format_parser 1.1.0 → 1.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e49394510e191185e3be1f3560ca182a103b9446225a1163294f2db6a912ae76
-  data.tar.gz: 4f52d761857020a0957711528ad4d10ee885c74745897fa14a5e69292abf0cc2
+  metadata.gz: d90a6eee951feb5017bdedc0fb6dd4be49fc0a7f218972c0fb423b1985bf9a97
+  data.tar.gz: ee6caab359b0e01450897d00abd1f190b131e40d842d487d1f29107ac3b374f0
 SHA512:
-  metadata.gz: 829bd488d021cb0db5ce2b718aa661bdf3d4f86a55e55808c8279367e955d35b2bdccdd2bdf805ff736cef6d49f8c0f425c1ec65c0d7f07ada2db684581d4c1f
-  data.tar.gz: ed76f04c157a865de03df0d614d0d353b53e1e9c9b0a164ac6b2a6612a358643ae210b00fd92b91fec6cefd8c8c71e6ff559e63df3343b13ce161175425dbd3a
+  metadata.gz: 72aca621c20dfb24443e32c52a3d27f64a4887c1254f6d2295f77b39ce57e1a5ef0aa52f365630badcb2e3aa90d32c3546c8fcea3b83e025e2c4fc3606dd2dd7
+  data.tar.gz: 90da48352579c4044035732fd1f837d86db1a2d3c8f325b60ee688630f6cb4ea8efdb2ffd7887ac4e49cf9f9df2a22fe2189187c32d606b8f027a0a3d6a3ec5f

data/CHANGELOG.md

@@ -1,3 +1,12 @@
+## 1.3.0
+* Add `heif_parser` and support for `HEIF` and `HEIC` formats. Exif parsing is still missing.
+* 
+## 1.2.1
+* Resolve bug when `stts` atom is `nil`
+
+## 1.2.0
+* Add support for `codecs` in moov_parser for video metadata
+
 ## 1.1.0
 * Add support for `frame_rate` in moov_parser
 

data/Gemfile

@@ -1,8 +1,4 @@
 source 'https://rubygems.org'
 
-gem 'ruby-debug-ide'
-gem 'debase'
-gem 'solargraph', group: :development
-
 # Gem dependencies specified in the gemspec
 gemspec

data/lib/format_parser/version.rb

@@ -1,3 +1,3 @@
 module FormatParser
-  VERSION = '1.1.0'
+  VERSION = '1.3.0'
 end

data/lib/io_utils.rb

@@ -30,5 +30,22 @@ module FormatParser::IOUtils
     nil
   end
 
+  def read_int_8
+    safe_read(@buf, 1).unpack('C').first
+  end
+
+  def read_int_16
+    safe_read(@buf, 2).unpack('n').first
+  end
+
+  def read_int_32
+    safe_read(@buf, 4).unpack('N').first
+  end
+
+  # 'n' is the number of bytes to read
+  def read_string(n)
+    safe_read(@buf, n)
+  end
+
   ### TODO: Some kind of built-in offset for the read
 end

