Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(1)

Side by Side Diff: components/translate/core/browser/translate_ranker_impl.cc

Issue 2565873002: [translate] Add translate ranker model loader. (Closed)
Patch Set: comments from sdefresne Created 3 years, 9 months 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
OLDNEW
(Empty)
1 // Copyright 2016 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 "components/translate/core/browser/translate_ranker_impl.h"
6
7 #include <cmath>
8
9 #include "base/bind.h"
10 #include "base/bind_helpers.h"
11 #include "base/command_line.h"
12 #include "base/files/file_path.h"
13 #include "base/files/file_util.h"
14 #include "base/memory/ptr_util.h"
15 #include "base/metrics/histogram_macros.h"
16 #include "base/profiler/scoped_tracker.h"
17 #include "base/strings/string_number_conversions.h"
18 #include "base/strings/string_util.h"
19 #include "base/task_runner.h"
20 #include "base/threading/thread_task_runner_handle.h"
21 #include "components/metrics/proto/translate_event.pb.h"
22 #include "components/translate/core/browser/proto/ranker_model.pb.h"
23 #include "components/translate/core/browser/proto/translate_ranker_model.pb.h"
24 #include "components/translate/core/browser/ranker_model.h"
25 #include "components/translate/core/browser/translate_download_manager.h"
26 #include "components/translate/core/browser/translate_prefs.h"
27 #include "components/translate/core/browser/translate_url_fetcher.h"
28 #include "components/translate/core/common/translate_switches.h"
29 #include "components/variations/variations_associated_data.h"
30 #include "url/gurl.h"
31
32 namespace translate {
33
34 namespace {
35
36 using chrome_intelligence::RankerModel;
37 using chrome_intelligence::RankerModelProto;
38 using chrome_intelligence::TranslateRankerModel;
39
40 const double kTranslationOfferThreshold = 0.5;
41
42 const char kTranslateRankerModelFileName[] = "Translate Ranker Model";
43 const char kUmaPrefix[] = "Translate.Ranker";
44 const char kUnknown[] = "UNKNOWN";
45
46 double Sigmoid(double x) {
47 return 1.0 / (1.0 + exp(-x));
48 }
49
50 double SafeRatio(int numerator, int denominator) {
51 return denominator ? (numerator / static_cast<double>(denominator)) : 0.0;
52 }
53
54 double ScoreComponent(const google::protobuf::Map<std::string, float>& weights,
55 const std::string& key) {
56 auto i = weights.find(base::ToLowerASCII(key));
57 if (i == weights.end())
58 i = weights.find(kUnknown);
59 return i == weights.end() ? 0.0 : i->second;
60 }
61
62 RankerModelStatus ValidateModel(const RankerModel& model) {
63 if (model.proto().model_case() != RankerModelProto::kTranslate)
64 return RankerModelStatus::VALIDATION_FAILED;
65
66 if (model.proto().translate().model_revision_case() !=
67 TranslateRankerModel::kLogisticRegressionModel) {
68 return RankerModelStatus::INCOMPATIBLE;
69 }
70
71 return RankerModelStatus::OK;
72 }
73
74 } // namespace
75
76 const base::Feature kTranslateRankerQuery{"TranslateRankerQuery",
77 base::FEATURE_DISABLED_BY_DEFAULT};
78
79 const base::Feature kTranslateRankerEnforcement{
80 "TranslateRankerEnforcement", base::FEATURE_DISABLED_BY_DEFAULT};
81
82 const base::Feature kTranslateRankerLogging{"TranslateRankerLogging",
83 base::FEATURE_DISABLED_BY_DEFAULT};
84
85 TranslateRankerFeatures::TranslateRankerFeatures() {}
86
87 TranslateRankerFeatures::TranslateRankerFeatures(int accepted,
88 int denied,
89 int ignored,
90 const std::string& src,
91 const std::string& dst,
92 const std::string& cntry,
93 const std::string& locale)
94 : accepted_count(accepted),
95 denied_count(denied),
96 ignored_count(ignored),
97 total_count(accepted_count + denied_count + ignored_count),
98 src_lang(src),
99 dst_lang(dst),
100 country(cntry),
