Add pdf document load observer

Adds observer for pdf document load complete event.
Also changes the consumers (contextual cueing, page_content_annotations) to listen to the event before querying page count.

Change-Id: Iecb4571aeb61c52dde6fe1bc770e5bdb372a1525
Bug: 395086836
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6250270
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Raj T <rajendrant@chromium.org>
Reviewed-by: Will Harris <wfh@chromium.org>
Reviewed-by: Duncan Mercer <mercerd@google.com>
Cr-Commit-Position: refs/heads/main@{#1420001}

chrome/browser/contextual_cueing/contextual_cueing_page_data.cc

@@ -116,10 +116,31 @@ ContextualCueingPageData::DidMatchCueingConditions(
 
 #if BUILDFLAG(ENABLE_PDF)
 void ContextualCueingPageData::RequestPdfPageCount() {
-  CHECK(page().GetContentsMimeType() == pdf::kPDFMimeType);
+  CHECK_EQ(pdf::kPDFMimeType, page().GetContentsMimeType());
-  pdf::PDFDocumentHelper* pdf_helper =
+
-      pdf::PDFDocumentHelper::MaybeGetForWebContents(
+  auto* pdf_helper = pdf::PDFDocumentHelper::MaybeGetForWebContents(
-          content::WebContents::FromRenderFrameHost(&page().GetMainDocument()));
+      content::WebContents::FromRenderFrameHost(&page().GetMainDocument()));
+  if (!pdf_helper) {
+    return;
+  }
+  if (!pdf_helper->IsDocumentLoadComplete()) {
+    // Wait for the PDF to load.
+    pdf_load_obseration_.Observe(pdf_helper);
+    return;
+  }
+  // Fetch zero PDF bytes to just receive the total page count.
+  pdf_helper->GetPdfBytes(
+      /*size_limit=*/0,
+      base::BindOnce(&ContextualCueingPageData::OnPdfPageCountReceived,
+                     weak_factory_.GetWeakPtr()));
+}
+
+void ContextualCueingPageData::OnDocumentLoadComplete() {
+  CHECK_EQ(pdf::kPDFMimeType, page().GetContentsMimeType());
+  pdf_load_obseration_.Reset();
+
+  auto* pdf_helper = pdf::PDFDocumentHelper::MaybeGetForWebContents(
+      content::WebContents::FromRenderFrameHost(&page().GetMainDocument()));
   if (pdf_helper) {
     // Fetch zero PDF bytes to just receive the total page count.
     pdf_helper->GetPdfBytes(

chrome/browser/contextual_cueing/contextual_cueing_page_data.h

@@ -5,17 +5,25 @@
 #ifndef CHROME_BROWSER_CONTEXTUAL_CUEING_CONTEXTUAL_CUEING_PAGE_DATA_H_
 #define CHROME_BROWSER_CONTEXTUAL_CUEING_CONTEXTUAL_CUEING_PAGE_DATA_H_
 
+#include "base/scoped_observation.h"
 #include "components/optimization_guide/proto/contextual_cueing_metadata.pb.h"
 #include "components/pdf/common/constants.h"
 #include "content/public/browser/page_user_data.h"
 #include "pdf/buildflags.h"
 #include "pdf/mojom/pdf.mojom.h"
 
+#if BUILDFLAG(ENABLE_PDF)
+#include "components/pdf/browser/pdf_document_helper.h"
+#endif  // BUILDFLAG(ENABLE_PDF)
+
 namespace contextual_cueing {
 
 // Decider for contextual cueing that is scoped to `Page`.
-class ContextualCueingPageData
+class ContextualCueingPageData :
-    : public content::PageUserData<ContextualCueingPageData> {
+#if BUILDFLAG(ENABLE_PDF)
+    public pdf::PDFDocumentHelper::Observer,
+#endif  // BUILDFLAG(ENABLE_PDF)
+    public content::PageUserData<ContextualCueingPageData> {
  public:
   using CueingDecisionCallback = base::OnceCallback<void(const std::string&)>;
 
@@ -48,11 +56,16 @@ class ContextualCueingPageData
       const optimization_guide::proto::GlicCueingConfiguration& config);
 
 #if BUILDFLAG(ENABLE_PDF)
+  // Requests for page count if this is a PDF page.
   void RequestPdfPageCount();
 
+  // Invoked when page count is received.
   void OnPdfPageCountReceived(pdf::mojom::PdfListener::GetPdfBytesStatus status,
                               const std::vector<uint8_t>& bytes,
                               uint32_t page_count);
+
+  // pdf::PDFDocumentHelper::Observer:
+  void OnDocumentLoadComplete() override;
 #endif  // BUILDFLAG(ENABLE_PDF)
 
   const optimization_guide::proto::GlicContextualCueingMetadata metadata_;
@@ -64,6 +77,12 @@ class ContextualCueingPageData
 