data/lib/parsers/heif_parser.rb

@@ -0,0 +1,431 @@
+# HEIF stands for High-Efficiency Image File format, which is basically a container that is capable of storing an image, or a sequence of images in a single file.
+# There are a number of variants of HEIF, which can be used to store images, sequences of images, or videos using different codecs.
+# The variant that Apple uses to store images and sequences of images in its iOS and macOS operating systems is High Efficiency Image Coding (HEIC), which uses HEVC / H.265 for content compression.
+class FormatParser::HEIFParser
+  include FormatParser::IOUtils
+
+  HEIF_MARKER = [0x68, 0x65, 0x69, 0x63].pack('C4') # heif marker
+  FILE_TYPE_BOX_MARKER = [0x66, 0x74, 0x79, 0x70].pack('C4') # ftyp marker
+  META_BOX_MARKER = [0x6D, 0x65, 0x74, 0x61].pack('C4') # meta marker
+  MIF1_MARKER = [0x6D, 0x69, 0x66, 0x31].pack('C4') # mif1 marker
+  MSF1_MARKER = [0x6D, 0x73, 0x66, 0x31].pack('C4') # msf1 marker
+  MEANINGLESS_BYTE = [0x00, 0x00, 0x00, 0x00].pack('C4')
+  HANDLER_MARKER = [0x68, 0x64, 0x6C, 0x72].pack('C4') # hdlr marker
+  ITEM_PROPERTIES_BOX = [0x69, 0x70, 0x72, 0x70].pack('C4') # iprp marker
+  ITEM_PROPERTIES_CONTAINER_BOX = [0x69, 0x70, 0x63, 0x6F].pack('C4') # ipco marker
+  IMAGE_SPATIAL_EXTENTS_BOX = [0x69, 0x73, 0x70, 0x65].pack('C4') # ispe marker
+  PIXEL_ASPECT_RATIO_BOX = [0x70, 0x61, 0x73, 0x70].pack('C4') # pasp marker
+  ITEM_INFO_BOX = [0x69, 0x69, 0x6E, 0x66].pack('C4') # iinf marker
+  ITEM_INFO_ENTRY = [0x69, 0x6E, 0x66, 0x65].pack('C4') # infe marker
+  MIME_MARKER = [0x6D, 0x69, 0x6D, 0x65].pack('C4') # mime marker
+  COLOUR_INFO_BOX = [0x63, 0x6F, 0x6C, 0x72].pack('C4') # colr marker
+  PIXEL_INFO_BOX = [0x70, 0x69, 0x78, 0x69].pack('C4') # pixi marker
+  RELATIVE_LOCATION_BOX = [0x72, 0x6C, 0x6F, 0x63].pack('C4') # rloc marker
+  CLEAN_APERTURE_BOX = [0x63, 0x6C, 0x61, 0x70].pack('C4') # clap marker
+  PRIMARY_ITEM_BOX = [0x70, 0x69, 0x74, 0x6D].pack('C4') # pitm marker
+  ITEM_PROPERTIES_ASSOCIATION_BOX = [0x69, 0x70, 0x6D, 0x61].pack('C4') # ipma marker
+  IMAGE_ROTATION_BOX = [0x69, 0x72, 0x6F, 0x74].pack('C4') # irot marker
+  HEADER_LENGTH = 8 # every box header has a length of 8 bytes
+  HEIC_MIME_POSSIBLE_TYPES = {
+    'heic' => :heic,
+    'heix' => :heix,
+    'heim' => :heim,
+    'heis' => :heis
+  }
+  HEIC_MIME_TYPE = 'image/heic'
+  HEIF_MIME_TYPE = 'image/heif'
+  # TODO: use the following when adding image-sequence parsing
+  # HEIC_SEQUENCE_MIME_TYPE = 'image/heic-sequence'
+  # HEIF_SEQUENCE_MIME_TYPE = 'image/heif-sequence'
+
+  def self.call(io)
+    new.call(io)
+  end
+
+  def call(io)
+    @buf                = FormatParser::IOConstraint.new(io)
+    @format             = nil
+    @@major_brand       = nil
+    @width              = nil
+    @height             = nil
+    @exif_data_frames   = []
+    @compatible_brands  = nil
+    @metadata_start_pos = 0
+    @metadata_end_pos   = 0
+    @handler_type       = nil
+    @sub_items          = nil
+    @pixel_aspect_ratio = nil
+    @colour_info        = nil
+    @pixel_info         = nil
+    @horizontal_offset  = nil
+    @vertical_offset    = nil
+    @clean_aperture     = nil
+    @primary_item_id    = 0
+    @item_props         = {}
+    @rotation           = 0
+    @item_props_idxs    = []
+    @content_type       = nil
+    scan
+  end
+
+  def scan
+    # All HEIC files must be conform to ISO/IEC 23008-12:2017
+    # Moreover, all HEIC files are conform to ISO/IEC 14496-12:2015 and should be conform to the Clause 4 of such spec.
+    # Files are formed as a series of objects, called boxes. All data is contained in such boxes.
+    # All boxes start with a header which defines both size and type.
+    # The size is the entire size of the box, including the size and type header, fields, and all contained boxes.
+    # The fields in the objects are stored with the most significant byte first, commonly known as network byte order or big-endian format.
+    # A HEIC file must contain a File Type Box (ftyp).
+    # A file conforms to all the requirements of the brands listed in the compatible_brands.
+    scan_file_type_box
+
+    # file may be identified by MIME type of Annex C of ISO/IEC 23008-12 if 'mif1' is the major brand or Annex D if 'msf1' is the major brand.
