Add a histogram to approximate the amount of content translated

Add Translate.Translation.LanguageDetection.ContentLength to approximate
the amount of content for translation. This histogram records the length
of the page content that was extracted at language identification time
every time a successful translation is finished.

Bug: 1220804
Change-Id: Id474d0ecb782865c8f60c055c241e5af003c3946
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2965863
Reviewed-by: Max Curran <curranmax@chromium.org>
Commit-Queue: Megan Jablonski <megjablon@chromium.org>
Cr-Commit-Position: refs/heads/master@{#894427}

components/translate/content/renderer/translate_agent.cc

@@ -179,6 +179,8 @@ void TranslateAgent::PageCaptured(const std::u16string& contents) {
     return;
   }
 
+  page_contents_length_ = contents.size();
+
   WebLanguageDetectionDetails web_detection_details =
       WebLanguageDetectionDetails::CollectLanguageDetectionDetails(document);
   std::string content_language = web_detection_details.content_language.Utf8();
@@ -491,6 +493,7 @@ void TranslateAgent::CheckTranslateStatus() {
     // Check JavaScript performance counters for UMA reports.
     ReportTimeToTranslate(
         ExecuteScriptAndGetDoubleResult("cr.googleTranslate.translationTime"));
+    ReportTranslatedLanguageDetectionContentLength(page_contents_length_);
 
     // Notify the browser we are done.
     std::move(translate_callback_pending_)
@@ -572,6 +575,7 @@ void TranslateAgent::ResetPage() {
   receiver_.reset();
   translate_callback_pending_.Reset();
   CancelPendingTranslation();
+  page_contents_length_ = 0;
 }
 
 void TranslateAgent::OnDestruct() {

components/translate/content/renderer/translate_agent.h

@@ -175,6 +175,8 @@ class TranslateAgent : public content::RenderFrameObserver,
   // The URL scheme for translate extensions.
   std::string extension_scheme_;
 
+  size_t page_contents_length_;
+
   // The task runner responsible for the translation task, freezing it
   // when the frame is backgrounded.
   scoped_refptr<base::SingleThreadTaskRunner> translate_task_runner_;

components/translate/core/common/translate_metrics.cc

@@ -29,6 +29,8 @@ const char kTranslatePageScheme[] = "Translate.PageScheme";
 const char kTranslateSimilarLanguageMatch[] = "Translate.SimilarLanguageMatch";
 const char kTranslateLanguageDeterminedDuration[] =
     "Translate.LanguageDeterminedDuration";
+const char kTranslatedLanguageDetectionContentLength[] =
+    "Translate.Translation.LanguageDetection.ContentLength";
 
 }  // namespace metrics_internal
 
@@ -82,4 +84,10 @@ void ReportLanguageDeterminedDuration(base::TimeTicks begin,
       metrics_internal::kTranslateLanguageDeterminedDuration, end - begin);
 }
 
+void ReportTranslatedLanguageDetectionContentLength(size_t content_length) {
+  base::UmaHistogramCounts100000(
+      metrics_internal::kTranslatedLanguageDetectionContentLength,
+      content_length);
+}
+
 }  // namespace translate

components/translate/core/common/translate_metrics.h

@@ -25,6 +25,7 @@ extern const char kTranslatePageScheme[];
 extern const char kTranslateSimilarLanguageMatch[];
 extern const char kTranslateLanguageDetectionConflict[];
 extern const char kTranslateLanguageDeterminedDuration[];
+extern const char kTranslatedLanguageDetectionContentLength[];
 
 }  // namespace metrics_internal
 
@@ -77,6 +78,9 @@ void ReportSimilarLanguageMatch(bool match);
 void ReportLanguageDeterminedDuration(base::TimeTicks begin,
                                       base::TimeTicks end);
 
+// Called after when a translation starts.
+void ReportTranslatedLanguageDetectionContentLength(size_t content_length);
+
 }  // namespace translate
 
 #endif  // COMPONENTS_TRANSLATE_CORE_COMMON_TRANSLATE_METRICS_H_

components/translate/core/common/translate_metrics_unittest.cc

@@ -231,4 +231,13 @@ TEST(TranslateMetricsTest, ReportSimilarLanguageMatch) {
   EXPECT_EQ(1, recorder.GetCount(kFalse));
 }
 
+TEST(TranslateMetricsTest, ReportTranslatedLanguageDetectionContentLength) {
+  MetricsRecorder recorder(
+      translate::metrics_internal::kTranslatedLanguageDetectionContentLength);
+  recorder.CheckTotalCount(0);
+  translate::ReportTranslatedLanguageDetectionContentLength(12345);
+  recorder.CheckValueInLogs(12345);
+  recorder.CheckTotalCount(1);
+}
+
 }  // namespace translate

tools/metrics/histograms/histograms_xml/translate/histograms.xml

@@ -1055,6 +1055,17 @@ reviews. Googlers can read more about this at go/gwsq-gerrit.
   </summary>
 </histogram>
 
+<histogram name="Translate.Translation.LanguageDetection.ContentLength"
+    units="characters" expires_after="2021-12-12">
+  <owner>megjablon@chromium.org</owner>
+  <owner>chrome-language@google.com</owner>
+  <summary>
+    The number of characters of page content extracted at language detection
+    time, recorded when a translation successfully finishes. This can be used as
+    a rough estimate for the amount of content translated.
+  </summary>
+</histogram>
+
 <histogram name="Translate.Translation.SourceLanguage" enum="CLD3LanguageCode"
     expires_after="2021-08-01">
   <owner>curranmax@google.com</owner>