ('c') | Hide Comments ('s')

OLD	NEW
(Empty)
	1 // Copyright 2015 The Chromium Authors. All rights reserved.

	2 // Use of this source code is governed by a BSD-style license that can be

	3 // found in the LICENSE file.

	4

	5 #include "config.h"

	6 #include "core/dom/DocumentStatisticsCollector.h"

	7

	8 #include "core/dom/Document.h"

	9 #include "core/dom/DocumentVisibilityObserver.h"

	10 #include "core/frame/FrameView.h"

	11 #include "core/html/HTMLHeadElement.h"

	12 #include "core/html/HTMLLinkElement.h"

	13 #include "core/testing/DummyPageHolder.h"

	14 #include "public/platform/WebDistillability.h"

	15 #include <gmock/gmock.h>

	16 #include <gtest/gtest.h>

	17

	18 namespace blink {

	19

	20 class DocumentStatisticsCollectorTest : public ::testing::Test {

	21 protected:

	22 void SetUp() override;

	23

	24 #if ENABLE(OILPAN)

	25 void TearDown() override

	26 {

	27 Heap::collectAllGarbage();

	28 }

	29 #endif

	30

	31 Document& document() const { return m_dummyPageHolder->document(); }

	32

	33 void setHtmlInnerHTML(const String&);

	34

	35 private:

	36 OwnPtr<DummyPageHolder> m_dummyPageHolder;

	37 };

	38

	39 void DocumentStatisticsCollectorTest::SetUp()

	40 {

	41 m_dummyPageHolder = DummyPageHolder::create(IntSize(800, 600));

	42 }

	43

	44 void DocumentStatisticsCollectorTest::setHtmlInnerHTML(const String& htmlContent )

	45 {

	46 document().documentElement()->setInnerHTML((htmlContent), ASSERT_NO_EXCEPTIO N);

	47 document().view()->updateAllLifecyclePhases();

	48 }

	49

	50 // This test checks open graph articles can be recognized.

	51 TEST_F(DocumentStatisticsCollectorTest, HasOpenGraphArticle)

	52 {

	53 setHtmlInnerHTML(

	54 "<head>"

	55 // Note the case-insensitive matching of the word "article".

	56 " <meta property='og:type' content='arTiclE' />"

	57 "</head>"

	58 );

	59 WebDistillabilityFeatures features = DocumentStatisticsCollector::collectSta tistics(document());

	60

	61 EXPECT_EQ(true, features.openGraph);

	62 }

	63

	64 // This test checks non-existence of open graph articles can be recognized.

	65 TEST_F(DocumentStatisticsCollectorTest, NoOpenGraphArticle)

	66 {

	67 setHtmlInnerHTML(

	68 "<head>"

	69 " <meta property='og:type' content='movie' />"

	70 "</head>"

	71 );

	72 WebDistillabilityFeatures features = DocumentStatisticsCollector::collectSta tistics(document());

	73

	74 EXPECT_EQ(false, features.openGraph);

	75 }

	76

	77 // This test checks element counts are correct.

	78 TEST_F(DocumentStatisticsCollectorTest, CountElements)

	79 {

	80 setHtmlInnerHTML(

	81 "<form>"

	82 " <input type='text'>"

	83 " <input type='password'>"

	84 "</form>"

	85 "<pre></pre>"

	86 "<p><a> </a></p>"

	87 "<ul><li><p><a> </a></p></li></ul>"

	88 );

	89 WebDistillabilityFeatures features = DocumentStatisticsCollector::collectSta tistics(document());

	90

	91 EXPECT_EQ(false, features.openGraph);

	92

	93 EXPECT_EQ(10u, features.elementCount);

	94 EXPECT_EQ(2u, features.anchorCount);

	95 EXPECT_EQ(1u, features.formCount);

	96 EXPECT_EQ(1u, features.textInputCount);

	97 EXPECT_EQ(1u, features.passwordInputCount);

	98 EXPECT_EQ(2u, features.pCount);

	99 EXPECT_EQ(1u, features.preCount);

	100 }

	101

	102 // This test checks score calculations are correct.

	103 TEST_F(DocumentStatisticsCollectorTest, CountScore)

	104 {

	105 setHtmlInnerHTML(

	106 "<p class='menu' id='article'> 1 </p>" // trimmedTextContentLength = 1

	107 "<ul><li><p>12</p></li></ul>" // trimmedTextContentLength = 2, skipped b ecause under li

	108 "<p class='menu'>123</p>" // trimmedTextContentLength = 3, skipped becau se unlikelyCandidates

	109 "<p>"

	110 "12345678901234567890123456789012345678901234567890"

	111 "12345678901234567890123456789012345678901234567890"

	112 "12345678901234567890123456789012345678901234"

	113 "</p>" // trimmedTextContentLength = 144

	114 "<p style='display:none'>12345</p>" // trimmedTextContentLength = 5, ski pped because invisible

	115 "<div style='visibility:hidden'><p>123456</p></div>" // trimmedTextConte ntLength = 6, skipped because invisible

	116 "<p style='opacity:0'>1234567</p>" // trimmedTextContentLength = 7, skip ped because invisible

	117 "<p> <a href='#'> 12345 </a> 9 <b> </b> </p>" // trimmedTextContentLen gth = 9

	118 );

	119 WebDistillabilityFeatures features = DocumentStatisticsCollector::collectSta tistics(document());

	120

	121 const unsigned kParagraphLengthThreshold = 140;

	122

	123 EXPECT_DOUBLE_EQ(features.mozScore, sqrt(144 - kParagraphLengthThreshold));

	124 EXPECT_DOUBLE_EQ(features.mozScoreAllSqrt, 1 + sqrt(144) + sqrt(9));

	125 EXPECT_DOUBLE_EQ(features.mozScoreAllLinear, 1 + 144 + 9);

	126 }

	127

	128 } // namespace blink

OLD	NEW