+    # the MIME indicates the nature and format of our assortment of bytes
+    # note particularly that the brand 'mif1' doesn't mandate a MovieBox ("moov").
+    # One or more brands must be included in the list of compatible brands
+    return if @compatible_brands.nil?
+    if @compatible_brands&.include?(MIF1_MARKER)
+      scan_meta_level_box
+      if @major_brand == MIF1_MARKER
+        @content_type = HEIF_MIME_TYPE
+        @format = :heif
+      elsif (@compatible_brands & HEIC_MIME_POSSIBLE_TYPES.keys).length > 0
+        @format = :heic
+        @content_type = HEIC_MIME_TYPE
+      end
+    end
+    if @compatible_brands&.include?(MSF1_MARKER)
+      # TODO
+    end
+
+    result = FormatParser::Image.new(
+      format: @format,
+      width_px: @width,
+      height_px: @height,
+      intrinsics: {
+        compatible_brands: @compatible_brands,
+        handler_type: @handler_type,
+        # 'sub_items': @sub_items, # enable this if you want to output all the sub-items in the image
+        pixel_aspect_ratio: @pixel_aspect_ratio,
+        colour_info: @colour_info,
+        pixel_info: @pixel_info,
+        horizontal_offset: @horizontal_offset,
+        vertical_offset: @vertical_offset,
+        clean_aperture: @clean_aperture,
+        rotation: @rotation
+      },
+      content_type: @content_type
+    )
+
+    result
+  end
+
+  def scan_file_type_box
+    file_type_box_length = read_int_32
+    return unless read_string(4) == FILE_TYPE_BOX_MARKER
+    @major_brand = read_string(4)
+    return unless @major_brand == HEIF_MARKER || @major_brand == MIF1_MARKER
+    read_string(4) # minor_brand
+
+    # Subtracting from the total length of the box specified in the header the size header itself (8 bytes = header length and length of ftyp)
+    # and the length of the major and minor brand, we obtain the compatible brands
+    data_left_length = file_type_box_length - HEADER_LENGTH - HEIF_MARKER.length - 4
+
+    @compatible_brands = []
+    (data_left_length / 4).times do
+      @compatible_brands << read_string(4)
+    end
+  end
+
+  def scan_meta_level_box
+    metadata_length = read_int_32
+    return unless read_string(4) == META_BOX_MARKER
+    @metadata_start_pos = @buf.pos
+    @metadata_end_pos = @buf.pos + metadata_length - HEADER_LENGTH # the real data is always without the 8 initial bytes of the handler
+    read_nil_version_and_flag
+
+    # we are looking for box/containers right beneath the Meta box
+    # we start with the HDLR (Handler) box..
+    handler_length = read_int_32
+    return unless read_string(4) == HANDLER_MARKER
+    handler_length -= HEADER_LENGTH # subtract the header as usual (will not be mentioned anymore from now on)
+    handler_start = @buf.pos
+    # the handler type declares the type of metadata and thus the process by which the media-data in the track is presented
+    # it also indicates the structure or format of the ‘meta’ box contents
+    read_nil_version_and_flag
+    read_string(4) # pre_defined bytes, always 4 null bytes in the hdlr box
+    @handler_type = read_string(4)
+    @buf.seek(handler_start + handler_length) # the remaining part is reserved
+
+    # ..continue looking for the IINF box and especially for the IPRP box, containing info about the image itself
+    next_box_length = read_int_32
+    next_box = read_string(4)
+    next_box_start_pos = @buf.pos
+    while @buf.pos < @metadata_end_pos # we iterate over all next incoming boxed but without going outside the meta-box
+      case next_box
+      when PRIMARY_ITEM_BOX
+        read_primary_item_box
+      when ITEM_INFO_BOX
+        read_item_info_box
+      when ITEM_PROPERTIES_BOX
+        read_item_properties_box
+        fill_primary_values
+      when next_box == ''
+        break
+      end
+      next_box_length, next_box, next_box_start_pos = get_next_box(next_box_start_pos, next_box_length, @metadata_end_pos)
+    end
+  end
+
+  def read_item_info_box
+    version = read_int_8
+    safe_skip(@buf, 3) # 0 flags
+    entry_count = if version == 0
+      read_int_16
+    else
+      read_int_32
+    end
+    @sub_items = []
+    entry_count.times {
+      item_info_entry_length = read_int_32
