pdfrb 0.6.0 → 0.7.10

data/data/pdfrb/layout/hyphenation_en.txt

@@ -0,0 +1,408 @@
+.ab1
+ab2c
+a2ble
+.a1
+a2b
+ac1c
+ad2d
+ag2
+al1s
+an3c
+a3p2
+ar1d
+ar1t
+as1c
+as1s
+at1c
+au1
+a1v
+a3y
+be2s4
+1ba
+2b1
+be2
+bi2z
+1b2
+bo2
+bp1
+bs1
+b5t2
+b5u2
+1ca
+c2h2
+ci2
+1c2
+ck1
+co3
+con1
+cp1
+cs2
+cu2
+cy1
+1da
+dd2
+de2
+1d2
+d4f3
+d5i6
+dg2
+d1i4
+d5le
+d3o2
+ds4
+1ea
+ea2
+e1b
+e2c1
+e1d
+e4ed
+e1f
+e1g
+e1h
+ei2
+e3j
+e1k
+e1l
+e2m1
+e1n
+eo2
+ep2
+e1q
+e1r
+e4s1
+e1t
+eu2
+e1v
+e1w
+e1x
+ey1
+e1z
+1fa
+fb1
+fe2
+1f2
+ff2
+4f3i
+fi2
+4f5l
+1f2
+fo1
+f4p1
+fr2
+1ga
+g2a
+1g2
+gd1
+ge3
+g2gel
+g4g3
+gi2
+1g2
+go1
+g4p1
+g3ra
+1ha
+he2
+1h2
+hi2
+1h2
+ho2
+hp1
+hs2
+hu2
+hy1
+4i1a
+2i1b
+i1c4
+2i1d
+i1e
+4if
+i1g2
+4ih
+ii2
+i1j4
+i1k2
+i1l2
+2im
+2in
+4i1o
+i1p4
+i1q
+ir2
+4is
+i1t4
+i1u
+2iv
+i1w
+iw4
+ix2
+i1y
+iz2
+1ja
+1je
+1ji
+1jo
+1ju
+1ka
+kb1
+1kc
+1kd
+ke2
+1k2
+kf1
+kg1
+k2h
+1ki
+1kj
+kk1
+1kl
+km1
+1kn
+ko1
+kp1
+1kr
+ks2
+1ku
+1kv
+1kw
+1la
+lb1
+lc1
+ld2
+le2
+1l2
+lf1
+lg2
+1li
+1lj
+1lk1
+1ll2
+1lm1
+1ln1
+1lo
+1lp1
+1lq
+1lr
+1ls
+1lt1
+1lu
+1lv
+1lw
+1lx
+1ly
+1lz
+4ma
+1m2b1
+me2
+m1f1
+m4g3
+mi2
+4m1l
+1mo
+1mp
+m2s1
+mu2
+my1
+1na
+nb1
+nc2
+nd2
+ne2
+1n2
+nf1
+n2g3
+1ni
+1nj
+1nk1
+n1l2
+1nm1
+1nn
+1no
+n1p1
+n1q
+1nr
+ns2
+nt1
+1nu
+1nv
+nw1
+ny1
+1ob
+oc2
+o1d4
+o1f
+og2
+oh2
+oi2
+o1j
+o1k
+ol2
+om1
+on4
+op2
+o1q
+o1r
+o4s3
+o1t
+ou2
+o1v
+ow2
+ox1
+o1y
+oz2
+1pa
+1p2b
+pe2
+1p2
+pf1
+4ph
+p2h2
+4p1l
+1p2
+po1
+p2p
+p1p1
+p4p3
+ps2
+p1t2
+4q
+1qu
+1ra
+1r2b1
+r1c
+rd2
+re2
+r1d4
+1re
+r1f4
+r1g
+r1h4
+ri2
+4r1l
+1rm1
+1rn1
+ro2
+r1p4
+r1r4
+1rs2
+rt1
+r1t4
+r1u
+ru2
+r1v
+r1w4
+1sa
+1s2b1
+s1c
+s1d
+se2
+s1e
+s1f
+s1g
+sh2
+si2
+1s2
+s1k1
+sl1
+s1l1
+sm2
+s1m1
+sn1
+1so
+s1p
+s1q
+1sr1
+s1s1
+s1t1
+st1
+su2
+s1w
+sy2
+4ta
+1t2b
+tc2
+t1d4
+te2
+t1e
+t1f
+t1g4
+th1
+1ti
+t1j
+t1k
+1tl1
+1tn1
+to2
+t1p
+t1q
+t1r4
+t1s1
+t1t1
+tu2
+t1u
+t1v
+tw1
+1ua
+ub1
+u1c4
+u1d
+4uf
+u1g2
+4uh
+ui2
+u1j4
+u1k2
+u1l2
+4un
+u1p4
+u1q
+ur2
+4us
+u1t4
+u1u
+4uv
+u1w
+ux1
+u1y
+uz2
+1va
+1v2b1
+vc1
+ve2
+1v2
+v1f1
+vi2
+v1l
+v1o
+v1p1
+1vs2
+vu1
+1wa
+1w2b
+we2
+wh1
+1wi
+1w2
+wl1
+wn1
+1wo
+1w2
+1wr
+1wu
+1xa
+1xe
+1xi
+1xo
+1xu
+1ya
+1ye
+yi2
+1yo
+1yu
+4za
+1z2b1
+ze2
+1z2
+zi2
+zo2
+zu2

