Text char codes extraction

diff --git a/src/core.js b/src/core.js
index 9f3e6b837e29ae8daacbcd376a153554807eb5e5..6a90357e8a42cde58f3aaf0ba2cd9fc41e29677f 100644 (file)
--- a/src/core.js
+++ b/src/core.js
@@ -200,10 +200,12 @@ var Page = (function PageClosure() {
       if (isArray(content)) {
         // fetching items
         var i, n = content.length;
+        var streams = [];
         for (i = 0; i < n; ++i)
-          content[i] = xref.fetchIfRef(content[i]);
-        content = new StreamsSequenceStream(content);
-      }
+          streams.push(xref.fetchIfRef(content[i]));
+        content = new StreamsSequenceStream(streams);
+      } else if (isStream(content))
+        content.pos = 0;
 
       var pe = this.pe = new PartialEvaluator(
                                 xref, handler, 'p' + this.pageNumber + '_');
@@ -212,6 +214,36 @@ var Page = (function PageClosure() {
                                            dependency));
     },
 
+    extractTextContent: function pageExtractPageContent() {
+      if ('textContent' in this) {
+        // text content was extracted
+        return this.textContent;
+      }
+
+      var handler = {
+        on: function () {},
+        send: function() {}
+      };
+
+      var xref = this.xref;
+      var content = xref.fetchIfRef(this.content);
+      var resources = xref.fetchIfRef(this.resources);
+      if (isArray(content)) {
+        // fetching items
+        var i, n = content.length;
+        var streams = [];
+        for (i = 0; i < n; ++i)
+          streams.push(xref.fetchIfRef(content[i]));
+        content = new StreamsSequenceStream(streams);
+      } else if (isStream(content))
+        content.pos = 0;
+
+      var pe = new PartialEvaluator(
+                     xref, handler, 'p' + this.pageNumber + '_');
+      var text = pe.getTextContent(content, resources);
+      return (this.textContent = text);
+    },
+
     ensureFonts: function pageEnsureFonts(fonts, callback) {
       // Convert the font names to the corresponding font obj.
       for (var i = 0, ii = fonts.length; i < ii; i++) {
@@ -614,6 +646,12 @@ var PDFDoc = (function PDFDocClosure() {
           throw data.error;
       }, this);
 
+      messageHandler.on('text_extracted', function pdfDocError(data) {
+        var index = data.index;
+        if (this.textExtracted)
+          this.textExtracted(index);
+      }, this);
+
       setTimeout(function pdfDocFontReadySetTimeout() {
         messageHandler.send('doc', this.data);
         this.workerReadyPromise.resolve(true);
@@ -643,6 +681,12 @@ var PDFDoc = (function PDFDocClosure() {
       return (this.pageCache[n] = page);
     },
 
+    extractText: function pdfDocExtractExtractText() {
+      this.workerReadyPromise.then(function pdfDocStartRenderingThen() {
+        this.messageHandler.send('extract_text');
+      }.bind(this));
+    },
+
     destroy: function pdfDocDestroy() {
       if (this.worker)
         this.worker.terminate();

diff --git a/src/evaluator.js b/src/evaluator.js
index a5ca627c5687db16600b1dd3d59201ea04f95b0d..588da5084ba1f6bc4cc8a253aaa02cbe1ff4da8c 100644 (file)
--- a/src/evaluator.js
+++ b/src/evaluator.js
@@ -144,7 +144,7 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
         fontRef = fontRef || fontRes.get(fontName);
         var font = xref.fetchIfRef(fontRef);
         assertWellFormed(isDict(font));
-        if (!font.translated) {
+        if (!font.loadedName) {
           font.translated = self.translateFont(font, xref, resources,
                                                dependency);
           if (font.translated) {
@@ -464,6 +464,65 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
       };
     },
 
