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
|
// Copyright 2019 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "components/page_image_annotation/content/renderer/content_page_annotator_driver.h"
#include <optional>
#include "base/base64.h"
#include "base/functional/bind.h"
#include "base/task/single_thread_task_runner.h"
#include "content/public/renderer/render_frame.h"
#include "crypto/sha2.h"
#include "services/image_annotation/public/mojom/image_annotation.mojom.h"
#include "third_party/blink/public/platform/browser_interface_broker_proxy.h"
#include "third_party/blink/public/platform/web_string.h"
#include "third_party/blink/public/web/web_document.h"
#include "third_party/blink/public/web/web_local_frame.h"
#include "third_party/blink/public/web/web_node.h"
namespace page_image_annotation {
namespace {
namespace ia_mojom = image_annotation::mojom;
// The number of milliseconds to wait before traversing the DOM to find image
// elements.
constexpr int kDomCrawlDelayMs = 3000;
// Attempts to produce image metadata for the given element. Will produce a null
// value if the element has a missing or malformed src attribute.
std::optional<PageAnnotator::ImageMetadata> ProduceMetadata(
const GURL& page_url,
const blink::WebElement element,
const uint64_t node_id) {
const std::string source_id = ContentPageAnnotatorDriver::GenerateSourceId(
page_url, element.GetAttribute("src").Utf8());
if (source_id.empty())
return std::nullopt;
return PageAnnotator::ImageMetadata{node_id, source_id};
}
mojo::PendingRemote<ia_mojom::Annotator> RequestAnnotator(
content::RenderFrame* const render_frame) {
mojo::PendingRemote<ia_mojom::Annotator> annotator;
render_frame->GetBrowserInterfaceBroker().GetInterface(
annotator.InitWithNewPipeAndPassReceiver());
return annotator;
}
} // namespace
ContentPageAnnotatorDriver::ContentPageAnnotatorDriver(
content::RenderFrame* const render_frame)
: RenderFrameObserver(render_frame),
RenderFrameObserverTracker<ContentPageAnnotatorDriver>(render_frame),
next_node_id_(1),
page_annotator_(RequestAnnotator(render_frame)) {}
ContentPageAnnotatorDriver::~ContentPageAnnotatorDriver() = default;
// static
ContentPageAnnotatorDriver* ContentPageAnnotatorDriver::GetOrCreate(
content::RenderFrame* const render_frame) {
ContentPageAnnotatorDriver* const existing = Get(render_frame);
if (existing)
return existing;
return new ContentPageAnnotatorDriver(render_frame);
}
PageAnnotator& ContentPageAnnotatorDriver::GetPageAnnotator() {
return page_annotator_;
}
blink::WebElement ContentPageAnnotatorDriver::GetElement(
const uint64_t node_id) {
const auto lookup = tracked_elements_.find(node_id);
if (lookup == tracked_elements_.end())
return blink::WebElement();
return lookup->second.second;
}
// static
std::string ContentPageAnnotatorDriver::GenerateSourceId(
const GURL& page_url,
const std::string& uri_fragment) {
if (uri_fragment.empty())
return std::string();
const GURL src_url = page_url.Resolve(uri_fragment);
if (!src_url.is_valid())
return std::string();
// Assign a source ID: either the URL of this image (if it can be resolved) or
// a hash of its data URI.
if (src_url.SchemeIs("data")) {
const std::string& content = src_url.GetContent();
if (!content.empty()) {
// We use SHA256 since it has comparable (<2x) speed to e.g. crc32, but
// has no known collisions (which could lead to cached results for another
// image being returned for this one).
return base::Base64Encode(crypto::SHA256HashString(content));
}
} else if (src_url.SchemeIs("http") || src_url.SchemeIs("https")) {
return src_url.spec();
}
return std::string();
}
void ContentPageAnnotatorDriver::DidDispatchDOMContentLoadedEvent() {
if (!render_frame()->IsMainFrame())
return;
// Cancel any pending DOM crawl.
weak_ptr_factory_.InvalidateWeakPtrs();
// Stop tracking old elements. After a page refresh, we're the only thing
// keeping these elements alive (i.e. they are not still being displayed to
// the user); we need to let Blink garbage collect them.
for (const auto& entry : tracked_elements_) {
page_annotator_.ImageRemoved(entry.first);
}
tracked_elements_.clear();
// Schedule new DOM crawl after page has likely reached a stable state.
//
// TODO(crbug.com/41432474): this approach is ad-hoc (e.g. uses a heuristic
// delay to wait for a stable DOM) and can cause jank;
// reinvestigate it once we are done prototyping the
// feature.
base::SingleThreadTaskRunner::GetCurrentDefault()->PostDelayedTask(
FROM_HERE,
base::BindOnce(&ContentPageAnnotatorDriver::FindAndTrackImages,
weak_ptr_factory_.GetWeakPtr()),
base::Milliseconds(kDomCrawlDelayMs));
}
void ContentPageAnnotatorDriver::OnDestruct() {
delete this;
}
void ContentPageAnnotatorDriver::FindImages(const GURL& page_url,
blink::WebElement element) {
if (element.ImageContents().isNull()) {
// This element is not an image but it could have children that are.
for (blink::WebNode child = element.FirstChild(); !child.IsNull();
child = child.NextSibling()) {
if (child.IsElementNode())
FindImages(page_url, child.To<blink::WebElement>());
}
} else {
// This element is an image; attempt to produce metadata for it and begin
// tracking.
const std::optional<PageAnnotator::ImageMetadata> metadata =
ProduceMetadata(page_url, element, next_node_id_);
if (metadata.has_value())
tracked_elements_.insert({next_node_id_++, {*metadata, element}});
}
}
void ContentPageAnnotatorDriver::FindAndTrackImages() {
const blink::WebDocument doc = render_frame()->GetWebFrame()->GetDocument();
if (doc.IsNull() || doc.Body().IsNull())
return;
const GURL page_url(doc.Url().GetString().Utf8(), doc.Url().GetParsed(),
doc.Url().IsValid());
FindImages(page_url, doc.Body());
// Inform the PageAnnotator of the new images.
for (const auto& entry : tracked_elements_) {
page_annotator_.ImageAddedOrPossiblyModified(
entry.second.first,
base::BindRepeating(&ContentPageAnnotatorDriver::GetBitmapForId,
base::Unretained(this), entry.first));
}
}
SkBitmap ContentPageAnnotatorDriver::GetBitmapForId(const uint64_t node_id) {
return GetElement(node_id).ImageContents();
}
} // namespace page_image_annotation
|