+      return unless read_string(4) == ITEM_INFO_ENTRY
+      item_info_end_pos = @buf.pos + item_info_entry_length - HEADER_LENGTH
+      version = read_int_8
+      safe_skip(@buf, 3) # 0 flags
+      case version
+      when 2
+        item_id = read_int_16
+      when 3
+        item_id = read_int_32
+      else
+        return # wrong version according to standards, hence return
+      end
+      safe_skip(@buf, 2) # not interested in the item_protection_index
+      item_type = read_string(4)
+      content_encoding = ''
+      if item_type == MIME_MARKER
+        content_encoding = read_string(item_info_end_pos - @buf.pos).delete!("\0") # remove the null-termination part for output visualization reason
+      end
+      @sub_items << {item_id: item_id, item_type: item_type, content_encoding: content_encoding}
+      @buf.seek(item_info_end_pos) # we are not interested in anything else, go directly to the end of this 'infe' box
+    }
+  end
+
+  def read_nil_version_and_flag
+    safe_skip(@buf, 1) # version, always 0 in this current box
+    safe_skip(@buf, 3) # flags, always 0 in this current box
+  end
+
+  def read_primary_item_box
+    version = read_int_8
+    safe_read(@buf, 3) # flags, always 0 in this current box
+    @primary_item_id = if version == 0
+      read_int_16
+    else
+      read_int_32
+    end
+  end
+
+  # the ITEM_PROPERTIES_CONTAINER_BOX contains an implicitely 1-based index list of item properties.
+  # While parsing such box we are storing the properties with its own index.
+  # Reason behind is that the primary_item will be associated to some of these properties through the same index
+  # and in order to output relevant data from the format_parser we need all the properties associated to the primary_item.
+  # Hence the need of the association between an item and its properties, found in the ITEM_PROPERTIES_ASSOCIATION_BOX
+  def read_item_properties_box
+    ipco_length = read_int_32
+    return unless read_string(4) == ITEM_PROPERTIES_CONTAINER_BOX
+    read_item_properties_container_box(ipco_length)
+    read_int_32 # ipma_length
+    return unless read_string(4) == ITEM_PROPERTIES_ASSOCIATION_BOX
+    read_item_properties_association_box
+  end
+
+  def read_item_properties_container_box(box_length)
+    end_of_ipco_box = @buf.pos + box_length - HEADER_LENGTH
+    item_prop_length = read_int_32
+    item_prop_name = read_string(4)
+    item_prop_start_pos = @buf.pos
+    item_prop_index = 1
+    while @buf.pos < end_of_ipco_box
+      case item_prop_name
+      when IMAGE_SPATIAL_EXTENTS_BOX
+        read_nil_version_and_flag
+        width = read_int_32
+        height = read_int_32
+        @item_props[item_prop_index] = {
+          type: IMAGE_SPATIAL_EXTENTS_BOX,
+          width: width,
+          height: height
+        }
+      when PIXEL_ASPECT_RATIO_BOX
+        h_spacing = read_int_32
+        v_spacing = read_int_32
+        pixel_aspect_ratio = "#{h_spacing}/#{v_spacing}"
+        @item_props[item_prop_index] = {
+          type: PIXEL_ASPECT_RATIO_BOX,
+          pixel_aspect_ratio: pixel_aspect_ratio
+        }
+      when COLOUR_INFO_BOX
+        colour_info = {
+          colour_primaries: read_int_16,
+          transfer_characteristics: read_int_16,
+          matrix_coefficients: read_int_16
+        }
+        @item_props[item_prop_index] = {
+          type: COLOUR_INFO_BOX,
+          colour_info: colour_info
+        }
+      when PIXEL_INFO_BOX
+        pixel_info = []
+        read_nil_version_and_flag
+        num_channels = read_int_8
+        channel = 1
+        while channel <= num_channels
+          channel += 1
+          pixel_info << {
+            "bits_in_channel_#{channel}": read_int_8
+          }
+        end
+        @item_props[item_prop_index] = {
+          type: PIXEL_INFO_BOX,
+          pixel_info: pixel_info
+        }
+      when RELATIVE_LOCATION_BOX
+        read_nil_version_and_flag
+        horizontal_offset = read_int_32
+        vertical_offset = read_int_32
+        @item_props[item_prop_index] = {
+          type: RELATIVE_LOCATION_BOX,
+          horizontal_offset: horizontal_offset,
+          vertical_offset: vertical_offset
+        }
+      when CLEAN_APERTURE_BOX
+        clean_aperture = []