+    getTextContent: function partialEvaluatorGetIRQueue(stream, resources) {
+
+      var self = this;
+      var xref = this.xref;
+
+      function handleSetFont(fontName, fontRef) {
+        var fontRes = resources.get('Font');
+
+        // TODO: TOASK: Is it possible to get here? If so, what does
+        // args[0].name should be like???
+        assert(fontRes, 'fontRes not available');
+
+        fontRes = xref.fetchIfRef(fontRes);
+        fontRef = fontRef || fontRes.get(fontName);
+        var font = xref.fetchIfRef(fontRef), tra;
+        assertWellFormed(isDict(font));
+        if (!font.translated) {
+          font.translated = self.translateFont(font, xref, resources);
+        }
+        return font;
+      }
+
+      resources = xref.fetchIfRef(resources) || new Dict();
+
+      var parser = new Parser(new Lexer(stream), false);
+      var res = resources;
+      var args = [], obj;
+
+      var text = '';
+      var font = null;
+      while (!isEOF(obj = parser.getObj())) {
+        if (isCmd(obj)) {
+          var cmd = obj.cmd;
+          switch (cmd) {
+            case 'Tf':
+              font = handleSetFont(args[0].name);
+              break;
+            case 'TJ':
+              var items = args[0];
+              for (var j = 0, jj = items.length; j < jj; j++) {
+                if (typeof items[j] === 'string')
+                  text += items[j];
+              }
+              break;
+            case 'Tj':
+              text += args[0];
+              break;
+          } // switch
+
+          args = [];
+        } else if (obj != null) {
+          assertWellFormed(args.length <= 33, 'Too many arguments');
+          args.push(obj);
+        }
+      }
+
+      return text;
+    },
+
     extractDataStructures: function
       partialEvaluatorExtractDataStructures(dict, baseDict,
                                             xref, properties) {
@@ -837,15 +896,18 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
 
       if (type.name === 'Type3') {
         properties.coded = true;
-        var charProcs = xref.fetchIfRef(dict.get('CharProcs'));
-        var fontResources = xref.fetchIfRef(dict.get('Resources')) || resources;
-        properties.resources = fontResources;
-        properties.charProcIRQueues = {};
-        for (var key in charProcs.map) {
-          var glyphStream = xref.fetchIfRef(charProcs.map[key]);
-          var queueObj = {};
-          properties.charProcIRQueues[key] =
-            this.getIRQueue(glyphStream, fontResources, queueObj, dependency);
+        // read char procs only if dependency is specified
+        if (dependency) {
+          var charProcs = xref.fetchIfRef(dict.get('CharProcs'));
+          var fontResources = xref.fetchIfRef(dict.get('Resources')) || resources;
+          properties.resources = fontResources;
+          properties.charProcIRQueues = {};
+          for (var key in charProcs.map) {
+            var glyphStream = xref.fetchIfRef(charProcs.map[key]);
+            var queueObj = {};
+            properties.charProcIRQueues[key] =
+              this.getIRQueue(glyphStream, fontResources, queueObj, dependency);
+          }
         }
       }
 

diff --git a/src/worker.js b/src/worker.js
index 52e631c92026b26cd318e98de1606d717bbfdcbe..c3992e54f7504a22f0d802c7d5dcd583042f989b 100644 (file)
--- a/src/worker.js
+++ b/src/worker.js
@@ -160,6 +160,28 @@ var WorkerMessageHandler = {
 
       handler.send('font_ready', [objId, obj]);
     });
+
+    handler.on('extract_text', function wphExtractText() {
+      var numPages = pdfDoc.numPages;
+      var index = [];
+      for (var i = 0; i < numPages; i++) {
+        var start = Date.now();
+
+        var textContent = '';
+        try {
+          var page = pdfDoc.getPage(i + 1);
+          textContent = page.extractTextContent();
+        } catch (e) {
+          // Skip errored pages
+        }
+
+        index.push(textContent);
+      }
+
+      console.log('text indexing=: time=%dms', Date.now() - start);
+
+      handler.send('text_extracted', { index: index });
+    });
   }
 };
 

diff --git a/web/viewer.js b/web/viewer.js
index bdcac09d53831b34968ee7a815266e54fe62ba83..465df5ab5cd470f075677562973b79eb48a23e04 100644 (file)
--- a/web/viewer.js
+++ b/web/viewer.js
@@ -309,6 +309,17 @@ var PDFView = {
     }
     else
       this.page = 1;
+
+    setTimeout((function loadStartTextExtraction() {
+      this.startTextExtraction(pdf);
+    }).bind(this), 500);
+  },
+
+  startTextExtraction: function(pdf) {
+    pdf.textExtracted = function pdfTextExtracted(index) {
+      console.log(index.join());
+    };
+    pdf.extractText();
   },
 
   setHash: function pdfViewSetHash(hash) {