pdf-reader 2.5.0 → 2.9.0

data/lib/pdf/reader/encoding.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: true
 # frozen_string_literal: true
 
 ################################################################################
@@ -68,7 +69,7 @@ class PDF::Reader
     #
     #   [25, :A, :B]
     def differences=(diff)
-      raise ArgumentError, "diff must be an array" unless diff.kind_of?(Array)
+      PDF::Reader::Error.validate_type(diff, "diff", Array)
 
       @differences = {}
       byte = 0
@@ -208,7 +209,7 @@ class PDF::Reader
     def load_mapping(file)
       File.open(file, "r:BINARY") do |f|
         f.each do |l|
-          _m, single_byte, unicode = *l.match(/([0-9A-Za-z]+);([0-9A-F]{4})/)
+          _m, single_byte, unicode = *l.match(/\A([0-9A-Za-z]+);([0-9A-F]{4})/)
           @mapping["0x#{single_byte}".hex] = "0x#{unicode}".hex if single_byte
         end
       end

data/lib/pdf/reader/error.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 ################################################################################
@@ -33,19 +34,34 @@ class PDF::Reader
     def self.str_assert(lvalue, rvalue, chars=nil)
       raise MalformedPDFError, "PDF malformed, expected string but found #{lvalue.class} instead" if chars and !lvalue.kind_of?(String)
       lvalue = lvalue[0,chars] if chars
-      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found #{lvalue} instead"  if lvalue != rvalue
+      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found '#{lvalue}' instead"  if lvalue != rvalue
     end
     ################################################################################
     def self.str_assert_not(lvalue, rvalue, chars=nil)
       raise MalformedPDFError, "PDF malformed, expected string but found #{lvalue.class} instead" if chars and !lvalue.kind_of?(String)
       lvalue = lvalue[0,chars] if chars
-      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found #{lvalue} instead"  if lvalue == rvalue
+      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found '#{lvalue}' instead"  if lvalue == rvalue
     end
     ################################################################################
     def self.assert_equal(lvalue, rvalue)
-      raise MalformedPDFError, "PDF malformed, expected #{rvalue} but found #{lvalue} instead" if lvalue != rvalue
+      raise MalformedPDFError, "PDF malformed, expected '#{rvalue}' but found '#{lvalue}' instead" if lvalue != rvalue
     end
     ################################################################################
+    def self.validate_type(object, name, klass)
+      raise ArgumentError, "#{name} (#{object}) must be a #{klass}" unless object.is_a?(klass)
+    end
+    ################################################################################
+    def self.validate_type_as_malformed(object, name, klass)
+      raise MalformedPDFError, "#{name} (#{object}) must be a #{klass}" unless object.is_a?(klass)
+    end
+    ################################################################################
+    def self.validate_not_nil(object, name)
+      raise ArgumentError, "#{object} must not be nil" if object.nil?
+    end
+    ################################################################################
+    def self.validate_not_nil_as_malformed(object, name)
+      raise MalformedPDFError, "#{object} must not be nil" if object.nil?
+    end
   end
 
   ################################################################################

data/lib/pdf/reader/filter/ascii85.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 require 'ascii85'
@@ -7,6 +8,7 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the Ascii85 filter
     class Ascii85
+
       def initialize(options = {})
         @options = options
       end
@@ -17,7 +19,11 @@ class PDF::Reader
       #
       def filter(data)
         data = "<~#{data}" unless data.to_s[0,2] == "<~"
-        ::Ascii85::decode(data)
+        if defined?(::Ascii85Native)
+          ::Ascii85Native::decode(data)
+        else
+          ::Ascii85::decode(data)
+        end
       rescue Exception => e
         # Oops, there was a problem decoding the stream
         raise MalformedPDFError,