data/lib/pdfrb/color/default_profile.rb

@@ -0,0 +1,213 @@
+# frozen_string_literal: true
+
+require "digest"
+
+module Pdfrb
+  module Color
+    # Default sRGB ICC profile accessor. Builds a usable sRGB v4
+    # ICC profile (header + tag table + tagged elements) that real
+    # PDF readers can consume for ICCBased color spaces. The profile
+    # includes:
+    #
+    #   * desc  — profile description (sRGB by pdfrb)
+    #   * wtpt  — D50 media white point
+    #   * rXYZ, gXYZ, bXYZ — sRGB colorant matrix (D50-adapted)
+    #   * rTRC, gTRC, bTRC — gamma 2.2 tone reproduction curves
+    #
+    # Tagged-element types follow ICC.1:2022 §10 (multiLocalizedUnicodeType
+    # for desc, XYZType for colorants, curveType for TRCs).
+    module DefaultProfile
+      SRGB_SIGNATURE = "acsp".b
+      ICC_VERSION_V4 = [0x04, 0x40, 0x00, 0x00].freeze
+
+      module_function
+
+      # Returns the sRGB ICC profile bytes (~ 600 bytes). Includes
+      # the structural header + tag table + tagged elements needed
+      # for a real ICCBased color space.
+      def srgb_bytes
+        tags = build_srgb_tags
+        tag_table = build_tag_table(tags)
+        header = build_srgb_header(tag_table.bytesize + tags_total_size(tags))
+        profile = header + tag_table
+        tags.each_value { |tag_bytes| profile << tag_bytes }
+        # Compute MD5 ProfileID over the on-disk profile (sans the
+        # existing ID slot, which we'll patch in). ICC.1:2022 §7.1.
+        profile = patch_profile_id(profile)
+        profile.force_encoding(Encoding::BINARY)
+      end
+
+      # Build an ICCBased color space array referencing a new stream
+      # containing the sRGB profile bytes. Returns
+      # [:ICCBased, ref] suitable for /ColorSpace resources.
+      def srgb_color_space(document)
+        bytes = srgb_bytes
+        stream = document.add(
+          { N: 3, Length: bytes.bytesize },
+          type: Pdfrb::Model::Cos::Stream
+        )
+        stream.stream = bytes
+        [:ICCBased, Pdfrb::Model::Reference.new(stream.oid, stream.gen)]
+      end
+
+      # ---- Header construction ----
+
+      def build_srgb_header(tag_section_size)
+        header = +""
+        header << ([0] * 4).pack("C*")                # Profile size placeholder
+        header << ([0] * 4).pack("C*")                # Preferred CMM
+        header << ICC_VERSION_V4.pack("C*")           # Version 4.4
+        header << "mntr"                              # Device class
+        header << "RGB "                              # Color space
+        header << "XYZ "                              # PCS
+        header << date_bytes                          # Date
+        header << SRGB_SIGNATURE # 'acsp'
+        header << "APPL"                              # Primary platform
+        header << ([0] * 4).pack("C*")                # Profile flags
+        header << ([0] * 4).pack("C*")                # Device manufacturer
+        header << ([0] * 4).pack("C*")                # Device model
+        header << ([0] * 8).pack("C*")                # Device attributes
+        header << ([0] * 4).pack("C*")                # Rendering intent (perceptual)
+        header << d50_illuminant_bytes                # PCS illuminant (D50)
+        header << ([0] * 4).pack("C*")                # Profile creator
+        header << ([0] * 16).pack("C*")               # Profile ID (MD5, patched later)
+        header << ([0] * 28).pack("C*")               # Reserved
+
+        total_size = header.bytesize + tag_section_size
+        header[0, 4] = [total_size].pack("N")
+        header
+      end
+
+      def date_bytes
+        [0x07, 0xE8, 0x00, 0x01, 0x00, 0x01,
+         0x00, 0x00, 0x00, 0x00, 0x00, 0x00].pack("C*")
+      end
+
+      def d50_illuminant_bytes
+        # D50 X=0.9642, Y=1.0, Z=0.8249 in s15Fixed16.
+        [0x00, 0x00, 0xF6, 0xD6, 0x00, 0x01, 0x00, 0x00,