   CueingDecisionCallback cueing_decision_callback_;
 
+#if BUILDFLAG(ENABLE_PDF)
+  base::ScopedObservation<pdf::PDFDocumentHelper,
+                          pdf::PDFDocumentHelper::Observer>
+      pdf_load_obseration_{this};
+#endif  // BUILDFLAG(ENABLE_PDF)
+
   base::WeakPtrFactory<ContextualCueingPageData> weak_factory_{this};
 
   PAGE_USER_DATA_KEY_DECL();

chrome/browser/page_content_annotations/annotate_page_content_request.cc

@@ -228,8 +228,30 @@ void AnnotatedPageContentRequest::OnInnerTextReceived(
 
 #if BUILDFLAG(ENABLE_PDF)
 void AnnotatedPageContentRequest::RequestPdfPageCount() {
-  CHECK(web_contents_->GetContentsMimeType() == pdf::kPDFMimeType);
+  CHECK_EQ(pdf::kPDFMimeType, web_contents_->GetContentsMimeType());
-  pdf::PDFDocumentHelper* pdf_helper =
+  auto* pdf_helper =
+      pdf::PDFDocumentHelper::MaybeGetForWebContents(web_contents_);
+  if (!pdf_helper) {
+    return;
+  }
+  if (!pdf_helper->IsDocumentLoadComplete()) {
+    // Wait for the PDF to load.
+    pdf_load_obseration_.Observe(pdf_helper);
+    return;
+  }
+  // Fetch zero PDF bytes to just receive the total page count.
+  pdf_helper->GetPdfBytes(
+      /*size_limit=*/0,
+      base::BindOnce(
+          &RecordPdfPageCountMetrics,
+          web_contents_->GetPrimaryMainFrame()->GetPageUkmSourceId()));
+}
+
+void AnnotatedPageContentRequest::OnDocumentLoadComplete() {
+  CHECK_EQ(pdf::kPDFMimeType, web_contents_->GetContentsMimeType());
+  pdf_load_obseration_.Reset();
+
+  auto* pdf_helper =
       pdf::PDFDocumentHelper::MaybeGetForWebContents(web_contents_);
   if (pdf_helper) {
     // Fetch zero PDF bytes to just receive the total page count.

chrome/browser/page_content_annotations/annotate_page_content_request.h

@@ -1,20 +1,30 @@
 // Copyright 2025 The Chromium Authors
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
+
 #ifndef CHROME_BROWSER_PAGE_CONTENT_ANNOTATIONS_PAGE_CONTENT_ANNOTATIONS_ANNOTATE_PAGE_CONTENT_REQUEST_H_
 #define CHROME_BROWSER_PAGE_CONTENT_ANNOTATIONS_PAGE_CONTENT_ANNOTATIONS_ANNOTATE_PAGE_CONTENT_REQUEST_H_
 
+#include "base/scoped_observation.h"
 #include "chrome/browser/content_extraction/inner_text.h"
 #include "components/optimization_guide/proto/features/common_quality_data.pb.h"
 #include "content/public/browser/web_contents.h"
 #include "pdf/buildflags.h"
 #include "third_party/blink/public/mojom/content_extraction/ai_page_content.mojom.h"
 
+#if BUILDFLAG(ENABLE_PDF)
+#include "components/pdf/browser/pdf_document_helper.h"
+#endif  // BUILDFLAG(ENABLE_PDF)
+
 namespace page_content_annotations {
 
 // Class for deciding when a page is ready for getting page content, and
 // extracts page content.
-class AnnotatedPageContentRequest {
+class AnnotatedPageContentRequest
+#if BUILDFLAG(ENABLE_PDF)
+    : public pdf::PDFDocumentHelper::Observer
+#endif  // BUILDFLAG(ENABLE_PDF)
+{
  public:
   static std::unique_ptr<AnnotatedPageContentRequest> MaybeCreate(
       content::WebContents* web_contents);
@@ -25,7 +35,11 @@ class AnnotatedPageContentRequest {
   AnnotatedPageContentRequest(const AnnotatedPageContentRequest&) = delete;
   AnnotatedPageContentRequest& operator=(const AnnotatedPageContentRequest&) =
       delete;
+#if BUILDFLAG(ENABLE_PDF)
+  ~AnnotatedPageContentRequest() override;
+#else
   ~AnnotatedPageContentRequest();
+#endif  // BUILDFLAG(ENABLE_PDF)
 
   void PrimaryPageChanged();
 