data/lib/pdf/reader/filter/ascii_hex.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 #
@@ -6,6 +7,7 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the AsciiHex stream filter
     class AsciiHex
+
       def initialize(options = {})
         @options = options
       end
@@ -16,9 +18,12 @@ class PDF::Reader
       def filter(data)
         data.chop! if data[-1,1] == ">"
         data = data[1,data.size] if data[0,1] == "<"
+
+        return "" if data.nil?
+
         data.gsub!(/[^A-Fa-f0-9]/,"")
         data << "0" if data.size % 2 == 1
-        data.scan(/.{2}/).map { |s| s.hex.chr }.join("")
+        data.scan(/.{2}/).flatten.map { |s| s.hex.chr }.join("")
       rescue Exception => e
         # Oops, there was a problem decoding the stream
         raise MalformedPDFError,

data/lib/pdf/reader/filter/depredict.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: true
 # frozen_string_literal: true
 
 class PDF::Reader
@@ -6,8 +7,9 @@ class PDF::Reader
     # some filter implementations support preprocessing of the  data to
     # improve compression
     class Depredict
+
       def initialize(options = {})
-        @options = options || {}
+        @options = options
       end
 
       ################################################################################
@@ -34,7 +36,7 @@ class PDF::Reader
       ################################################################################
       def tiff_depredict(data)
         data        = data.unpack("C*")
-        unfiltered  = []
+        unfiltered  = ''
         bpc         = @options[:BitsPerComponent] || 8
         pixel_bits  = bpc * @options[:Colors]
         pixel_bytes = pixel_bits / 8
@@ -51,11 +53,11 @@ class PDF::Reader
             left = index < pixel_bytes ? 0 : row_data[index - pixel_bytes]
             row_data[index] = (byte + left) % 256
           end
-          unfiltered += row_data
+          unfiltered += row_data.pack("C*")
           pos += line_len
         end
 
-        unfiltered.pack("C*")
+        unfiltered
       end
       ################################################################################
       def png_depredict(data)
@@ -67,7 +69,7 @@ class PDF::Reader
         scanline_length = (pixel_bytes * @options[:Columns]) + 1
         row = 0
         pixels = []
-        paeth, pa, pb, pc = nil
+        paeth, pa, pb, pc = 0, 0, 0, 0
         until data.empty? do
           row_data = data.slice! 0, scanline_length
           filter = row_data.shift
@@ -94,17 +96,17 @@ class PDF::Reader
               row_data[index] = (byte + ((left + upper)/2).floor) % 256
             end
           when 4 # Paeth
-            left = upper = upper_left = nil
+            left = upper = upper_left = 0
             row_data.each_with_index do |byte, index|
               col = index / pixel_bytes
 
-              left = index < pixel_bytes ? 0 : row_data[index - pixel_bytes]
+              left = index < pixel_bytes ? 0 : Integer(row_data[index - pixel_bytes])
               if row.zero?
                 upper = upper_left = 0
               else
-                upper = pixels[row-1][col][index % pixel_bytes]
+                upper = Integer(pixels[row-1][col][index % pixel_bytes])
                 upper_left = col.zero? ? 0 :
-                  pixels[row-1][col-1][index % pixel_bytes]
+                  Integer(pixels[row-1][col-1][index % pixel_bytes])
               end
 
               p = left + upper - upper_left

data/lib/pdf/reader/filter/flate.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 
@@ -8,6 +9,7 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the Flate (zlib) stream filter
     class Flate
+
       ZLIB_AUTO_DETECT_ZLIB_OR_GZIP = 47  # Zlib::MAX_WBITS + 32
       ZLIB_RAW_DEFLATE              = -15 # Zlib::MAX_WBITS * -1
 