+        clean_aperture << {
+          clean_aperture_width_n: read_int_32,
+          clean_aperture_width_d: read_int_32,
+          clean_aperture_height_n: read_int_32,
+          clean_aperture_height_d: read_int_32,
+          horiz_off_n: read_int_32,
+          horiz_off_d: read_int_32,
+          vert_off_n: read_int_32,
+          vert_off_d: read_int_32
+        }
+        @item_props[item_prop_index] = {
+          type: CLEAN_APERTURE_BOX,
+          clean_aperture: clean_aperture
+        }
+      when IMAGE_ROTATION_BOX
+        read_nil_version_and_flag
+        binary = convert_byte_to_binary(read_int_8)
+        # we need only the last 2 bits to retrieve the angle multiplier. angle multiplier * 90 specifies the angle
+        rotation = binary.slice(6, 2).join.to_i(2) * 90
+        @item_props[item_prop_index] = {
+          type: IMAGE_ROTATION_BOX,
+          rotation: rotation
+        }
+      end
+      item_prop_length, item_prop_name, item_prop_start_pos = get_next_box(item_prop_start_pos, item_prop_length, end_of_ipco_box)
+      item_prop_index += 1
+    end
+  end
+
+  def read_item_properties_association_box
+    version = read_int_8
+    safe_read(@buf, 2) # we skip the first 2 bytes of the flags (total of 3 bytes) cause we care only about the least significant bit
+    flags = read_int_8
+    entry_count = read_int_32
+    item_id = 0
+    entry_count.times do
+      item_id = if version == 0
+        read_int_16
+      else
+        read_int_32
+      end
+
+      association_count = read_int_8
+      association_count.times do
+        # we need to retrieve the "essential" bit wich is just the first bit in the next byte
+        binary = convert_byte_to_binary(read_int_8)
+        # essential_bit = binary[0] # uncomment if needed
+        binary.concat(convert_byte_to_binary(read_int_8)) if (flags & 1) == 1 # if flag is 1 we need the next 15 bits instead of only the next 7 bits
+        # we need to nullify the 1st bit since that one was the essential bit and doesn't count now to calculate the property index
+        binary[0] = 0
+        item_property_index = binary.join.to_i(2)
+        # we are interested only in the primary item properties
+        @item_props_idxs << item_property_index if item_id == @primary_item_id
+      end
+
+      # we are interested only in the primary item
+      if item_id != @primary_item_id
+        next
+      else
+        return
+      end
+    end
+  end
+
+  def fill_primary_values
+    @item_props_idxs.each { |x|
+      next if @item_props[x].nil?
+      prop = @item_props[x]
+      case prop[:type]
+      when IMAGE_SPATIAL_EXTENTS_BOX
+        @width = prop[:width]
+        @height = prop[:height]
+      when PIXEL_ASPECT_RATIO_BOX
+        @pixel_aspect_ratio = prop[:pixel_aspect_ratio]
+      when COLOUR_INFO_BOX
+        @colour_info = prop[:colour_info]
+      when PIXEL_INFO_BOX
+        @pixel_info = prop[:pixel_info]
+      when RELATIVE_LOCATION_BOX
+        @horizontal_offset = prop[:horizontal_offset]
+        @vertical_offset = prop[:vertical_offset]
+      when CLEAN_APERTURE_BOX
+        @clean_aperture = prop[:clean_aperture]
+      when IMAGE_ROTATION_BOX
+        @rotation = prop[:rotation]
+      end
+    }
+  end
+
+  def next_meaningful_meta_byte
+    while @buf.pos < @metadata_end_pos
+      next_byte = read_string(4)
+      return next_byte if meaningful?(next_byte)
+    end
+  end
+
+  def get_next_box(box_start_pos, box_length, end_pos_upper_box)
+    skip_pos = box_start_pos + box_length - HEADER_LENGTH
+    @buf.seek(skip_pos)
+    return if skip_pos >= end_pos_upper_box
+    next_box_length = read_int_32
+    next_box_name = read_string(4)
+    [next_box_length, next_box_name, @buf.pos]
+  end
+
+  def meaningful?(byte)
+    byte != MEANINGLESS_BYTE
+  end
+
+  def convert_byte_to_binary(integer)
+    binary = []
+    while integer > 0
+      binary << integer % 2
+      integer /= 2
+    end
+    binary_value = binary.reverse
+    (8 - binary_value.length).times do
+      binary_value.prepend('0')
+    end
+    binary_value
+  end
+
+  def likely_match?(filename)
+    filename =~ /\.hei[cf]$/i
+  end
+
+  FormatParser.register_parser(new, natures: :image, formats: [:heif, :heic], priority: 2)
+end

