[api-minor] Add support for PageLabels in the API

Snuffleupagus · Snuffleupagus · commit 85cf90643f39 · 2016-01-19T22:49:04.000+01:00
diff --git a/src/core/obj.js b/src/core/obj.js
@@ -263,6 +263,96 @@ var Catalog = (function CatalogClosure() {
       }
       return dest;
     },
+
+    get pageLabels() {
+      var obj = null;
+      try {
+        obj = this.readPageLabels();
+      } catch (ex) {
+        if (ex instanceof MissingDataException) {
+          throw ex;
+        }
+        warn('Unable to read page labels.');
+      }
+      return shadow(this, 'pageLabels', obj);
+    },
+    readPageLabels: function Catalog_readPageLabels() {
+      var obj = this.catDict.getRaw('PageLabels');
+      if (!obj) {
+        return null;
+      }
+      var pageLabels = new Array(this.numPages);
+      var style = null;
+      var prefix = '';
+      var start = 1;
+
+      var numberTree = new NumberTree(obj, this.xref);
+      var nums = numberTree.getAll();
+      var currentLabel = '', currentIndex = 1;
+
+      for (var i = 0, ii = this.numPages; i < ii; i++) {
+        if (nums.hasOwnProperty(i)) {
+          var labelDict = nums[i];
+          assert(isDict(labelDict), 'The PageLabel is not a dictionary.');
+
+          var type = labelDict.get('Type');
+          assert(!type || (isName(type) && type.name === 'PageLabel'),
+                 'Invalid type in PageLabel dictionary.');
+
+          var s = labelDict.get('S');
+          assert(!s || isName(s), 'Invalid style in PageLabel dictionary.');
+          style = (s ? s.name : null);
+
+          prefix = labelDict.get('P') || '';
+          assert(isString(prefix), 'Invalid prefix in PageLabel dictionary.');
+
+          start = labelDict.get('St') || 1;
+          assert(isInt(start), 'Invalid start in PageLabel dictionary.');
+          currentIndex = start;
+        }
+
+        switch (style) {
+          case 'D':
+            currentLabel = currentIndex;
+            break;
+          case 'R':
+          case 'r':
+            currentLabel = Util.toRoman(currentIndex, style === 'r');
+            break;
+          case 'A':
+          case 'a':
+            var LIMIT = 26; // Use only the characters A--Z, or a--z.
+            var A_UPPER_CASE = 0x41, A_LOWER_CASE = 0x61;
+
+            var baseCharCode = (style === 'a' ? A_LOWER_CASE : A_UPPER_CASE);
+            var letterIndex = currentIndex - 1;
+            var character = String.fromCharCode(baseCharCode +
+                                                (letterIndex % LIMIT));
+            var charBuf = [];
+            for (var j = 0, jj = (letterIndex / LIMIT) | 0; j <= jj; j++) {
+              charBuf.push(character);
+            }
+            currentLabel = charBuf.join('');
+            break;
+          default:
+            assert(!style,
+                   'Invalid style "' + style + '" in PageLabel dictionary.');
+        }
+        pageLabels[i] = prefix + currentLabel;
+
+        currentLabel = '';
+        currentIndex++;
+      }
+
+      // Ignore PageLabels if they correspond to standard page numbering.
+      for (i = 0, ii = this.numPages; i < ii; i++) {
+        if (pageLabels[i] !== (i + 1).toString()) {
+          break;
+        }
+      }
+      return (i === ii ? [] : pageLabels);
+    },
+
     get attachments() {
       var xref = this.xref;
       var attachments = null, nameTreeRef;
diff --git a/src/core/worker.js b/src/core/worker.js
@@ -451,6 +451,12 @@ var WorkerMessageHandler = PDFJS.WorkerMessageHandler = {
       }
     );
 
+    handler.on('GetPageLabels',
+      function wphSetupGetPageLabels(data) {
+        return pdfManager.ensureCatalog('pageLabels');
+      }
+    );
+
     handler.on('GetAttachments',
       function wphSetupGetAttachments(data) {
         return pdfManager.ensureCatalog('attachments');
diff --git a/src/display/api.js b/src/display/api.js
@@ -690,6 +690,16 @@ var PDFDocumentProxy = (function PDFDocumentProxyClosure() {
     getDestination: function PDFDocumentProxy_getDestination(id) {
       return this.transport.getDestination(id);
     },
+    /**
+     * @return {Promise} A promise that is resolved with: an Array containing
+     *   the pageLabels that correspond to the pageIndexes; or null, when no
+     *   pageLabels are present in the PDF file.
+     *   NOTE: If the pageLabels are all identical to standard page numbering,
+     *         i.e. [1, 2, 3, ...], the promise is resolved with an empty Array.
+     */
+    getPageLabels: function PDFDocumentProxy_getPageLabels() {
+      return this.transport.getPageLabels();
+    },
     /**
      * @return {Promise} A promise that is resolved with a lookup table for
      * mapping named attachments to their content.
@@ -1804,6 +1814,10 @@ var WorkerTransport = (function WorkerTransportClosure() {
       return this.messageHandler.sendWithPromise('GetDestination', { id: id });
     },
 