@@ -32,7 +34,7 @@ class PDF::Reader
       def zlib_inflate(data)
         begin
           return Zlib::Inflate.new(ZLIB_AUTO_DETECT_ZLIB_OR_GZIP).inflate(data)
-        rescue Zlib::DataError
+        rescue Zlib::Error
           # by default, Ruby's Zlib assumes the data it's inflating
           # is RFC1951 deflated data, wrapped in a RFC1950 zlib container. If that
           # fails, swallow the exception and attempt to inflate the data as a raw
@@ -41,7 +43,7 @@ class PDF::Reader
 
         begin
           return Zlib::Inflate.new(ZLIB_RAW_DEFLATE).inflate(data)
-        rescue StandardError
+        rescue Zlib::Error
           # swallow this one too, so we can try some other fallback options
         end
 

data/lib/pdf/reader/filter/lzw.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 #
@@ -6,6 +7,7 @@ class PDF::Reader
   module Filter # :nodoc:
     # implementation of the LZW stream filter
     class Lzw
+
       def initialize(options = {})
         @options = options
       end

data/lib/pdf/reader/filter/null.rb

@@ -1,7 +1,7 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
-#
 class PDF::Reader
   module Filter # :nodoc:
     # implementation of the null stream filter

data/lib/pdf/reader/filter/run_length.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 #
@@ -6,6 +7,7 @@ class PDF::Reader # :nodoc:
   module Filter # :nodoc:
     # implementation of the run length stream filter
     class RunLength
+
       def initialize(options = {})
         @options = options
       end
@@ -20,19 +22,23 @@ class PDF::Reader # :nodoc:
           length = data.getbyte(pos)
           pos += 1
 
-          case
-          when length == 128
-            break
-          when length < 128
-            # When the length is < 128, we copy the following length+1 bytes
-            # literally.
-            out << data[pos, length + 1]
-            pos += length
-          else
-            # When the length is > 128, we copy the next byte (257 - length)
-            # times; i.e., "\xFA\x00" ([250, 0]) will expand to
-            # "\x00\x00\x00\x00\x00\x00\x00".
-            out << data[pos, 1] * (257 - length)
+          unless length.nil?
+            case
+              # nothing
+            when length == 128
+              break
+            when length < 128
+              # When the length is < 128, we copy the following length+1 bytes
+              # literally.
+              out << data[pos, length + 1]
+              pos += length
+            else
+              # When the length is > 128, we copy the next byte (257 - length)
+              # times; i.e., "\xFA\x00" ([250, 0]) will expand to
+              # "\x00\x00\x00\x00\x00\x00\x00".
+              previous_byte = data[pos, 1] || ""
+              out << previous_byte * (257 - length)
+            end
           end
 
           pos += 1

data/lib/pdf/reader/filter.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: strict
 # frozen_string_literal: true
 
 ################################################################################
@@ -41,7 +42,7 @@ class PDF::Reader
     # returned untouched. At this stage PDF::Reader has no need to decode images.
     #
     def self.with(name, options = {})
-      case name.to_sym
+      case name
       when :ASCII85Decode   then PDF::Reader::Filter::Ascii85.new(options)
       when :ASCIIHexDecode  then PDF::Reader::Filter::AsciiHex.new(options)
       when :CCITTFaxDecode  then PDF::Reader::Filter::Null.new(options)

data/lib/pdf/reader/font.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: true
 # frozen_string_literal: true
 
 ################################################################################
@@ -42,6 +43,7 @@ class PDF::Reader
       @tounicode = nil
 
       extract_base_info(obj)
+      extract_type3_info(obj)
       extract_descriptor(obj)
       extract_descendants(obj)
       @width_calc = build_width_calculator
@@ -72,8 +74,44 @@ class PDF::Reader
       @cached_widths[code_point] ||= @width_calc.glyph_width(code_point)
     end
 
+    # In most cases glyph width is converted into text space with a simple divide by 1000.
+    #
+    # However, Type3 fonts provide their own FontMatrix that's used for the transformation.
+    #
+    def glyph_width_in_text_space(code_point)
+      glyph_width_in_glyph_space = glyph_width(code_point)
+
+      if @subtype == :Type3
+        x1, y1 = font_matrix_transform(0,0)
+        x2, y2 = font_matrix_transform(glyph_width_in_glyph_space, 0)
+        (x2 - x1).abs.round(2)
+      else
+        glyph_width_in_glyph_space / 1000.0
+      end
+    end
+
     private
 