data/lib/parsers/moov_parser/decoder.rb

@@ -130,6 +130,23 @@ class FormatParser::MOOVParser::Decoder
     stts
   end
 
+  def parse_stsd_atom(io, _)
+    version = read_byte_value(io)
+    is_v1 = version == 1
+    stsd = {
+      version: version,
+      flags: read_bytes(io, 3),
+      number_of_entries: is_v1 ? read_64bit_uint(io) : read_32bit_uint(io),
+      codecs: []
+    }
+    stsd[:number_of_entries].times {
+      codec_length = read_32bit_uint(io)
+      stsd[:codecs] << read_bytes(io, 4)
+      io.seek(io.pos + codec_length - 8) # 8 bytes is the header length containing the codec length and the codec name that we just did read
+    }
+    stsd
+  end
+
   def parse_mdhd_atom(io, _)
     version = read_byte_value(io)
     is_v1 = version == 1

data/lib/parsers/moov_parser.rb

@@ -64,14 +64,14 @@ class FormatParser::MOOVParser
         intrinsics: atom_tree,
       )
     else
-      frame_rate = parse_sample_atom(decoder, atom_tree)&.truncate(2)
       FormatParser::Video.new(
         format: format_from_moov_type(file_type),
         width_px: width,
         height_px: height,
-        frame_rate: frame_rate,
+        frame_rate: parse_time_to_sample_atom(decoder, atom_tree)&.truncate(2),
         media_duration_seconds: media_duration_s,
         content_type: MP4_MIXED_MIME_TYPE,
+        codecs: parse_sample_description_atom(decoder, atom_tree),
         intrinsics: atom_tree
       )
     end
@@ -119,7 +119,7 @@ class FormatParser::MOOVParser
 