101 app_locale(locale),
102 accepted_ratio(SafeRatio(accepted_count, total_count)),
103 denied_ratio(SafeRatio(denied_count, total_count)),
104 ignored_ratio(SafeRatio(ignored_count, total_count)) {}
105
106 TranslateRankerFeatures::TranslateRankerFeatures(const TranslatePrefs& prefs,
107 const std::string& src,
108 const std::string& dst,
109 const std::string& locale)
110 : TranslateRankerFeatures(prefs.GetTranslationAcceptedCount(src),
111 prefs.GetTranslationDeniedCount(src),
112 prefs.GetTranslationIgnoredCount(src),
113 src,
114 dst,
115 prefs.GetCountry(),
116 locale) {}
117
118 TranslateRankerFeatures::~TranslateRankerFeatures() {}
119
120 void TranslateRankerFeatures::WriteTo(std::ostream& stream) const {
121 stream << "src_lang='" << src_lang << "', "
122 << "dst_lang='" << dst_lang << "', "
123 << "country='" << country << "', "
124 << "app_locale='" << app_locale << "', "
125 << "accept_count=" << accepted_count << ", "
126 << "denied_count=" << denied_count << ", "
127 << "ignored_count=" << ignored_count << ", "
128 << "total_count=" << total_count << ", "
129 << "accept_ratio=" << accepted_ratio << ", "
130 << "decline_ratio=" << denied_ratio << ", "
131 << "ignore_ratio=" << ignored_ratio;
132 }
133
134 TranslateRankerImpl::TranslateRankerImpl(const base::FilePath& model_path,
135 const GURL& model_url)
136 : weak_ptr_factory_(this) {
137 model_loader_ = base::MakeUnique<RankerModelLoader>(
138 base::Bind(&ValidateModel),
139 base::Bind(&TranslateRankerImpl::OnModelAvailable,
140 weak_ptr_factory_.GetWeakPtr()),
141 model_path, model_url, kUmaPrefix);
142 model_loader_->Start();
143 }
144
145 TranslateRankerImpl::~TranslateRankerImpl() {}
146
147 // static
148 base::FilePath TranslateRankerImpl::GetModelPath(
149 const base::FilePath& data_dir) {
150 if (data_dir.empty())
151 return base::FilePath();
152
153 // Otherwise, look for the file in data dir.
154 return data_dir.AppendASCII(kTranslateRankerModelFileName);
155 }
156
157 // static
158 GURL TranslateRankerImpl::GetModelURL() {
159 // Allow override of the ranker model URL from the command line.
160 std::string raw_url;
161 base::CommandLine* command_line = base::CommandLine::ForCurrentProcess();
162 if (command_line->HasSwitch(switches::kTranslateRankerModelURL)) {
163 raw_url =
164 command_line->GetSwitchValueASCII(switches::kTranslateRankerModelURL);
165 } else {
166 // Otherwise take the ranker model URL from the ranker query variation.
167 raw_url = variations::GetVariationParamValueByFeature(
168 kTranslateRankerQuery, switches::kTranslateRankerModelURL);
169 }
170
171 DVLOG(3) << switches::kTranslateRankerModelURL << " = " << raw_url;
172
173 return GURL(raw_url);
174 }
175
176 bool TranslateRankerImpl::IsLoggingEnabled() {
177 return base::FeatureList::IsEnabled(kTranslateRankerLogging);
178 }
179
180 bool TranslateRankerImpl::IsQueryEnabled() {
181 return base::FeatureList::IsEnabled(kTranslateRankerQuery);
182 }
183
184 bool TranslateRankerImpl::IsEnforcementEnabled() {
185 return base::FeatureList::IsEnabled(kTranslateRankerEnforcement);
186 }
187
188 int TranslateRankerImpl::GetModelVersion() const {
189 return model_ ? model_->proto().translate().version() : 0;
190 }
191
192 bool TranslateRankerImpl::ShouldOfferTranslation(
193 const TranslatePrefs& translate_prefs,
194 const std::string& src_lang,
195 const std::string& dst_lang) {
196 DCHECK(sequence_checker_.CalledOnValidSequence());
197 // The ranker is a gate in the "show a translation prompt" flow. To retain
198 // the pre-existing functionality, it defaults to returning true in the
199 // absence of a model or if enforcement is disabled. As this is ranker is
200 // subsumed into a more general assist ranker, this default will go away
201 // (or become False).