+    # Only valid for Type3 fonts
+    def font_matrix_transform(x, y)
+      return x, y if @font_matrix.nil?
+
+      matrix = TransformationMatrix.new(
+        @font_matrix[0], @font_matrix[1],
+        @font_matrix[2], @font_matrix[3],
+        @font_matrix[4], @font_matrix[5],
+      )
+
+      if x == 0 && y == 0
+        [matrix.e, matrix.f]
+      else
+        [
+          (matrix.a * x) + (matrix.c * y) + (matrix.e),
+          (matrix.b * x) + (matrix.d * y) + (matrix.f)
+        ]
+      end
+    end
+
     def default_encoding(font_name)
       case font_name.to_s
       when "Symbol" then
@@ -111,37 +149,55 @@ class PDF::Reader
       end
     end
 
-    def extract_base_info(obj)
-      @subtype  = @ohash.object(obj[:Subtype])
-      @basefont = @ohash.object(obj[:BaseFont])
-      if @ohash.object(obj[:Encoding])
-        @encoding = PDF::Reader::Encoding.new(@ohash.object(obj[:Encoding]))
+    def build_encoding(obj)
+      if obj[:Encoding].is_a?(Symbol)
+        # one of the standard encodings, referenced by name
+        # TODO pass in a standard shape, always a Hash
+        PDF::Reader::Encoding.new(obj[:Encoding])
+      elsif obj[:Encoding].is_a?(Hash) || obj[:Encoding].is_a?(PDF::Reader::Stream)
+        PDF::Reader::Encoding.new(obj[:Encoding])
+      elsif obj[:Encoding].nil?
+        default_encoding(@basefont)
       else
-        @encoding = default_encoding(@basefont)
+        raise MalformedPDFError, "Unexpected type for Encoding (#{obj[:Encoding].class})"
       end
-      @widths   = @ohash.object(obj[:Widths]) || []
-      @first_char = @ohash.object(obj[:FirstChar])
-      @last_char = @ohash.object(obj[:LastChar])
+    end
+
+    def extract_base_info(obj)
+      @subtype  = @ohash.deref_name(obj[:Subtype])
+      @basefont = @ohash.deref_name(obj[:BaseFont])
+      @encoding = build_encoding(obj)
+      @widths   = @ohash.deref_array_of_numbers(obj[:Widths]) || []
+      @first_char = @ohash.deref_integer(obj[:FirstChar])
+      @last_char = @ohash.deref_integer(obj[:LastChar])
 
       # CID Fonts are not required to have a W or DW entry, if they don't exist,
       # the default cid width = 1000, see Section 9.7.4.1 PDF 32000-1:2008 pp 269
-      @cid_widths         = @ohash.object(obj[:W])  || []
-      @cid_default_width  = @ohash.object(obj[:DW]) || 1000
+      @cid_widths         = @ohash.deref_array(obj[:W])  || []
+      @cid_default_width  = @ohash.deref_number(obj[:DW]) || 1000
 
       if obj[:ToUnicode]
         # ToUnicode is optional for Type1 and Type3
-        stream = @ohash.object(obj[:ToUnicode])
-        if stream.is_a?(PDF::Reader::Stream)
+        stream = @ohash.deref_stream(obj[:ToUnicode])
+        if stream
           @tounicode = PDF::Reader::CMap.new(stream.unfiltered_data)
         end
       end
     end
 