+         0x00, 0x00, 0xD3, 0x2D].pack("C*")
+      end
+
+      # ---- Tag table + tagged elements ----
+
+      # Returns a Hash of { tag_signature => tag_bytes } in the
+      # order they should appear after the tag table.
+      def build_srgb_tags
+        {
+          "desc" => desc_tag_bytes("sRGB by pdfrb"),
+          "wtpt" => xyz_tag_bytes(d50_xyz),
+          "rXYZ" => xyz_tag_bytes(srgb_r_xyz),
+          "gXYZ" => xyz_tag_bytes(srgb_g_xyz),
+          "bXYZ" => xyz_tag_bytes(srgb_b_xyz),
+          "rTRC" => curve_tag_bytes,
+          "gTRC" => curve_tag_bytes,
+          "bTRC" => curve_tag_bytes,
+        }
+      end
+
+      def build_tag_table(tags)
+        out = +""
+        count = tags.length
+        out << [count].pack("N")
+
+        # Compute offsets: tag table is 4 + count*12 bytes; each tag
+        # element is appended after the table in order.
+        offset = 4 + (count * 12)
+        tags.each do |sig, bytes|
+          out << sig
+          out << [offset].pack("N")
+          out << [bytes.bytesize].pack("N")
+          offset += bytes.bytesize
+        end
+        out
+      end
+
+      def tags_total_size(tags)
+        tags.values.sum(&:bytesize)
+      end
+
+      # ---- Tag element builders ----
+
+      # multiLocalizedUnicodeType (mluc) for desc, per ICC.10.13.
+      def desc_tag_bytes(text)
+        utf16 = text.encode("UTF-16BE").bytes
+        body = +""
+        body << "mluc"                              # type signature
+        body << ([0] * 4).pack("C*")                # reserved
+        body << [1].pack("N")                        # number of records
+        body << [12].pack("N")                       # record size in bytes
+        body << "enUS" # language + country
+        body << [utf16.length].pack("N")             # string length in bytes
+        body << [28].pack("N")                       # string offset (from tag start)
+        body << utf16.pack("C*")
+        body
+      end
+
+      # XYZType for colorants and white points, per ICC.10.22.
+      def xyz_tag_bytes(xyz_triplet)
+        body = +""
+        body << "XYZ "                              # type signature
+        body << ([0] * 4).pack("C*")                # reserved
+        xyz_triplet.each { |v| body << s15fixed16(v) }
+        body
+      end
+
+      # curveType with a single gamma value, per ICC.10.16.
+      # 2.2 is the canonical sRGB approximation (real sRGB uses a
+      # piecewise curve, but a single 2.2 gamma is acceptable for
+      # most rendering).
+      def curve_tag_bytes
+        body = +""
+        body << "curv"                              # type signature
+        body << ([0] * 4).pack("C*")                # reserved
+        body << [0].pack("N") # count = 0 → gamma follows in u8Fixed8
+        body << u8fixed8(2.2)
+        body
+      end
+
+      # Encode a Float as ICC s15Fixed16 (signed 16.16).
+      def s15fixed16(value)
+        v = (value * 65536.0).round.to_i
+        [v].pack("N")
+      end
+
+      # Encode a Float as ICC u8Fixed8 (unsigned 8.8).
+      def u8fixed8(value)
+        v = (value * 256.0).round.to_i
+        [v].pack("n")
+      end
+
+      # ---- sRGB colorants (D50-adapted, per ICC.1:2022 Annex A) ----
+
+      def d50_xyz
+        [0.9642, 1.0000, 0.8249]
+      end
+
+      def srgb_r_xyz
+        [0.4360, 0.2225, 0.0139]
+      end
+
+      def srgb_g_xyz
+        [0.3851, 0.7169, 0.0971]
+      end
+
+      def srgb_b_xyz
+        [0.1431, 0.0606, 0.7141]
+      end
+
+      # ---- Profile ID (MD5) ----
+
+      # Patch the Profile ID (offset 84..99) with the MD5 of the
+      # profile bytes computed with the ID slot zeroed (per ICC.1
+      # §7.1.1).
+      def patch_profile_id(profile)
+        copy = profile.b
+        copy[84, 16] = ([0] * 16).pack("C*")
+        md5 = Digest::MD5.digest(copy)
+        copy[84, 16] = md5
+        copy
+      end
+    end
+  end
+end