   # Sample information is found in the 'time-to-sample' stts atom.
   # The media atom mdhd is needed too in order to get the movie timescale
-  def parse_sample_atom(decoder, atom_tree)
+  def parse_time_to_sample_atom(decoder, atom_tree)
     video_trak_atom = decoder.find_video_trak_atom(atom_tree)
 
     stts = if video_trak_atom
@@ -136,7 +136,7 @@ class FormatParser::MOOVParser
 
     if stts && mdhd
       timescale = mdhd.atom_fields[:tscale]
-      sample_duration = stts.field_value(:entries).first[:sample_duration]
+      sample_duration = stts.field_value(:entries).dig(0, :sample_duration)
       if timescale.nil? || timescale == 0 || sample_duration.nil? || sample_duration == 0
         nil
       else
@@ -147,5 +147,21 @@ class FormatParser::MOOVParser
     end
   end
 
-  FormatParser.register_parser new, natures: :video, formats: FTYP_MAP.values, priority: 1
+  def parse_sample_description_atom(decoder, atom_tree)
+    video_trak_atom = decoder.find_video_trak_atom(atom_tree)
+
+    stsd = if video_trak_atom
+      decoder.find_first_atom_by_path([video_trak_atom], 'trak', 'mdia', 'minf', 'stbl', 'stsd')
+    else
+      decoder.find_first_atom_by_path(atom_tree, 'moov', 'trak', 'mdia', 'minf', 'stbl', 'stsd')
+    end
+
+    if stsd
+      stsd.field_value(:codecs)
+    else
+      nil
+    end
+  end
+
+  FormatParser.register_parser new, natures: :video, formats: FTYP_MAP.values, priority: 3
 end

data/lib/parsers/pdf_parser.rb

@@ -21,5 +21,5 @@ class FormatParser::PDFParser
     FormatParser::Document.new(format: :pdf, content_type: PDF_CONTENT_TYPE)
   end
 
-  FormatParser.register_parser new, natures: :document, formats: :pdf, priority: 1
+  FormatParser.register_parser new, natures: :document, formats: :pdf, priority: 3
 end

data/lib/parsers/zip_parser.rb

@@ -60,5 +60,5 @@ class FormatParser::ZIPParser
     end
   end
 
-  FormatParser.register_parser new, natures: [:archive, :document], formats: :zip, priority: 2
+  FormatParser.register_parser new, natures: [:archive, :document], formats: :zip, priority: 4
 end

data/lib/video.rb

@@ -13,6 +13,8 @@ module FormatParser
     # Type of the file (e.g :mp3)
     attr_accessor :format
 
+    attr_accessor :codecs
+
     # Duration of the media object (be it audio or video) in seconds,
     # as a Float
     attr_accessor :media_duration_seconds