202 const bool kDefaultResponse = true;
203
204 // If we don't have a model, request one and return the default.
205 if (model_ == nullptr) {
206 return kDefaultResponse;
207 }
208
209 SCOPED_UMA_HISTOGRAM_TIMER("Translate.Ranker.Timer.ShouldOfferTranslation");
210
211 // TODO(rogerm): Remove ScopedTracker below once crbug.com/646711 is closed.
212 tracked_objects::ScopedTracker tracking_profile(
213 FROM_HERE_WITH_EXPLICIT_FUNCTION(
214 "646711 translate::TranslateRankerImpl::ShouldOfferTranslation"));
215
216 TranslateRankerFeatures features(
217 translate_prefs, src_lang, dst_lang,
218 TranslateDownloadManager::GetInstance()->application_locale());
219
220 double score = CalculateScore(features);
221
222 DVLOG(2) << "TranslateRankerImpl::ShouldOfferTranslation: "
223 << "Score = " << score << ", Features=[" << features << "]";
224
225 bool result = (score >= kTranslationOfferThreshold);
226
227 UMA_HISTOGRAM_BOOLEAN("Translate.Ranker.QueryResult", result);
228
229 return result;
230 }
231
232 double TranslateRankerImpl::CalculateScore(
233 const TranslateRankerFeatures& features) {
234 DCHECK(sequence_checker_.CalledOnValidSequence());
235 SCOPED_UMA_HISTOGRAM_TIMER("Translate.Ranker.Timer.CalculateScore");
236 DCHECK(model_ != nullptr);
237 const TranslateRankerModel::LogisticRegressionModel& logit =
238 model_->proto().translate().logistic_regression_model();
239
240 double dot_product =
241 (features.accepted_count * logit.accept_count_weight()) +
242 (features.denied_count * logit.decline_count_weight()) +
243 (features.ignored_count * logit.ignore_count_weight()) +
244 (features.accepted_ratio * logit.accept_ratio_weight()) +
245 (features.denied_ratio * logit.decline_ratio_weight()) +
246 (features.ignored_ratio * logit.ignore_ratio_weight()) +
247 ScoreComponent(logit.source_language_weight(), features.src_lang) +
248 ScoreComponent(logit.dest_language_weight(), features.dst_lang) +
249 ScoreComponent(logit.country_weight(), features.country) +
250 ScoreComponent(logit.locale_weight(), features.app_locale);
251
252 return Sigmoid(dot_product + logit.bias());
253 }
254
255 void TranslateRankerImpl::FlushTranslateEvents(
256 std::vector<metrics::TranslateEventProto>* events) {
257 DCHECK(sequence_checker_.CalledOnValidSequence());
258 DVLOG(3) << "Flushing translate ranker events.";
259 events->swap(event_cache_);
260 event_cache_.clear();
261 }
262
263 void TranslateRankerImpl::AddTranslateEvent(
264 const metrics::TranslateEventProto& event) {
265 DCHECK(sequence_checker_.CalledOnValidSequence());
266 DVLOG(3) << "Adding translate ranker event.";
267 if (IsLoggingEnabled())
268 event_cache_.push_back(event);
269 }
270
271 void TranslateRankerImpl::OnModelAvailable(std::unique_ptr<RankerModel> model) {
272 DCHECK(sequence_checker_.CalledOnValidSequence());
273 model_ = std::move(model);
274 }
275
276 } // namespace translate
277
278 std::ostream& operator<<(std::ostream& stream,
279 const translate::TranslateRankerFeatures& features) {
280 features.WriteTo(stream);
281 return stream;
282 }
OLDNEW

Powered by Google App Engine
This is Rietveld 408576698