+    getPageLabels: function WorkerTransport_getPageLabels() {
+      return this.messageHandler.sendWithPromise('GetPageLabels', null);
+    },
+
     getAttachments: function WorkerTransport_getAttachments() {
       return this.messageHandler.sendWithPromise('GetAttachments', null);
     },
diff --git a/src/shared/util.js b/src/shared/util.js
@@ -808,6 +808,42 @@ var Util = PDFJS.Util = (function UtilClosure() {
     return num < 0 ? -1 : 1;
   };
 
+  var ROMAN_NUMBER_MAP = [
+    '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM',
+    '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC',
+    '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX'
+  ];
+  /**
+   * Converts positive integers to (upper case) Roman numerals.
+   * @param {integer} number - The number that should be converted.
+   * @param {boolean} lowerCase - Indicates if the result should be converted
+   *   to lower case letters. The default is false.
+   * @return {string} The resulting Roman number.
+   */
+  Util.toRoman = function Util_toRoman(number, lowerCase) {
+    assert(isInt(number) && number > 0,
+           'The number should be a positive integer.');
+    var pos, romanBuf = [];
+    // Thousands
+    while (number >= 1000) {
+      number -= 1000;
+      romanBuf.push('M');
+    }
+    // Hundreds
+    pos = (number / 100) | 0;
+    number %= 100;
+    romanBuf.push(ROMAN_NUMBER_MAP[pos]);
+    // Tens
+    pos = (number / 10) | 0;
+    number %= 10;
+    romanBuf.push(ROMAN_NUMBER_MAP[10 + pos]);
+    // Ones
+    romanBuf.push(ROMAN_NUMBER_MAP[20 + number]);
+
+    var romanStr = romanBuf.join('');
+    return (lowerCase ? romanStr.toLowerCase() : romanStr);
+  };
+
   Util.appendToArray = function Util_appendToArray(arr1, arr2) {
     Array.prototype.push.apply(arr1, arr2);
   };
diff --git a/test/pdfs/.gitignore b/test/pdfs/.gitignore
@@ -24,6 +24,7 @@
 !arial_unicode_ab_cidfont.pdf
 !arial_unicode_en_cidfont.pdf
 !asciihexdecode.pdf
+!bug793632.pdf
 !bug1020858.pdf
 !bug1050040.pdf
 !bug1200096.pdf
diff --git a/test/pdfs/bug793632.pdf b/test/pdfs/bug793632.pdf
diff --git a/test/unit/api_spec.js b/test/unit/api_spec.js
@@ -329,6 +329,36 @@ describe('api', function() {
         expect(data).toEqual(null);
       });
     });
+    it('gets non-existent page labels', function () {
+      var promise = doc.getPageLabels();
+      waitsForPromiseResolved(promise, function (data) {
+        expect(data).toEqual(null);
+      });
+    });
+    it('gets page labels', function () {
+      // PageLabels with Roman/Arabic numerals.
+      var url0 = combineUrl(window.location.href, '../pdfs/bug793632.pdf');
+      var promise0 = PDFJS.getDocument(url0).promise.then(function (pdfDoc) {
+        return pdfDoc.getPageLabels();
+      });
+      // PageLabels with only a label prefix.
+      var url1 = combineUrl(window.location.href, '../pdfs/issue1453.pdf');
+      var promise1 = PDFJS.getDocument(url1).promise.then(function (pdfDoc) {
+        return pdfDoc.getPageLabels();
+      });
+      // PageLabels identical to standard page numbering.
+      var url2 = combineUrl(window.location.href, '../pdfs/rotation.pdf');
+      var promise2 = PDFJS.getDocument(url2).promise.then(function (pdfDoc) {
+        return pdfDoc.getPageLabels();
+      });
+
+      waitsForPromiseResolved(Promise.all([promise0, promise1, promise2]),
+          function (pageLabels) {
+        expect(pageLabels[0]).toEqual(['i', 'ii', 'iii', '1']);
+        expect(pageLabels[1]).toEqual(['Front Page1']);
+        expect(pageLabels[2]).toEqual([]);
+      });
+    });
     it('gets attachments', function() {
       var promise = doc.getAttachments();
       waitsForPromiseResolved(promise, function (data) {