data/lib/pdfrb/color/icc_validator.rb

@@ -0,0 +1,69 @@
+# frozen_string_literal: true
+
+module Pdfrb
+  module Color
+    # Validates an ICC profile binary against the structural rules in
+    # ICC.1:2022 §7. The validator checks:
+    #   * Profile size >= 128 (header) + 4 (tag count)
+    #   * 'acsp' signature at offset 36
+    #   * Profile version is 2.x, 4.x, or 5.x
+    #   * Device class is one of the known classes
+    #   * Color space is one of the known spaces
+    #   * Tag table offsets don't exceed the file size
+    #
+    # Returns an Array of Violation strings (empty if valid).
+    module ICCValidator
+      VALID_VERSIONS = [(2.0..2.99), (4.0..4.99), (5.0..5.99)].freeze
+      VALID_DEVICE_CLASSES = %w[
+        scnr mntr prtr link spac abst nmcl nkpr
+      ].freeze
+      VALID_COLOR_SPACES = %w[
+        XYZ Lab Luv YCbCr Yxy RGB GRAY HSV HLS CMYK CMY 2CLR 3CLR
+        4CLR 5CLR 6CLR 7CLR 8CLR 9CLR ACLR BCLR CCLR DCLR ECLR FCLR
+      ].freeze
+
+      module_function
+
+      def validate(profile_bytes)
+        violations = []
+        unless profile_bytes.is_a?(::String) && profile_bytes.bytesize >= 132
+          violations << "ICC profile too small (need >= 132 bytes, got #{profile_bytes&.bytesize})"
+          return violations
+        end
+
+        unless profile_bytes.byteslice(36, 4) == "acsp"
+          violations << "ICC profile signature at offset 36 is not 'acsp'"
+        end
+
+        version_hi = profile_bytes.getbyte(8)
+        version_lo = profile_bytes.getbyte(9)
+        version = (version_hi * 1.0) + (version_lo / 16.0 / 10.0)
+        unless VALID_VERSIONS.any? { |range| range.cover?(version) }
+          violations << "ICC profile version #{version} not in 2.x/4.x/5.x"
+        end
+
+        device_class = profile_bytes.byteslice(12, 4).to_s.strip
+        unless VALID_DEVICE_CLASSES.include?(device_class)
+          violations << "ICC device class '#{device_class}' is not recognised"
+        end
+
+        color_space = profile_bytes.byteslice(16, 4).to_s.strip
+        unless VALID_COLOR_SPACES.include?(color_space)
+          violations << "ICC color space '#{color_space}' is not recognised"
+        end
+
+        # Tag table sanity check
+        tag_count = profile_bytes.unpack1("N", offset: 128)
+        if tag_count > 256
+          violations << "ICC tag count #{tag_count} looks bogus (max 256)"
+        end
+
+        violations
+      end
+
+      def valid?(profile_bytes)
+        validate(profile_bytes).empty?
+      end
+    end
+  end
+end

data/lib/pdfrb/color.rb

@@ -8,5 +8,7 @@ module Pdfrb
   module Color
     autoload :ICCProfile, "pdfrb/color/icc_profile"
     autoload :ColorSpace, "pdfrb/color/color_space"
+    autoload :DefaultProfile, "pdfrb/color/default_profile"
+    autoload :ICCValidator, "pdfrb/color/icc_validator"
   end
 end