@@ -53,6 +67,9 @@ class AnnotatedPageContentRequest {
 
 #if BUILDFLAG(ENABLE_PDF)
   void RequestPdfPageCount();
+
+  // pdf::PDFDocumentHelper::Observer:
+  void OnDocumentLoadComplete() override;
 #endif  // BUILDFLAG(ENABLE_PDF)
 
   const raw_ptr<content::WebContents> web_contents_;
@@ -66,6 +83,12 @@ class AnnotatedPageContentRequest {
   bool waiting_for_load_ = false;
   bool waiting_for_fcp_ = false;
 
+#if BUILDFLAG(ENABLE_PDF)
+  base::ScopedObservation<pdf::PDFDocumentHelper,
+                          pdf::PDFDocumentHelper::Observer>
+      pdf_load_obseration_{this};
+#endif  // BUILDFLAG(ENABLE_PDF)
+
   base::WeakPtrFactory<AnnotatedPageContentRequest> weak_factory_{this};
 };
 

components/pdf/browser/pdf_document_helper.cc

@@ -262,6 +262,14 @@ void PDFDocumentHelper::GetMostVisiblePageIndex(
   remote_pdf_client_->GetMostVisiblePageIndex(std::move(callback));
 }
 
+void PDFDocumentHelper::AddObserver(Observer* observer) {
+  observers_.AddObserver(observer);
+}
+
+void PDFDocumentHelper::RemoveObserver(Observer* observer) {
+  observers_.RemoveObserver(observer);
+}
+
 void PDFDocumentHelper::OnSelectionEvent(ui::SelectionEventType event) {
   // Should be handled by `TouchSelectionControllerClientAura`.
   NOTREACHED();
@@ -381,6 +389,14 @@ void PDFDocumentHelper::InitTouchSelectionClientManager() {
   touch_selection_controller_client_manager_->AddObserver(this);
 }
 
+void PDFDocumentHelper::OnDocumentLoadComplete() {
+  // Only notify the consumers on first load complete.
+  if (!is_document_load_complete_) {
+    is_document_load_complete_ = true;
+    observers_.Notify(&Observer::OnDocumentLoadComplete);
+  }
+}
+
 void PDFDocumentHelper::SaveUrlAs(const GURL& url,
                                   network::mojom::ReferrerPolicy policy) {
   client_->OnSaveURL(&GetWebContents());

components/pdf/browser/pdf_document_helper.h

@@ -8,6 +8,7 @@
 #include <memory>
 
 #include "base/memory/raw_ptr.h"
+#include "base/observer_list.h"
 #include "content/public/browser/document_user_data.h"
 #include "content/public/browser/render_frame_host_receiver_set.h"
 #include "content/public/browser/render_widget_host_observer.h"
@@ -38,6 +39,15 @@ class PDFDocumentHelper
       public ui::TouchSelectionMenuClient,
       public content::TouchSelectionControllerClientManager::Observer {
  public:
+  class Observer : public base::CheckedObserver {
+   public:
+    // Invoked when the document load is completed successfully. Will not be
+    // invoked when the PDF is already loaded. Will not be invoked when the load
+    // fails. This is useful to wait for document metadata to be loaded, before
+    // calls to `GetPdfBytes`, `GetPageText` can be made.
+    virtual void OnDocumentLoadComplete() {}
+  };
+
   PDFDocumentHelper(const PDFDocumentHelper&) = delete;
   PDFDocumentHelper& operator=(const PDFDocumentHelper&) = delete;
 
@@ -81,6 +91,7 @@ class PDFDocumentHelper
 
   // pdf::mojom::PdfHost:
   void SetListener(mojo::PendingRemote<mojom::PdfListener> listener) override;
+  void OnDocumentLoadComplete() override;
   void SaveUrlAs(const GURL& url,
                  network::mojom::ReferrerPolicy policy) override;
   void UpdateContentRestrictions(int32_t content_restrictions) override;
@@ -93,6 +104,10 @@ class PDFDocumentHelper
   void OnSearchifyStarted() override;
 #endif
 
+  // Returns whether document is loaded, at which point, the other calls to
+  // document metadata such  as `GetPdfBytes`, `GetPageText` can be made.
+  bool IsDocumentLoadComplete() const { return is_document_load_complete_; }
+
   void GetPdfBytes(uint32_t size_limit,
                    pdf::mojom::PdfListener::GetPdfBytesCallback callback);
 
@@ -101,6 +116,9 @@ class PDFDocumentHelper
   void GetMostVisiblePageIndex(
       pdf::mojom::PdfListener::GetMostVisiblePageIndexCallback callback);
 
+  void AddObserver(Observer* observer);
+  void RemoveObserver(Observer* observer);
+
  private:
   friend class content::DocumentUserData<PDFDocumentHelper>;
 
@@ -130,8 +148,12 @@ class PDFDocumentHelper
   int32_t selection_right_height_ = 0;
   bool has_selection_ = false;
 
+  bool is_document_load_complete_ = false;
+
   mojo::Remote<mojom::PdfListener> remote_pdf_client_;
 