+    def extract_type3_info(obj)
+      if @subtype == :Type3
+        @font_matrix = @ohash.deref_array_of_numbers(obj[:FontMatrix]) || [
+          0.001, 0, 0, 0.001, 0, 0
+        ]
+      end
+    end
+
     def extract_descriptor(obj)
       if obj[:FontDescriptor]
         # create a font descriptor object if we can, in other words, unless this is
         # a CID Font
-        fd = @ohash.object(obj[:FontDescriptor])
+        fd = @ohash.deref_hash(obj[:FontDescriptor])
         @font_descriptor = PDF::Reader::FontDescriptor.new(@ohash, fd)
       else
         @font_descriptor = nil
@@ -153,9 +209,9 @@ class PDF::Reader
       # per PDF 32000-1:2008 pp. 280 :DescendentFonts is:
       # A one-element array specifying the CIDFont dictionary that is the
       # descendant of this Type 0 font.
-      descendants = @ohash.object(obj[:DescendantFonts])
+      descendants = @ohash.deref_array(obj[:DescendantFonts])
       @descendantfonts = descendants.map { |desc|
-        PDF::Reader::Font.new(@ohash, @ohash.object(desc))
+        PDF::Reader::Font.new(@ohash, @ohash.deref_hash(desc))
       }
     end
 

data/lib/pdf/reader/font_descriptor.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: true
 # frozen_string_literal: true
 
 require 'ttfunk'
@@ -14,22 +15,23 @@ class PDF::Reader
                 :x_height, :font_flags
 
     def initialize(ohash, fd_hash)
-      @ascent                = ohash.object(fd_hash[:Ascent])    || 0
-      @descent               = ohash.object(fd_hash[:Descent])   || 0
-      @missing_width         = ohash.object(fd_hash[:MissingWidth]) || 0
-      @font_bounding_box     = ohash.object(fd_hash[:FontBBox])  || [0,0,0,0]
-      @avg_width             = ohash.object(fd_hash[:AvgWidth])  || 0
-      @cap_height            = ohash.object(fd_hash[:CapHeight]) || 0
-      @font_flags            = ohash.object(fd_hash[:Flags])     || 0
-      @italic_angle          = ohash.object(fd_hash[:ItalicAngle])
-      @font_name             = ohash.object(fd_hash[:FontName]).to_s
-      @leading               = ohash.object(fd_hash[:Leading])   || 0
-      @max_width             = ohash.object(fd_hash[:MaxWidth])  || 0
-      @stem_v                = ohash.object(fd_hash[:StemV])
-      @x_height              = ohash.object(fd_hash[:XHeight])
-      @font_stretch          = ohash.object(fd_hash[:FontStretch]) || :Normal
-      @font_weight           = ohash.object(fd_hash[:FontWeight])  || 400
-      @font_family           = ohash.object(fd_hash[:FontFamily])
+      # TODO change these to typed derefs
+      @ascent                = ohash.deref_number(fd_hash[:Ascent])    || 0
+      @descent               = ohash.deref_number(fd_hash[:Descent])   || 0
+      @missing_width         = ohash.deref_number(fd_hash[:MissingWidth]) || 0
+      @font_bounding_box     = ohash.deref_array_of_numbers(fd_hash[:FontBBox])  || [0,0,0,0]
+      @avg_width             = ohash.deref_number(fd_hash[:AvgWidth])  || 0
+      @cap_height            = ohash.deref_number(fd_hash[:CapHeight]) || 0
+      @font_flags            = ohash.deref_integer(fd_hash[:Flags])     || 0
+      @italic_angle          = ohash.deref_number(fd_hash[:ItalicAngle])
+      @font_name             = ohash.deref_name(fd_hash[:FontName]).to_s
+      @leading               = ohash.deref_number(fd_hash[:Leading])   || 0
+      @max_width             = ohash.deref_number(fd_hash[:MaxWidth])  || 0
+      @stem_v                = ohash.deref_number(fd_hash[:StemV])
+      @x_height              = ohash.deref_number(fd_hash[:XHeight])
+      @font_stretch          = ohash.deref_name(fd_hash[:FontStretch]) || :Normal
+      @font_weight           = ohash.deref_number(fd_hash[:FontWeight])  || 400
+      @font_family           = ohash.deref_string(fd_hash[:FontFamily])
 
