Extract one page after the other and not all pages at once

diff --git a/src/core.js b/src/core.js
index 4bd2cb23443104ec9ab748ad8411d20c34bcedce..38f264bf59a21d21af5eb192d70b2faf1931b68c 100644 (file)
--- a/src/core.js
+++ b/src/core.js
@@ -698,6 +698,9 @@ var PDFDoc = (function PDFDocClosure() {
     this.fontsLoading = {};
     this.workerReadyPromise = new Promise('workerReady');
 
+    this.pageText = [];
+    this.startedTextExtraction = false;
+
     // If worker support isn't disabled explicit and the browser has worker
     // support, create a new web worker and test if it/the browser fullfills
     // all requirements to run parts of pdf.js in a web worker.
@@ -769,7 +772,6 @@ var PDFDoc = (function PDFDocClosure() {
       WorkerMessageHandler.setup(messageHandler);
     },
 
-
     setupMessageHandler: function PDFDoc_setupMessageHandler(messageHandler) {
       this.messageHandler = messageHandler;
 
@@ -825,9 +827,18 @@ var PDFDoc = (function PDFDocClosure() {
       }, this);
 
       messageHandler.on('text_extracted', function pdfTextExtracted(data) {
-        var index = data[0];
+        var pageNum = data[0];
+        var content = data[1];
+        if (pageNum !== this.pageText.length + 1)
+          error('pdfTextExtracted: pageIdx and pageText length got to fit');
+
+        this.pageText.push(content);
+
         if (this.textExtracted)
-          this.textExtracted(index);
+          this.textExtracted(pageNum, content);
+
+        if (pageNum < this.numPages)
+          this.extractTextPage(pageNum + 1);
       }, this);
 
       messageHandler.on('jpeg_decode', function(data, promise) {
@@ -895,9 +906,19 @@ var PDFDoc = (function PDFDocClosure() {
       return (this.pageCache[n] = page);
     },
 
+    extractTextPage: function PDFDoc_extractTextPage(pageNum) {
+      this.messageHandler.send('extract_text', pageNum);
+    },
+
     extractText: function PDFDoc_extractText() {
+      if (this.startedTextExtraction)
+        return;
+
+      this.startedTextExtraction = true;
+
       this.workerReadyPromise.then(function pdfDocStartRenderingThen() {
-        this.messageHandler.send('extract_text');
+        // Start the text extraction process.
+        this.extractTextPage(1);
       }.bind(this));
     },
 

diff --git a/src/worker.js b/src/worker.js
index b75fc66e87a93eaac152c11d78aa54ff4e259983..b7679bdbe7add6f65a5050c026cee90612b197ea 100644 (file)
--- a/src/worker.js
+++ b/src/worker.js
@@ -94,7 +94,6 @@ var WorkerMessageHandler = {
     handler.on('page_request', function wphSetupPageRequest(pageNum) {
       pageNum = parseInt(pageNum);
 
-
       // The following code does quite the same as
       // Page.prototype.startRendering, but stops at one point and sends the
       // result back to the main thread.
@@ -156,37 +155,20 @@ var WorkerMessageHandler = {
       });
     }, this);
 
-    handler.on('extract_text', function wphExtractText() {
-      var numPages = pdfModel.numPages;
-      var index = [];
+    handler.on('extract_text', function wphExtractText(pageNum) {
       var start = Date.now();
 
-      function indexPage(pageNum) {
-        if (pageNum > numPages) {
-          console.log('text indexing: time=%dms', Date.now() - start);
-
-          handler.send('text_extracted', [index]);
-          return;
-        }
-
-        var textContent = '';
-        // try {
-          var page = pdfModel.getPage(pageNum);
-          textContent = page.extractTextContent();
-        // } catch (e) {
-        //   // Skip errored pages
-        // }
-
-        index.push(textContent);
-
-        // processing one page, interrupting thread to process
-        // other requests
-        setTimeout(function extractTextNextPage() {
-          indexPage(pageNum + 1);
-        }, 0);
+      var textContent = '';
+      try {
+        var page = pdfModel.getPage(pageNum);
+        textContent = page.extractTextContent();
+      } catch (e) {
+        // Skip errored pages
       }
 
-      indexPage(1);
+      console.log('text indexing: page=%d - time=%dms',
+                      pageNum, Date.now() - start);
+      handler.send('text_extracted', [pageNum, textContent]);
     });
   }
 };

diff --git a/web/viewer.js b/web/viewer.js
index 91639d9ee084f535a4180c48038b6801bc1b25d2..c827b5bce0ad4bb12805f467182c1c8be6b01034 100644 (file)
--- a/web/viewer.js
+++ b/web/viewer.js
@@ -491,7 +491,7 @@ var PDFView = {
 
     var pdf;
     try {
-      pdf = new PDFJS.PDFDoc(data);
+      this.pdfDoc = pdf = new PDFJS.PDFDoc(data);
     } catch (e) {
       this.error('An error occurred while reading the PDF.', e);
     }
@@ -576,22 +576,18 @@ var PDFView = {
 
     if (pdfTitle)
       document.title = pdfTitle + ' - ' + document.title;
-
-    // loosing pdf reference here, starting text indexing in 500ms
-    setTimeout((function loadStartTextExtraction() {
-      this.startTextExtraction(pdf);
-    }).bind(this), 500);
-    delete PDFView.extractedText;
   },
 
   startTextExtraction: function pdfViewStartTextExtraction(pdf) {
     var searchResults = document.getElementById('searchResults');
     searchResults.textContent = '';
 
-    pdf.textExtracted = function pdfTextExtracted(index) {
-      PDFView.extractedText = index;
-    };
+    pdf.textExtracted = (function pdfTextExtracted(pageIdx, content) {
+      this.search();
+    }).bind(this);
     pdf.extractText();
+
+    this.pdfDoc = pdf;
   },
 
   search: function pdfViewStartSearch() {
@@ -604,21 +600,19 @@ var PDFView = {
     }
 
     var searchResults = document.getElementById('searchResults');
-    if (!('extractedText' in PDFView)) {
-      // not indexed yet, repeat in 1 second
-      searchResults.textContent = 'Searching...';
-      setTimeout(this.search.bind(this), 1000);
-      return;
-    }
 
     var searchTermsInput = document.getElementById('searchTermsInput');
     searchResults.removeAttribute('hidden');
     searchResults.textContent = '';
 
     var terms = searchTermsInput.value;
+
+    if (!terms)
+      return;
+
     // simple search: removing spaces and hyphens, then scanning every
     terms = terms.replace(/\s-/g, '').toLowerCase();
-    var index = PDFView.extractedText;
+    var index = PDFView.pdfDoc.pageText;
     var pageFound = false;
     for (var i = 0, ii = index.length; i < ii; i++) {
       var pageText = index[i].replace(/\s-/g, '').toLowerCase();
@@ -708,6 +702,9 @@ var PDFView = {
 
       var searchTermsInput = document.getElementById('searchTermsInput');
       searchTermsInput.focus();
+
+      // Start text extraction as soon as the search gets displayed.
+      this.pdfDoc.extractText();
     } else {
       searchScrollView.setAttribute('hidden', 'true');
       searchSwitchButton.removeAttribute('data-selected');