+  base::ObserverList<Observer> observers_;
+
   DOCUMENT_USER_DATA_KEY_DECL();
 };
 

components/pdf/browser/pdf_document_helper_browsertest.cc

@@ -55,6 +55,11 @@ class FakePdfListener : public pdf::mojom::PdfListener {
               (override));
 };
 
+class FakePdfLoadObserver : public PDFDocumentHelper::Observer {
+ public:
+  MOCK_METHOD(void, OnDocumentLoadComplete, (), (override));
+};
+
 class TestPDFDocumentHelperClient : public PDFDocumentHelperClient {
  public:
   TestPDFDocumentHelperClient() = default;
@@ -266,6 +271,19 @@ IN_PROC_BROWSER_TEST_P(PDFDocumentHelperTest, DefaultImplementation) {
   EXPECT_TRUE(pdf_document_helper()->GetSelectedText().empty());
 }
 
+IN_PROC_BROWSER_TEST_P(PDFDocumentHelperTest, DocumentLoadComplete) {
+  NiceMock<FakePdfLoadObserver> listener;
+  EXPECT_FALSE(pdf_document_helper()->IsDocumentLoadComplete());
+  pdf_document_helper()->AddObserver(&listener);
+  EXPECT_CALL(listener, OnDocumentLoadComplete);
+  pdf_document_helper()->OnDocumentLoadComplete();
+  EXPECT_TRUE(pdf_document_helper()->IsDocumentLoadComplete());
+
+  // Subsequent load complete should not trigger listener calls.
+  EXPECT_CALL(listener, OnDocumentLoadComplete).Times(0);
+  pdf_document_helper()->OnDocumentLoadComplete();
+}
+
 // TODO(crbug.com/40268279): Stop testing both modes after OOPIF PDF viewer
 // launches.
 INSTANTIATE_FEATURE_OVERRIDE_TEST_SUITE(PDFDocumentHelperTest);

pdf/mojom/pdf.mojom

@@ -48,6 +48,10 @@ interface PdfListener {
 interface PdfHost {
   SetListener(pending_remote<PdfListener> client);
 
+  // Invoked when document load is completed successfully. Will not be invoked
+  // if the load fails.
+  OnDocumentLoadComplete();
+
   // Updates the content restrictions, i.e. to disable print/copy.
   UpdateContentRestrictions(int32 restrictions);
 

pdf/pdf_view_web_plugin.cc

@@ -1309,6 +1309,7 @@ void PdfViewWebPlugin::DocumentLoadComplete() {
     return;
 
   DidStopLoading();
+  pdf_host_->OnDocumentLoadComplete();
   pdf_host_->UpdateContentRestrictions(GetContentRestrictions());
 }
 

pdf/pdf_view_web_plugin_unittest.cc

@@ -161,8 +161,9 @@ MATCHER(SearchStringResultEq, "") {
 }
 
 MATCHER_P(IsExpectedImeKeyEvent, expected_text, "") {
-  if (arg.GetType() != blink::WebInputEvent::Type::kChar)
+  if (arg.GetType() != blink::WebInputEvent::Type::kChar) {
     return false;
+  }
 
   const auto& event = static_cast<const blink::WebKeyboardEvent&>(arg);
   return event.GetModifiers() == blink::WebInputEvent::kNoModifiers &&
@@ -365,6 +366,7 @@ class FakePdfHost : public pdf::mojom::PdfHost {
               SetListener,
               (mojo::PendingRemote<pdf::mojom::PdfListener>),
               (override));
+  MOCK_METHOD(void, OnDocumentLoadComplete, (), (override));
   MOCK_METHOD(void, UpdateContentRestrictions, (int32_t), (override));
   MOCK_METHOD(void,
               SaveUrlAs,
@@ -739,6 +741,7 @@ TEST_F(PdfViewWebPluginFullFrameTest, DocumentLoadComplete) {
                                                    kContentRestrictionPaste |
                                                    kContentRestrictionCut |
                                                    kContentRestrictionCopy));
+  EXPECT_CALL(pdf_host_, OnDocumentLoadComplete);
   plugin_->DocumentLoadComplete();
 
   EXPECT_EQ(PdfViewWebPlugin::DocumentLoadState::kComplete,
@@ -751,6 +754,7 @@ TEST_F(PdfViewWebPluginTest, DocumentLoadFailed) {
 
   EXPECT_CALL(*client_ptr_, RecordComputedAction("PDF.LoadFailure"));
   EXPECT_CALL(*client_ptr_, DidStopLoading).Times(0);
+  EXPECT_CALL(pdf_host_, OnDocumentLoadComplete).Times(0);
   plugin_->DocumentLoadFailed();
 
   EXPECT_EQ(PdfViewWebPlugin::DocumentLoadState::kFailed,