       # A FontDescriptor may have an embedded font program in FontFile
       # (Type 1 Font Program), FontFile2 (TrueType font program), or
@@ -39,7 +41,7 @@ class PDF::Reader
       # 2) CIDFontType0C: Type 0 Font Program in Compact Font Format
       # 3) OpenType:      OpenType Font Program
       # see Section 9.9, PDF 32000-1:2008, pp 288-292
-      @font_program_stream = ohash.object(fd_hash[:FontFile2])
+      @font_program_stream = ohash.deref_stream(fd_hash[:FontFile2])
       #TODO handle FontFile and FontFile3
 
       @is_ttf = true if @font_program_stream

data/lib/pdf/reader/form_xobject.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: true
 # frozen_string_literal: true
 
 require 'digest/md5'
@@ -14,15 +15,24 @@ module PDF
     # This behaves and looks much like a limited PDF::Reader::Page class.
     #
     class FormXObject
-      include ResourceMethods
+      extend Forwardable
 
       attr_reader :xobject
 
+      def_delegators :resources, :color_spaces
+      def_delegators :resources, :fonts
+      def_delegators :resources, :graphic_states
+      def_delegators :resources, :patterns
+      def_delegators :resources, :procedure_sets
+      def_delegators :resources, :properties
+      def_delegators :resources, :shadings
+      def_delegators :resources, :xobjects
+
       def initialize(page, xobject, options = {})
         @page    = page
         @objects = page.objects
         @cache   = options[:cache] || {}
-        @xobject = @objects.deref(xobject)
+        @xobject = @objects.deref_stream(xobject)
       end
 
       # return a hash of fonts used on this form.
@@ -33,9 +43,9 @@ module PDF
       # to most available metrics for each font.
       #
       def font_objects
-        raw_fonts = @objects.deref(resources[:Font] || {})
+        raw_fonts = @objects.deref_hash(fonts)
         ::Hash[raw_fonts.map { |label, font|
-          [label, PDF::Reader::Font.new(@objects, @objects.deref(font))]
+          [label, PDF::Reader::Font.new(@objects, @objects.deref_hash(font))]
         }]
       end
 
@@ -60,7 +70,7 @@ module PDF
       # Returns the resources that accompany this form.
       #
       def resources
-        @resources ||= @objects.deref(@xobject.hash[:Resources]) || {}
+        @resources ||= Resources.new(@objects, @objects.deref_hash(@xobject.hash[:Resources]) || {})
       end
 
       def callback(receivers, name, params=[])

data/lib/pdf/reader/glyph_hash.rb

@@ -1,4 +1,5 @@
 # coding: utf-8
+# typed: true
 # frozen_string_literal: true
 
 ################################################################################
@@ -103,19 +104,25 @@ class PDF::Reader
 
     # returns a hash that maps glyph names to unicode codepoints. The mapping is based on
     # a text file supplied by Adobe at:
-    # http://www.adobe.com/devnet/opentype/archives/glyphlist.txt
+    # https://github.com/adobe-type-tools/agl-aglfn
     def load_adobe_glyph_mapping
       keyed_by_name      = {}
       keyed_by_codepoint = {}
 
-      File.open(File.dirname(__FILE__) + "/glyphlist.txt", "r:BINARY") do |f|
-        f.each do |l|
-          _m, name, code = *l.match(/([0-9A-Za-z]+);([0-9A-F]{4})/)
-          if name && code
-            cp = "0x#{code}".hex
-            keyed_by_name[name.to_sym]   = cp
-            keyed_by_codepoint[cp]     ||= []
-            keyed_by_codepoint[cp]     << name.to_sym
+      paths = [
+        File.dirname(__FILE__) + "/glyphlist.txt",
+        File.dirname(__FILE__) + "/glyphlist-zapfdingbats.txt",
+      ]
+      paths.each do |path|
+        File.open(path, "r:BINARY") do |f|
+          f.each do |l|
+            _m, name, code = *l.match(/([0-9A-Za-z]+);([0-9A-F]{4})/)
+            if name && code
+              cp = "0x#{code}".hex
+              keyed_by_name[name.to_sym]   = cp
+              keyed_by_codepoint[cp]     ||= []
+              keyed_by_codepoint[cp]     << name.to_sym
+            end
           end
         end
       end