data/spec/parsers/heif_parser_spec.rb

@@ -0,0 +1,75 @@
+require 'spec_helper'
+
+describe FormatParser::HEIFParser do
+  it 'is able to parse single heif image with heic major brand' do
+    heif_path = fixtures_dir + 'HEIF/SingleImage.heic'
+
+    result = subject.call(File.open(heif_path, 'rb'))
+    expect(result).not_to be_nil
+    expect(result.nature).to eq(:image)
+    expect(result.format).to eq(:heic)
+    expect(result.width_px).to eq(4000)
+    expect(result.height_px).to eq(3000)
+    expect(result.content_type).to eq('image/heic')
+    expect(result.intrinsics[:compatible_brands].should =~ ['mif1', 'heic'])
+  end
+
+  it 'is able to parse single heif image with mif1 major brand' do
+    heif_path = fixtures_dir + 'HEIF/SingleImage_Autumn.heic'
+
+    result = subject.call(File.open(heif_path, 'rb'))
+    expect(result).not_to be_nil
+    expect(result.nature).to eq(:image)
+    expect(result.format).to eq(:heif)
+    expect(result.width_px).to eq(1440)
+    expect(result.height_px).to eq(960)
+    expect(result.content_type).to eq('image/heif')
+    expect(result.intrinsics[:compatible_brands].should =~ ['mif1', 'heic'])
+  end
+
+  it 'is able to parse image collection with mif1 major brand' do
+    heif_path = fixtures_dir + 'HEIF/ImageCollection.heic'
+
+    result = subject.call(File.open(heif_path, 'rb'))
+    expect(result).not_to be_nil
+    expect(result.nature).to eq(:image)
+    expect(result.format).to eq(:heif)
+    expect(result.width_px).to eq(1440)
+    expect(result.height_px).to eq(960)
+    expect(result.content_type).to eq('image/heif')
+  end
+
+  it 'is able to parse image collection with colour info' do
+    heif_path = fixtures_dir + 'HEIF/SingleImage_Autumn_WithColourInfo.heic'
+
+    result = subject.call(File.open(heif_path, 'rb'))
+    expect(result).not_to be_nil
+    expect(result.nature).to eq(:image)
+    expect(result.format).to eq(:heic)
+    expect(result.width_px).to eq(1440)
+    expect(result.height_px).to eq(960)
+    colour_info = result.intrinsics[:colour_info]
+    expect(colour_info[:colour_primaries]).to eq(28259)
+    expect(colour_info[:transfer_characteristics]).to eq(27768)
+    expect(colour_info[:matrix_coefficients]).to eq(2)
+    expect(result.content_type).to eq('image/heic')
+    expect(result.intrinsics[:compatible_brands].should =~ ['mif1', 'heic'])
+  end
+
+  it 'is able to parse image collection with pixel info' do
+    heif_path = fixtures_dir + 'HEIF/SingleImage_Autumn_WithColourInfo.heic'
+
+    result = subject.call(File.open(heif_path, 'rb'))
+    expect(result).not_to be_nil
+    expect(result.nature).to eq(:image)
+    expect(result.format).to eq(:heic)
+    expect(result.width_px).to eq(1440)
+    expect(result.height_px).to eq(960)
+    pixel_info = result.intrinsics[:pixel_info]
+    expect(pixel_info[0][:bits_in_channel_2]).to eq(8)
+    expect(pixel_info[1][:bits_in_channel_3]).to eq(8)
+    expect(pixel_info[2][:bits_in_channel_4]).to eq(8)
+    expect(result.content_type).to eq('image/heic')
+    expect(result.intrinsics[:compatible_brands].should =~ ['mif1', 'heic'])
+  end
+end

data/spec/parsers/moov_parser_spec.rb

@@ -94,6 +94,7 @@ describe FormatParser::MOOVParser do
     expect(result.format).to eq(:mov)
     expect(result.width_px).to eq(1920)
     expect(result.height_px).to eq(1080)
+    expect(result.codecs).to eq(['apcn'])
   end
 
   it 'parses an MP4 video file and provides the necessary metadata' do
@@ -107,6 +108,7 @@ describe FormatParser::MOOVParser do
     expect(result.width_px).to eq(160)
     expect(result.height_px).to eq(90)
     expect(result.frame_rate).to eq(14.98)
+    expect(result.codecs).to eq(['avc1'])
   end
 
   it 'provides filename hints' do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: format_parser
 version: !ruby/object:Gem::Version
-  version: 1.1.0
+  version: 1.3.0
 platform: ruby
 authors:
 - Noah Berman
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2022-04-01 00:00:00.000000000 Z
+date: 2022-05-31 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ks
@@ -238,6 +238,7 @@ files:
 - lib/parsers/fdx_parser.rb
 - lib/parsers/flac_parser.rb
 - lib/parsers/gif_parser.rb
+- lib/parsers/heif_parser.rb
 - lib/parsers/jpeg_parser.rb
 - lib/parsers/m3u_parser.rb
 - lib/parsers/moov_parser.rb
@@ -278,6 +279,7 @@ files:
 - spec/parsers/fdx_parser_spec.rb
 - spec/parsers/flac_parser_spec.rb
 - spec/parsers/gif_parser_spec.rb
+- spec/parsers/heif_parser_spec.rb
 - spec/parsers/jpeg_parser_spec.rb
 - spec/parsers/m3u_parser_spec.rb
 - spec/parsers/moov_parser_spec.rb
@@ -315,7 +317,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.4
+rubygems_version: 3.1.6
 signing_key: 
 specification_version: 4
 summary: A library for efficient parsing of file metadata